Post on 22-Apr-2023
int j remote sensing 2002 vol 23 no 20 4403ndash4438
Astronaut-acquired orbital photographs as digital data for remotesensing spatial resolution
JULIE A ROBINSONdagger DAVID L AMSBURYDaggerDONN A LIDDLEdagger and CYNTHIA A EVANSdagger
daggerEarth Sciences and Image Analysis Laboratory NASA Johnson Space Centerand Lockheed Martin Space Operations 2400 NASA Road 1 C23 HoustonTX 77058 USADagger128 Homestead Kerrville TX 78028 USA
(Received 15 November 2000 in nal form 9 August 2001 )
Abstract Astronaut-acquired orbital photographs (astronaut photographs) area useful complement to images taken by orbiting satellites They are in the publicdomain and have been particularly useful for scientists in developing countriesas supplementary low-cloud data and for studies requiring large numbers ofimages Depending on camera lens and look angle digitized astronaut photo-graphs can have pixel sizes representing areas on the Earth as small as 10 m orless although most photographs suitable for digital remote sensing have pixelsizes between 30 m and 60 m The objective of this paper is to provide a practicalreference for scientists in a variety of disciplines who want to use astronautphotographs as remote sensing data The characteristics of astronaut photographysystems that in uence spatial resolution are detailed and previous image acqui-sitions relative to these elements are summarized Methods are presented forestimating ground coverage under three diVerent levels of assumptions to meetaccuracy needs of diVerent users Of the more than 375 000 photographs takento date at least half have the potential to be used as a source of digital remotesensing data
1 Introduction11 Overview of NASA astronaut photography
Astronaut photography of Earth is produced and archived by the NationalAeronautics amp Space Administration (NASA) and provides an important record ofthe state of the Earth that has not been used to its potential (Lulla et al 1996 ) Thepractice was the foundation for the development of other forms of orbital remotesensing (Lowman 1999) Although the geometry is more complex than that of avertical aerial photograph astronaut photographs still provide information that canbe interpreted by knowledgeable observers (Ring and Eyre 1983 Lowman 1985Rasher and Weaver 1990 Drury 1993 Campbell 1996 121ndash156 Arnold 1997)
OYcial NASA campaigns of terrain ocean and atmospheric photography werecarried out during the Gemini missions (Underwood 1967 Lowman and Tiedemann
email juliearobinson1jscnasagov
Internationa l Journal of Remote SensingISSN 0143-1161 printISSN 1366-590 1 online copy 2002 US Crown
httpwwwtandfcoukjournalsDOI 10108001431160110107798
J A Robinson et al4404
1971) the Earth-orbiting Apollo missions (Colwell 1971) the Apollo-Soyuz mission(El-Baz 1977 El-Baz and Warner 1979) Skylab (NASA 1974 Wilmarth et al 1977 )a few Shuttle missions (eg the two Space Radar Laboratory missions of 1994 Joneset al 1996) and the Shuttle-Mir missions (Evans et al 2000) Extensive training inphotography is available to members of all ight crews during their general trainingperiod and during intensive training for speci c missions (Jones et al 1996) Mostphotographs have been taken by astronauts on a time-available basis Astronautphotographs are thus a subset of the potential scenes selected both by opportunity(orbital parameters lighting and crew workloads and schedules) and by the trainingexperience and interest of the photographers
As remote sensing and geographic information systems have become more widelyavailable tools we have collaborated with a number of scientists interested in usingastronaut photography for quantitative remote sensing applications Digitized imagesfrom lm are suitable for geometric recti cation and image enhancement followed byclassi cation and other remote sensing techniques (Lulla and Helfert 1989 Mohler et al1989 Helfert et al 1990 Lulla et al 1991 Eckardt et al 2000 Robinson et al 2000a2000c and in press Webb et al in press) The photographs are in the public domainthus they provide a low-cost alternative data source for cases where commercial imagerycannot be acquired Such cases often include studies in developing countries in areasthat have not usually been targets for major satellites needing supplemental low-clouddata requiring a time series or requiring a large number of images
12 Extent of the datasetAs of 30 September 1999 378 461 frames were included in the photograph
database (OYce of Earth Sciences 2000) comprising 99 missions from Mercury 3(21 July 1961) through STS-96 (27 May to 6 June 1999) The database was reviewedremoving photographs that were deemed unsuitable for remote sensing analysisbecause they were not Earth-looking (no entry for tilt angle) had no estimated focallength were over- or underexposed or were taken at oblique tilt angles (furtherdiscussion of these characteristics follows) The remaining 190 911 frames (504 ofthe records present in the database) represent photographs that are potentiallysuitable for remote sensing data As the database is always having new photographyadded statistics can be updated on request using the web link lsquoSummary of DatabaseContentsrsquo at the Gateway to Astronaut Photography of Earth (OYce of EarthSciences 2000)
13 ObjectivesThe overall purpose of this paper is to provide understanding of the properties of
astronaut-acquired orbital photographs to help scientists evaluate their applicabilityfor digital remote sensing Previous general descriptions of astronaut photographyhave been much less extensive and have tended to focus on numbers of photographstaken and geographic coverage and emphasized interpretative applications (Helfertand Wood 1989 Lulla et al 1993 1994 1996) By providing a unique compilationof the background information necessary to extend the use of astronaut photographybeyond interpretation to rigorous analysis it is hoped to provide a resource forscientists who could use this data in a variety of disciplines In keeping with apotential interdisciplinary audience we have tried to provide suYcient backgroundinformation for scientists who do not have extensive training in remote sensing
We focus on spatial resolution because it is one of the most important factors
Astronaut photography as digital data for remote sensing 4405
determining the suitability of an image for a remote sensing objective In remotesensing literature spatial resolution for aerial photography is often treated verydiVerently from spatial resolution of multispectral scanning sensors To providesuYcient background for a discussion of spatial resolution for a data source thatshares properties with both aerial photography and satellite remote sensing we rst summarize the ways that spatial resolution is typically quanti ed for aerialphotography and satellite remote sensors
Given this background information we (1 ) describe and illustrate factors thatin uence the spatial resolution of astronaut photographs (2) show examples ofestimating spatial resolution for a variety of photographs in a way that willenable users to make these calculations whether or not they have a background inphotogrammetry and (3) compare spatial resolution of digital data extracted fromastronaut photographs to data obtained from other satellites
2 BackgroundmdashSpatial resolution of imaging systemsSpatial resolution is a fundamental property of any imaging system used to
collect remote sensing data and directly determines the spatial scale of the resultantinformation Resolution seems intuitively obvious but its technical de nition andprecise application in remote sensing have been complex For example Townshend(1980) summarized 13 diVerent ways to estimate the resolving power of the LandsatMultiSpectral Scanner (MSS) Simonett (1983 p 20) stated lsquoIn the simplest casespatial resolution may be de ned as the minimum distance between two objects thata sensor can record distinctly[but] it is the format of the sensor system thatdetermines how spatial resolution is measuredrsquo By focusing attention on the proper-ties of the system (and not the images acquired) this de nition can be applied to avariety of types of images provided by diVerent systems In order to provide ascomplete a treatment as possible background is provided on several views of spatialresolution that are relevant when considering the spatial resolution of astronautphotography
21 Ground resolved distancePhotogrammetrists were measuring spatial resolution of aerial photographs using
empirical methods long before the rst scanning sensor was placed in orbit Thesemethods integrated properties of the imaging system with the other external factorsthat determine spatial resolution Ground resolved distance (GRD) is the parameterof most interest because it measures the applicability of an image to a speci c taskThe GRD of an image is de ned as the dimensions of the smallest discernible objectThe GRD is a function of geometry (altitude focal length of optics) equipment(internal system spatial resolution of the camera or scanner) and also on re ectancecharacteristics of the object compared to its surroundings (contrast)
Performance of a lm lm and camera or deployed aerial photography systemis measured empirically using standard targets that consist of black-and-white barsof graduated widths and spacings ( gure 6ndash9 in Slater et al 1983) The area-weightedaverage resolution (AWAR) in lp mm Otilde 1 ( line pairs mm Otilde 1 ) at the lm plane is deter-mined by measuring the smallest set of line pairs that can be discriminated on anoriginal lm negative or transparency Line pairs are quoted because it is necessaryto discriminate between one object and another to detect it and measure it
Film resolving power (in lp mm Otilde 1 ) is measured by manufacturers under standardphotographi c conditions (Smith and Anson 1968 Eastman Kodak Company 1998)
J A Robinson et al4406
at high contrast (objectbackground ratio 10001) and low contrast (objectback-ground ratio 161) Most terrestrial surfaces recorded from orbit are low contrastfor the purpose of estimating resolving power of lm Kodak no longer measures lm resolving power for non-aerial photographic lms (Karen Teitelbaum EastmanKodak Company personal communication) including lms that NASA routinelyuses for Earth photography
AWAR can be measured for the static case of lm and camera or for thecamerandashaircraft system in motion For example AWAR for the National AerialPhotography Program includes eVects of the lens resolving power of original lmimage blur due to aircraft motion and spatial resolution of duplicating lm (Light1996) Given AWAR for a system in motion the GRD can be calculated bytrigonometry (see equation (3) in sect5 d=1AWAR and D=GRD)
An impediment to similar rigorous measurement of GRD for orbital remotesensing systems is the lack of a target of suitable scale on the ground thus spatialresolution for most orbiting sensors is described in terms of a less all-encompassingmeasure instantaneous eld of view (see below) Additional challenges to measuringAWAR for a complete astronaut photography system include the number of diVerentoptions for aspects of the system including diVerent cameras lms and orbitalaltitudes These elements that must be standardized to determine AWAR providea useful list of those characteristics of astronaut photography that will mostin uence GRD
22 Instantaneous eld of viewThe instantaneous eld of view (IFOV) is generally used to represent the spatial
resolution of automated satellite systems and is commonly used interchangeablywith the term spatial resolution when comparing diVerent sensors IFOV is a combina-tion of geometric mechanical and electronic properties of the imaging systemGeometric properties including satellite orbital altitude detector size and the focallength of the optical system (Simonett 1983) The sensitivity of each detector elementat the wavelength desired plus the signal-to-noise level desired are electronic proper-ties that determine a minimum time for energy absorption For a linear sensor array(pushbroom scanning) this minimum time is translated to areal coverage by theforward velocity of the platform (Campbell 1996 p 97) Usually each detectorelement in the array corresponds to a pixel in the image Thus for a given altitudethe width of the pixel is determined by the optics and sensor size and the height ofthe pixel is determined by the rate of forward motion When magni ed by the ratioof the sensor altitude to the focal length of the optics of the sensor system IFOV isthe size of the area on the ground represented by an individual detector element(pixel Slater 1980 p 27)
The equation of IFOV with spatial resolution can be misleading because IFOVdoes not include factors other than geometrymdashfactors that largely determine thelevel of detail that can be distinguished in an image For example IFOV does notinclude characteristics of the target (contrast with surroundings shape of an objectcolour) atmospheric conditions illumination and characteristics of the interpreter(machine or human) Factors in uencing spatial resolution that are included in IFOVcan be calculated from design speci cations before the system has been built whereasthe actual spatial resolution of an image captured by the sensor will be unique tothat image IFOV represents the best spatial resolution possible given optimalconditions
Astronaut photography as digital data for remote sensing 4407
23 Imagery from scanners versus lm from camerasLike aerial photography systems the GRD of early remote sensing scanners
(electro-optical imaging systems) was calibrated empirically ( gures 1ndash6 in Simonett1983) but such methods are impractical for routine applications Satellites are nowcommon platforms for collecting remote sensing data small-scale aerial photographsare widely available within the USA and Europe and astronaut photographshave become available worldwide Straightforward comparison of resolutions ofphotographs to IFOVs of scanner images is necessary for many applications
How can spatial resolutions of data from diVerent sources be compared usingcommon units Oft-quoted lsquoresolution numbersrsquo actually IFOV of digital scanningsystems should be at least doubled for comparison to the standard method ofexpressing photographi c spatial resolution as GRD simply because of the objectcontrast requirement The following rule of thumb has been used in geographicapplications to compare spatial resolution of scanners and aerial photography (Welch1982 Jensen 1983) a low-contrast target may be converted to an approximate IFOVby dividing the GRD (of the photograph) by 24 Such a rule of thumb might alsobe reasonably applied to astronaut photography making it possible to convertobserved GRD to IFOV and IFOV to an estimated GRD
In this paper we discuss the spatial resolution of astronaut photographs in termsof system properties and compute the area on the ground represented by a singlepixel the equivalent to IFOV To complete our treatment of spatial resolution wealso provide estimates of the sizes of small features identi able in images for severalcases where this can be readily done
3 Factors that determine the footprint (area covered by the photograph )The most fundamental metric that forms the basis for estimating spatial resolution
of astronaut photographs is the size of the footprint or area on the ground capturedin a photograph (see review of satellite photogrammetry by Light 1980) The basicgeometric variables that in uence the area covered by an astronaut photograph are(1) the altitude of the orbit H (2) the focal length of the lens f (3) the actual sizeof the image on the lm d and (4) the orientation of the camera axis relative to theground (the obliquity or look angle t ) The relationships among these parametersare illustrated in gure 1 (also see equation (3) in sect5)
31 AltitudeHuman space ight missions have had a variety of primary objectives that required
diVerent orbital altitudes The higher the altitude the larger the footprint of thephotographs The diVering scale of photographs taken at diVerent altitudes is illus-trated in gure 2 The two photographs of Lake Eyre Australia were taken with thesame camera and lens but on diVerent dates and from diVerent altitudes In (a) thelake is relatively ooded while in (b) it is dry A 23timesdiVerence in altitude leads toa corresponding diVerence in the scales of the resulting photographs
32 L ensesThe longer the lens focal length the more magni cation the greater detail and
the smaller footprint A variety of lenses with diVerent focal lengths are own oneach space mission (table 1) The eVect of lens length on spatial coverage and imagedetail is shown in gure 3 In these views of Houston (a)ndash(c) taken from approxi-mately similar altitudes with the same camera most of the diVerence in scale of the
J A Robinson et al4408
Figure 1 Geometric relationship between look angle (t) spacecraft nadir point (SN) photo-graph centre point (PC) and photograph principal point (PP) The dark shaded arearepresents Earthrsquos surface The distance covered on the ground D corresponds totable 5 Original image size d is listed for various lms in table 1 Variables correspondto equations (2)ndash(9) in sect5
Astronaut photography as digital data for remote sensing 4409
(a) (b)
Figure 2 Example of the eVect of altitude on area covered in a photograph Both photographsof Lake Eyre Australia were taken using a Hasselblad camera with 100 mm lens(a) altitude 283 km STS093-702-062 (b) altitude 617 km STS082-754-61
photographs is due to the diVerent magni cations of the lenses Although taken froma diVerent altitude and using a diVerent camera with a diVerent original image size(see table 2) an electronic still camera image taken with a 300 mm lens is alsoincluded for comparison
For remote sensing longer focal length lenses are generally preferred (250 mm or350 mm for Hasselblad 300 mm or 400 mm lenses for 35 mm format cameras andESCs) Unfortunately longer-focal-length lenses exhibit poorer performance towardthe edge of a frame For example a 250 mm lens (Distagon CF 56) used on theHasselblad camera has spatial resolution of 57 lp mm Otilde 1 at the centre but only51 lp mm Otilde 1 at the edge and 46 lp mm Otilde 1 at the corner (tested with Ektachrome 5017f8 aperture and high contrast) A longer 300 mm lens used on the Nikon camerahas a greater spatial resolution diVerence between centre (82 lp mm Otilde 1 ) and corner(49 lp mm Otilde 1 ) using Kodak 5017 Ektachrome f4 aperture and high contrast FredPearce (unpublished data) There are tradeoVs among lens optics and speed Forexample the lenses for the Linhof system and the 250 mm lens for the Hasselblad(Distagon CF 56 see footnotes to table 1) are limited to apertures smaller thanf56 This becomes an important constraint in selecting shutter speeds (see discussionof shutter speed in sect42)
33 Cameras and actual image sizesAfter passing through the lens the photographic image is projected onto lm
inside the camera The size of this original image is another important propertydetermining spatial resolution and is determined by the camera used Cameraformats include 35 mm and 70 mm (Lowman 1980 Amsbury 1989) and occasionally5times4 inch (127times100 mm) and larger (table 1) Cameras own on each mission arenot metricmdashthey lack vacuum platens or reseau grids image-motion compensationor gyro-stabilized mounts The workhorse for engineering and Earth photographyon NASA missions has been a series of 70 mm Hasselblad cameras (table 1) chosenfor their reliability The modi ed magazine databack imprints a unique number and
J A Robinson et al4410
Tab
le1
Det
ails
on
the
cam
eras
use
dfo
rast
ron
aut
ph
oto
gra
phy
of
Eart
h
Data
incl
ud
ep
ho
togr
aphy
from
all
NA
SA
hu
man
spac
eig
ht
mis
sion
sth
rough
ST
S-9
6(J
un
e19
99
378
461
tota
lfr
am
esin
the
dat
abas
e)1
Imag
esi
zeT
ypic
al
lense
sN
um
ber
of
Cam
era
make
(mm
timesm
m)
use
d(m
m)
ph
oto
gra
ph
sP
erio
dof
use
No
tes
Has
selb
lad
55times
55
40
50100
2502
288
494
1963ndashp
rese
nt
Lin
ho
f10
5times120
90
250
29
373
1984ndashp
rese
nt
No
won
lyo
wn
occ
asio
nal
lyM
ult
ispec
tral
Cam
era
(S19
0A
)57times
57152
32
913
1973ndash19
74S
kyla
b
xed
-mou
nt3
cam
era
(NA
SA
1974
)E
art
hT
erra
inC
amer
a(S
190B
)11
5times11
5457
5478
1973ndash19
74S
kyla
b
xed
-mou
nt3
cam
era
(NA
SA
1974
)R
ole
iex
55times
55
50
100250
4352
1990ndash19
92L
arg
eF
orm
at
Cam
era4
229
times457
305
2143
1984
Mis
sio
nST
Sndash41G
only
(Mer
kel
etal
198
5)
Maure
r55
times55
76
80818
1961ndash19
68
1 Sm
all-f
orm
atca
mer
as
(35
mm
lm
and
ES
C)
are
no
tin
clud
edin
this
table
bec
ause
of
the
larg
enu
mb
erof
len
ses
and
con
gu
rati
on
sth
at
have
bee
nu
sed
and
bec
au
seth
ese
data
are
no
tcu
rren
tly
incl
uded
inth
edata
bas
efo
rall
mis
sion
s2 L
ense
suse
dfo
rH
ass
elbla
dm
anufa
cture
dby
Carl
Zei
ss
Dis
tago
nC
F4
(40
mm
)D
ista
gon
CF
4(5
0m
m)
Dis
tagon
CF
35
(100
mm
)D
ista
go
nC
F5
6(2
50
mm
)S
tand
ard
Has
selb
lad
len
ses
ow
nw
ill
chan
ge
inth
eyea
r20
00
toD
ista
gon
FE
28
(50
mm
)D
ista
go
nF
E2
(110
mm
)T
ele-
Tes
sar
FE
4(2
50m
m)
and
Tel
e-T
essa
rF
E4
(350
mm
)3 T
hes
eca
mer
as
wer
ein
xed
mou
nti
ngs
on
the
outs
ide
of
the
Skyla
bsp
ace
stati
on
A
ltho
ugh
not
hand
hel
d
the
lm
isca
talo
gued
wit
ho
ther
ast
ron
aut
ph
oto
gra
ph
yan
dis
incl
uded
her
efo
rco
mp
lete
nes
s4 A
NA
SA
-des
igned
cam
era
wit
hatt
itud
ere
fere
nce
syst
emfo
rd
eter
min
ing
pre
cise
nadir
poin
ts
Dat
ais
no
tcu
rren
tly
inco
rpora
ted
into
on
lin
ed
atab
ase
bu
tis
cata
loged
by
Mer
kel
etal
(1
985
)
Astronaut photography as digital data for remote sensing 4411
(a) (b)
(c) (d)
Figure 3 Example of the eVect of lens focal length on resolving power Hasselblad imagesof Houston were all taken from an altitude of 294 kmndash302 km with diVerent lenses(a) 40 mm STS062-97-151 (b) 100 mm STS094-737-28 (c) 250 mm STS055-71-41 Forcomparison an Electronic Still Camera image with 300 mm lens at 269 km altitude isalso shown (d ) S73E5145
timestamp on each frame at the time of exposure Cameras are serviced between ights Occasionally there is enough volume and mass allowance so that a Linhof5times4 inch (127times101 mm) format camera can be own Nikon 35 mm cameras are own routinely also because of proven reliability Electronic still cameras (ESC)were tested for Earth photography beginning in 1992 (Lulla and Holland 1993) Inan ESC a CCD (charge-coupled device) is used as a digital replacement for lmrecording the image projected inside the camera An ESC (consisting of a KodakDCS 460c CCD in a Nikon N-90S body) was added as routine equipment forhandheld photographs in 1995 ESCs have also been operated remotely to captureand downlink Earth images through a NASA-sponsored educational programme(EarthKAM) Discussion of CCD spatial array and radiometric sensitivity arebeyond the scope of this paper but are summarized by Robinson et al (2000b) The
J A Robinson et al4412
format of the lm (or CCD) and image size projected onto the lm (or CCD) aresummarized for all the diVerent cameras own in table 2
34 L ook angle or obliquityNo handheld photographs can be considered perfectly nadir they are taken at a
variety of look angles ranging from near vertical ( looking down at approximatelythe nadir position of the spacecraft ) to high oblique (images that include the curvatureof Earth) Imaging at oblique look angles leads to an image where scale degradesaway from nadir A set of views of the same area from diVerent look angles is shownin gure 4 The rst two shots of the island of Hawaii were taken only a few secondsapart and with the same lens The third photograph was taken on a subsequentorbit and with a shorter lens The curvature of the Earth can be seen in the upperleft corner
Obliquity can be described qualitatively ( gure 4) or quantitatively as the lookangle (t gure 1 calculations described in formulation 2 (sect52)) Because obliquityand look angle have such a dramatic in uence on the footprint we summarize thedatabase characteristics relative to these two parameters Figure 5 is a breakdownof the spatial resolution characteristics of low oblique and near vertical photographsin the NASA Astronaut Photography database Number of photographs are grouped(1) by calculated values for look angle (t ) and (2) by altitude After observing theoverlap between near vertical and low oblique classes we are currently restructuringthis variable (lsquotiltrsquo) in the database to provide a measure of t when available Userswill still be able to do searches based on the qualitative measures but these measureswill be more closely tied to actual look angle
341 Obliquity and georeferencing digitised photographsOften the rst step in a remote sensing analysis of a digitized astronaut photo-
graph is to georeference the data and resample it to conform to a known mapprojection Details and recommendations for resampling astronaut photographydata are provided by Robinson et al (2000a 2000c) and a tutorial is also available(McRay et al 2000) Slightly oblique photographs can be geometrically correctedfor remote sensing purposes but extremely oblique photographs are not suited forgeometric correction When obliquity is too great the spatial scale far away fromnadir is much larger than the spatial scale closer to nadir resampling results areunsuitable because pixels near nadir are lost as the image is resampled while manypixels far away from nadir are excessively replicated by resampling
To avoid the generation of excess pixels during georeferencing the pixel sizes ofthe original digitized image should be smaller than the pixels in the nal resampledimage Calculations of original pixel size using methods presented below can beuseful in ensuring meaningful resampling For slightly oblique images formulation 3(sect53) can be used to estimate pixel sizes at various locations in a photograph (nearnadir and away from nadir) and these pixel sizes then used to determine a reasonablepixel scale following resampling
4 Other characteristics that in uence spatial resolutionBeyond basic geometry many other factors internal to the astronaut photography
system impact the observed ground resolved distance in a photograph The lensesand cameras already discussed are imperfect and introduce radiometric degradations(Moik 1980)Vignetting (slight darkening around the edges of an image) occurs in
Astronaut photography as digital data for remote sensing 4413
Tab
le2
Max
imu
mpo
ssib
led
igit
alsp
atia
lre
solu
tio
n(p
ixel
size
inm
)fo
rca
mer
asy
stem
scu
rren
tly
use
dby
ast
ron
auts
top
ho
togr
aph
eart
hb
ase
do
nth
ege
om
etry
of
alt
itud
ele
ns
and
lm
size
C
alc
ula
tion
sas
sum
eth
atco
lour
tran
spar
ency
lm
isdig
itiz
edat
2400
ppi
(pix
els
per
inch
10
6m
mpix
elOtilde
1 )
Min
imu
mgro
un
ddis
tance
repre
sente
dby
1p
ixel
(m)
As
afu
nct
ion
of
alti
tude1
Imag
esi
zeM
inim
um
alt
itud
eM
edia
nalt
itu
de
Maxi
mum
alt
itud
eC
amer
asy
stem
mm
timesm
mp
ixel
s2(2
22k
m)
(326
km
)(6
11
km
)
Lin
ho
f125
mm
90
mm
lens
105
times120
8978times
11
434
261
383
718
250
mm
lens
94
138
259
Has
selb
lad
70
mm
100
mm
lens
55times
55
5198
times519
823
534
564
725
0m
mle
ns
94
138
259
Nik
on
35m
m30
0m
mle
ns
36times
24
3308
times217
478
115
216
400
mm
lens
59
86
162
ESC
35m
m18
0m
mle
ns3
27
6times
18
530
69times
204
311
116
330
630
0m
mle
ns
67
98
184
400
mm
lens
50
74
138
1 Alt
itud
essh
ow
nar
em
inim
um
m
edia
nan
dm
axim
um
thro
ugh
ST
S-8
9(J
anuary
199
8N
=84
mis
sion
s)
2 Act
ual
pix
els
for
ES
C
oth
erw
ise
assu
med
dig
itiz
ati
on
at240
0pp
i(p
ixel
sp
erin
ch)
equ
alto
10
6m
mpix
elOtilde
1 3 T
he
mo
stco
mm
on
con
gura
tion
for
Eart
hph
oto
gra
ph
yas
part
of
the
Eart
hK
AM
pro
ject
J A Robinson et al4414
Figure 4 De nitions of look angle for photographs taken from low orbit around EarthThe example photographs of Hawaii were all taken from approximately the samealtitude (370ndash398 km) and with the same (250 mm) lens on a Hasselblad camera(STS069-701-69 STS069-701-70 and STS069-729-27 [100 mm lens])
astronaut photography due to light path properties of the lenses used A lack of atness of the lm at the moment of exposure (because cameras used in orbit donot incorporate a vacuum platen) also introduces slight degradations A numberof additional sources of image degradation that would aVect GRD of astronautphotographs are listed
41 Films and processing411 Colour reversal lms
Most lms used in NASA handheld cameras in orbit have been E-6 processcolour reversal lms (Ektachromes) that have a nominal speed of 64 to 100 ASAalthough many other lms have been own (table 3) Reversal lms are used becausedamage from radiation exposure while outside the atmospheric protection of Earthis more noticeable in negative lms than in positive lms (Slater 1996) The choiceof ASA has been to balance the coarser grains ( lower lm resolving power) of high-speed lms with the fact that slower lms require longer exposures and thus areaVected more by vehicle motion (approximately 73 km s Otilde 1 relative to Earthrsquos surfacefor the Space Shuttle) Extremely fast lms (gt400 ASA) are also more susceptibleto radiation damage in orbit (particularly during long-duration or high-altitudemissions Slater 1996) and have not traditionally been used for Earth photography The manufacturer stock numbers identifying the lm used are available for eachimage in the Astronaut Photography Database (OYce of Earth Sciences 2000)
412 Colour infrared lmsColour infrared lm (CIR) was used during an Earth-orbiting Apollo mission
(Colwell 1971) in the multispectral S-190A and the high-resolution S-190B camera
Astronaut photography as digital data for remote sensing 4415
Figure 5 (a) Distributions of astronaut photographs by look angle oV-nadir (calculated fromcentre point and nadir point using Formulation 2) Data included for all photographsfor which centre and nadir points are known with high oblique look angles (n=55 155) excluded (Total sample shown is 108 293 of 382 563 total records in thedatabase when compiled For records where nadir data was not available number ofobservations for qualitative look angles are near vertical 14 086 low oblique 115 834undetermined 89 195) (b) Altitude distribution for astronaut photographs of Earthtaken with the Hasselblad camera Data included for Hasselblad camera only focallength determined with high oblique look angles excluded (Total sample shown is159 046 of 206 971 Hasselblad records with known altitude and of 382 563 total recordsin the database when compiled For Hasselblad records with known altitude data notshown includes high oblique photographs using 100 mm and 250 mm lenses n=20 262and all look angles with other lenses n=27 663)
systems on Skylab (NASA 1974 Wilmarth et al 1977) and occasionally on Shuttlemissions and Shuttle-Mir missions (table 3) The CIR lm used is a three-layerAerochrome having one layer that is sensitive to re ected solar infrared energy toapproximately 900 nm (Eastman Kodak Company 1998) protective coatings onmost spacecraft windows also limit IR transmittance between 800 mm and 1000 nmThis layer is extremely sensitive to the temperature of the lm which createsunpredictable degradation of the IR signature under some space ight conditions
J A Robinson et al4416
Tab
le3
Det
ails
on
lm
suse
dfo
rast
ron
aut
pho
togr
aphy
of
Eart
h
Dat
ain
clud
ep
ho
togr
aphy
fro
mall
NA
SA
hum
an
spac
eig
ht
mis
sio
ns
thro
ugh
ST
S-9
6(J
un
e19
99
378
461
tota
lfr
ames
inth
edata
base
)F
ilm
sw
ith
lt50
00fr
am
esex
po
sed
and
lm
suse
dfo
r35
mm
cam
eras
only
are
no
tin
clud
ed
Res
olv
ing
Nu
mb
erof
Film
man
ufa
ctu
rer
AS
Ap
ow
er1
fram
esP
erio
dof
use
Cam
eras
Colo
ur
posi
tive
Kod
akA
eroch
rom
eII
(2447
2448
SO
-131S
O-3
68
)32
40ndash50
513
8196
8ndash19
85
Hass
elbla
dL
inh
of
Kod
akE
kta
chro
me
(5017
502
56
017
6118
SO
-121
64
50
113
932
196
5ndash19
68
Hass
elbla
dL
inh
of
Role
iex
SO
-217
Q
X-8
24
QX
-868
)19
81ndash1993
Mau
rer
Nik
on
Kod
akL
um
iere
(5046
5048
)100
105
743
199
4ndash19
98
Hass
elbla
dL
inho
fK
od
akE
lite
100
S(5
069
)100
41
486
199
6ndashp
rese
nt
Hass
elbla
d2
Fu
jiV
elvia
50C
S135
-36
32
13
882
1992ndash19
97H
ass
elbla
dN
iko
nK
od
akH
i-R
esolu
tio
nC
olo
r(S
O-2
42S
O-3
56
)10
096
74
1973ndash19
75H
ass
elbla
d
S19
0A
S190
B
Infr
are
dand
bla
ck-a
nd-w
hit
e
lms
Kod
akA
eroch
rom
eC
olo
rIn
frar
ed40
32
40
704
196
5ndashp
rese
nt2
Hass
elbla
dL
inh
of
Nik
on
S190
A
(8443
34
432443
)S
190B
Kod
akB
ampW
Infr
are
d(2
424
)32
10
553
197
3ndash19
74
S190
AK
od
akP
anato
mic
-XB
ampW
(SO
-022
)63
10
657
197
3ndash19
74
S190
A
1A
tlo
wco
ntr
ast
(ob
ject
back
grou
nd
rati
o16
1)
aspro
vid
edby
Ko
dak
and
list
edby
Sla
ter
etal
(1
983
256
)R
esolv
ing
pow
erw
as
dis
con
tin
ued
fro
mth
ete
chn
ical
test
sper
form
edb
yK
odak
inapp
roxim
atel
y19
93
an
dit
isn
ot
kn
ow
nif
curr
ent
lm
sh
ave
sim
ilar
reso
lvin
gpo
wer
too
lder
ver
sion
so
fsi
milar
lm
s(K
T
eite
lbaum
K
od
akp
erso
nal
com
mu
nic
atio
n)
Th
egra
nu
lari
tym
easu
reno
wpro
vid
edb
yK
od
ak
can
no
tb
ere
adily
con
ver
ted
toa
spat
ial
reso
luti
on
2 The
Lin
hof
was
use
dag
ain
on
ST
S-1
03in
Dec
emb
er19
99(d
ata
not
cata
logued
asof
the
pre
para
tion
of
this
tab
le)
usi
ng
Ko
dak
Elite
100S
lm
3 B
egin
nin
gin
July
1999
2443
colo
ur-
infr
are
d
lmpro
cess
was
chan
ged
from
EA
-5to
E-6
Astronaut photography as digital data for remote sensing 4417
413 Film duplicationThe original lm is archived permanently after producing about 20 duplicate
printing masters (second generation) Duplicate printing masters are disseminatedto regional archives and used to produce products (thirdndashfourth generation) for themedia the public and for scienti c use When prints slides transparencies anddigital products are produced for public distribution they are often colour adjustedto correspond to a more realistic look Digital products from second generationcopies can be requested by scienti c users (contact the authors for information onobtaining such products for a particular project) and these are recommended forremote sensing applications Care should be taken that the digital product acquiredfor remote sensing analysis has not been colour adjusted for presentation purposesBased on qualitative observations there is little visible increase in GRD in the thirdor fourth generation products There is a signi cant degradation in delity of colourin third and fourth generation duplicates because an increase in contrast occurswith every copy
42 Shutter speedsThe impact of camera motion (both due to the photographer and to the motion
of the spacecraft ) on image quality is determined by shutter speedmdash1250 to 1500second have been used because slower speeds record obvious blurring caused bythe rapid motion of the spacecraft relative to the surface of the Earth Generally1250 was used for ISO 64 lms (and slower) and after the switch to ISO 100 lmsa 1500 setting became standard New Hasselblad cameras that have been ownbeginning with STS-92 in October 2000 vary shutter speed using a focal plane shutterfor exposure bracketing (rather than varying aperture) and 1250 1500 and 11000second are used
Across an altitude range of 120ndash300 nautical miles (222ndash555 km) and orbitalinclinations of 285deg and 570deg the median relative ground velocity of orbitingspacecraft is 73 km s Otilde 1 At a nominal shutter speed of 1500 the expected blur dueto motion relative to the ground would be 146 m When cameras are handheld (asopposed to mounted in a bracket) blur can be reduced by physical compensationfor the motion (tracking) by the photographer many photographs show a level ofdetail indicating that blur at this level did not occur (eg gure 3(d )) Thus motionrelative to the ground is not an absolute barrier to ground resolution for handheldphotographs
43 Spacecraft windowsMost of the photographs in the NASA Astronaut Photography Database were
taken through window ports on the spacecraft used The transmittance of the windowport may be aVected by material inhomogeniety of the glass the number of layersof panes coatings used the quality of the surface polish environmentally inducedchanges in window materials (pressure loads thermal gradients) or deposited con-tamination Such degradation of the window cannot be corrected is diVerent foreach window and changes over time See Eppler et al (1996) and Scott (2000) fordiscussion of the spectral transmittance of the fused quartz window that is part ofthe US Laboratory Module (Destiny) of the International Space Station
44 Digitized images from lmWhen lm is scanned digitally the amount of information retained depends on
the spectral information extracted from the lm at the spatial limits of the scanner
J A Robinson et al4418
To date standard digitizing methodologies for astronaut photographs have not beenestablished and lm is digitized on a case-by-case basis using the equipment available
441 Digitizing and spatial resolutionLight (1993 1996) provided equations for determining the digitizing spatial
resolution needed to preserve spatial resolution of aerial photography based on thestatic resolving power (AWAR) for the system For lms where the manufacturer hasprovided data (Eastman Kodak 1998 K Teitelbaum personal communication) the resolving power of lms used for astronaut photography ranges from 32 lp mm Otilde 1to 100 lp mm Otilde 1 at low contrast (objectbackground ratio 161 table 3) The AWARfor the static case of the Hasselblad camera has been measured at high and lowcontrast (using lenses shown in table 1) with a maximum of approximately55 lp mm Otilde 1 (Fred Pearce unpublished data)
Based on the method of Light (1993 1996) the dimension of one spatial resolutionelement for a photograph with maximum static AWAR of 55 lp mm Otilde 1 would be18 mm lp Otilde 1 The acceptable range of spot size to preserve spatial information wouldthen be
18 mm
2 atilde 2aring scan spot size aring
18 mm
2(1)
and 6 mm aring scan spot size aring 9 mm Similarly for a more typical static AWAR of30 lp mm Otilde 1 (33 mm lpOtilde 1 low contrast Fred Pearce unpublished data) 11 mm aring scanspot sizearing 17 mm These scan spot sizes correspond to digitizing resolutions rangingfrom 4233ndash2822 ppi (pixels inch Otilde 1 ) for AWAR of 55 lp mm Otilde 1 and 2309ndash1494 ppi forAWAR of 30 lp mm Otilde 1 Scan spot sizes calculated for astronaut photography arecomparable to those calculated for the National Aerial Photography Program(9ndash13 mm Light 1996)
Widely available scanners that can digitize colour transparency lm currentlyhave a maximum spatial resolution of approximately 2400 ppi (106 mm pixel Otilde 1) Forexample we routinely use an Agfa Arcus II desktop scanner (see also Baltsavias1996) with 2400 ppi digitizing spatial resolution (2400 ppi optical resolution in onedirection and 1200 ppi interpolated to 2400 ppi in the other direction) and 2400 ppiwas used to calculate IFOV equivalents (tables 2 and 4) For some combinations oflens camera and contrast 2400 ppi will capture nearly all of the spatial informationcontained in the lm However for lm with higher resolving power thanEktachrome-64 for better lenses and for higher contrast targets digitizing at 2400 ppiwill not capture all of the spatial information in the lm
Improvements in digitizing technology will only produce real increases in IFOVto the limit of the AWAR of the photography system The incremental increase inspatial information above 3000 ppi (85 mm pixel Otilde 1 ) is not likely to outweigh thecosts of storing large images (Luman et al 1997) At 2400 ppi a Hasselblad frameis approximately 5200times5200 or 27 million pixels (table 2) while the same imagedigitized at 4000 ppi would contain 75 million pixels
Initial studies using astronaut photographs digitized at 2400 ppi (106 mm pixel Otilde 1 Webb et al 2000 Robinson et al 2000c) indicate that some GRD is lost comparedto photographic products Nevertheless the spatial resolution is still comparablewith other widely used data sources (Webb et al 2000) Robinson et al (2000c)found that digitizing at 21 mm pixel Otilde 1 provided information equivalent to 106 mmpixel Otilde 1 for identifying general urban area boundaries for six cities except for a single
Astronaut photography as digital data for remote sensing 4419
photo that required higher spatial resolution digitizing (it had been taken with ashorter lens and thus had less spatial resolution) Part of the equivalence observedin the urban areas study may be attributable to the fact that the atbed scannerused interpolates from 1200 ppi to 2400 ppi in one direction The appropriate digitiz-ing spatial resolution will in part depend on the scale of features of interest to theresearchers with a maximum upper limit set of a scan spot size of approximately6 mm (4233 ppi)
442 Digitizing and spectral resolutionWhen colour lm is digitized there will be loss of spectral resolution The three
lm emulsion layers (red green blue) have relatively distinct spectral responses butare fused together so that digitizers detect one colour signal and must convert itback into three (red green blue) digital channels Digital scanning is also subject tospectral calibration and reproduction errors Studies using digital astronaut photo-graphs to date have used 8 bits per channel However this is another parameterthat can be controlled when the lm is digitized We do not further address spectralresolution of digitally scanned images in this paper
45 External factors that in uence GRDAlthough not discussed in detail in this paper factors external to the spacecraft
and camera system (as listed by Moik (1980) for remote sensing in general) alsoimpact GRD These include atmospheric interference (due to scattering attenuationhaze) variable surface illumination (diVerences in terrain slope and orientation) andchange of terrain re ectance with viewing angle (bidirectional re ectance) For astro-naut photographs variable illumination is particularly important because orbits arenot sun-synchronous Photographs are illuminated by diVerent sun angles and imagesof a given location will have colour intensities that vary widely In addition theviewing angle has an eVect on the degree of object-to-backgroun d contrast andatmospheric interference
5 Estimating spatial resolution of astronaut photographsIn order to use astronaut photographs for digital remote sensing it is important
to be able to calculate the equivalent to an IFOVmdashthe ground area represented bya single pixel in a digitized orbital photograph The obliquity of most photographsmeans that pixel lsquosizesrsquo vary at diVerent places in an image Given a ground distanceD represented by a photograph in each direction (horizontal vertical ) an approxi-mate average pixel width (P the equivalent of IFOV) for the entire image can becalculated as follows
P=D
dS(2)
where D is the projected distance on the ground covered by the image in the samedirection as the pixel is measured d is the actual width of the image on the original lm (table 1) and S is the digitizing spatial resolution
Here we present three mathematical formulations for estimating the size of thefootprint or area on the ground covered by the image Example results from theapplication of all three formulations are given in table 5 The rst and simplestcalculation (formulation 1) gives an idea of the maximum spatial resolution attainableat a given altitude of orbit with a given lm format and a perfectly vertical (nadir)
J A Robinson et al4420
view downward Formulation 2 takes into account obliquity by calculating lookangle from the diVerence between the location on the ground represented at thecentre of the photograph and the nadir location of the spacecraft at the time thephotograph was taken ( gure 1) Formulation 3 describes an alternate solution tothe oblique look angle problem using coordinate-system transformations Thisformulation has been implemented in a documented spreadsheet and is availablefor download (httpeoljscnasagovsseopFootprintCalculatorxls) and is in theprocess of being implemented on a Web-based user interface to the AstronautPhotography Database (OYce of Earth Sciences 2000)
Although formulations 2 and 3 account for obliquity for the purposes of calcula-tion they treat the position of the spacecraft and position of the camera as one Inactuality astronauts are generally holding the cameras by hand (although camerasbracketed in the window are also used) and the selection of window position of theastronaut in the window and rotation of the camera relative to the movement ofthe spacecraft are not known Thus calculations using only the photo centre point(PC) and spacecraft nadir point (SN) give a locator ellipse and not the locations ofthe corners of the photograph A locator ellipse describes an estimated area on theground that is likely to be included in a speci c photograph regardless of the rotationof the lm plane about the camerarsquos optical axis ( gure 6)
Estimating the corner positions of the photo requires additional user input of asingle auxiliary pointmdasha location on the image that has a known location on theground Addition of this auxiliary point is an option available to users of thespreadsheet An example of the results of adding an auxiliary point is shown in gure 7 with comparisons of the various calculations in table 5
51 Formulation 1 Footprint for a nadir viewThe simplest way to estimate footprint size is to use the geometry of camera lens
and spacecraft altitude to calculate the scaling relationship between the image in the
Figure 6 Sketch representation of the locator ellipse an estimated area on the ground thatis likely to be included in a speci c photograph regardless of the rotation of the lmplane about the camerarsquos optical axis
Astronaut photography as digital data for remote sensing 4421
Figure 7 An example of the application of the Low Oblique Space Photo FootprintCalculator The photograph (STS093-704-50) of Limmen Bight Australia was takenfrom an altitude of 283 km using the Hasselblad camera with a 250 mm lens Mapused is 11 000 000 Operational Navigational Chart The distances shown equate toan average pixel size of 124 m for this photograph
lm and the area covered on the ground For a perfect nadir view the scalerelationship from the geometry of similar triangles is
d
D=
f
H(3)
where d=original image size D=distance of footprint on the ground f =focallength of lens and H=altitude Once D is known pixel size ( length=width) can becalculated from equation (2) These calculations represent the minimum footprintand minimum pixel size possible for a given camera system altitude and digitisingspatial resolution (table 2) Formulation 1 was used for calculating minimum pixelsizes shown in tables 2 and 4
By assuming digitizing at 2400 ppi (106 mm pixel Otilde 1 ) currently a spatial resolutioncommonly attainable from multipurpose colour transparency scanners (see sect441)this formulation was used to convert area covered to an IFOV equivalent for missionsof diVerent altitudes (table 2) Table 4 provides a comparison of IFOV of imagesfrom various satellites including the equivalent for astronaut photography Valuesin this table were derived by using formulation 1 to estimate the area covered becausea perfect nadir view represents the best possible spatial resolution and smallest eldof view that could be obtained
52 Formulation 2 Footprint for oblique views using simpli ed geometry and thegreat circle distance
A more realistic approach to determining the footprint of the photographaccounts for the fact that the camera is not usually pointing perfectly down at the
J A Robinson et al4422
Table 4 Comparison of pixel sizes (instantaneous elds of view) for satellite remote sensorswith pixel sizes for near vertical viewing photography from spacecraft Note that alldata returned from the automated satellites have the same eld of view while indicatedpixel sizes for astronaut photography are best-case for near vertical look angles See gure 5 for distributions of astronaut photographs of various look angles High-altitude aerial photography is also included for reference
Altitude PixelInstrument (km) width (mtimesm) Bands1
Landsat Multipectral Scanner2 (MSS 1974-) 880ndash940 79times56 4ndash5Landsat Thematic Mapper (TM 1982-) ~705 30times30 7Satellite pour lrsquoObservation de la Terre (SPOT 1986-) ~832
SPOT 1-3 Panchromatic 10times10 1SPOT 1-3 Multispectral (XS) 20times20 3
Astronaut Photography (1961-)Skylab3 (1973-1974 ) ~435
Multispectral Camera (6 camera stations) 17ndash19times17ndash19 3ndash8Earth Terrain Camera (hi-res colour) 875times875 3
Space Shuttle5 (1981-) 222ndash611Hasselblad 70 mm (250mm lens colour) vertical view 9ndash26times9ndash26 3Hasselblad 70 mm (100mm lens colour) vertical view 23ndash65times23ndash65 3Nikon 35 mm (400mm lens colour lm or ESC) vertical 5ndash16times5ndash16 3
High Altitude Aerial Photograph (1100 000) ~12 2times2 3
1Three bands listed for aerial photography obtainable by high-resolution colour digitizingFor high-resolution colour lm at 65 lp mmOtilde 1 (K Teitelbaum Eastman Kodak Co personalcommunication) the maximum information contained would be 3302 ppi
2Data from Simonett (1983Table 1-2)3Calculated as recommended by Jensen (19831596) the dividend of estimated ground
resolution at low contrast given in NASA (1974179-180) and 244Six lters plus colour and colour infrared lm (NASA 19747) 5Extracted from table 2
nadir point The look angle (the angle oV nadir that the camera is pointing) can becalculated trigonometrically by assuming a spherical earth and calculating the dis-tance between the coordinates of SN and PC ( gure 1) using the Great Circledistance haversine solution (Sinnott 1984 Snyder 198730ndash32 Chamberlain 1996)The diVerence between the spacecraft centre and nadir point latitudes Dlat=lat 2 shy lat 1 and the diVerence between the spacecraft centre and nadir pointlongitudes Dlon=lon 2 shy lon 1 enter the following equations
a=sin2ADlat
2 B+cos(lat 1) cos(lat 2) sin2ADlon
2 B (4)
c=2 arcsin(min[1 atilde a]) (5)
and oVset=R c with R the radius of a spherical Earth=6370 kmThe look angle (t ) is then given by
t=arctanAoffset
H B (6)
Assuming that the camera was positioned so that the imaginary line between thecentre and nadir points (the principal line) runs vertically through the centre of thephotograph the distance between the geometric centre of the photograph (principalpoint PP) and the top of the photograph is d2 ( gure 8(a)) The scale at any point
Astronaut photography as digital data for remote sensing 4423
(a) (b)
Figure 8 The geometric relationship between look angle lens focal length and the centrepoint and nadir points of a photograph (see also Estes and Simonett 1975) Note thatthe Earthrsquos surface and footprint represented in gure 1 are not shown here only arepresentation of the image area of the photograph Variables shown in the gurelook angle t focal length f PP centre of the image on the lm SN spacecraft nadird original image size (a) The special case where the camera is aligned squarely withthe isometric parallel In this case tilt occurs along a plane parallel with the centre ofthe photograph When the conditions are met calculations using Formulation 3 willgive the ground coordinates of the corner points of the photograph (b) The generalrelationship between these parameters and key photogrammetric points on the photo-graph For this more typical case coordinates of an additional point in the photographmust be input in order to adjust for twist of the camera and to nd the corner pointcoordinates
in the photograph varies as a function of the distance y along the principal linebetween the isocentre and the point according to the relationship
d
D=
f shy ysint
H(7)
(Wong 1980 equation (214) H amp the ground elevation h) Using gure 8(a) at thetop of the photo
y= f tanAt
2B+d
2(8)
and at the bottom of the photo
y= f tanAt
2Bshyd
2(9)
Thus for given PC and SN coordinates and assuming a photo orientation as in gure 8(a) and not gure 8(b) we can estimate a minimum D (using equations (7)and (8)) and a maximum D (using equations (7) and (9)) and then average the twoto determine the pixel size (P ) via equation (2)
J A Robinson et al4424
53 Formulation 3 T he L ow Oblique Space Photo Footprint CalculatorThe Low Oblique Space Photo Footprint Calculator was developed to provide
a more accurate estimation of the geographic coordinates for the footprint of a lowoblique photo of the Earthrsquos surface taken from a human-occupied spacecraft inorbit The calculator performs a series of three-dimensional coordinate transforma-tions to compute the location and orientation of the centre of the photo exposureplane relative to an earth referenced coordinate system The nominal camera focallength is then used to create a vector from the photorsquos perspective point througheach of eight points around the parameter of the image as de ned by the formatsize The geographical coordinates for the photo footprint are then computed byintersecting these photo vectors with a spherical earth model Although more sophist-icated projection algorithms are available no signi cant increase in the accuracy ofthe results would be produced by these algorithms due to inherent uncertainties inthe available input data (ie the spacecraft altitude photo centre location etc)
The calculations were initially implemented within a Microsoft Excel workbookwhich allowed one to embed the mathematical and graphical documentation nextto the actual calculations Thus interested users can inspect the mathematical pro-cessing A set of error traps was also built into the calculations to detect erroneousresults A summary of any errors generated is reported to the user with the resultsof the calculations Interested individuals are invited to download the Excel work-book from httpeoljscnasagovsseopFootprintCalculatorxls The calculations arecurrently being encoded in a high-level programming language and should soon beavailable alongside other background data provided for each photograph at OYceof Earth Sciences (2000)
For the purposes of these calculations a low oblique photograph was de ned as
one with the centre within 10 degrees of latitude and longitude of the spacecraftnadir point For the typical range of spacecraft altitudes to date this restricted thecalculations to photographs in which Earthrsquos horizon does not appear (the generalde nition of a low oblique photograph eg Campbell 199671 and gure 4)
531 Input data and resultsUpon opening the Excel workbook the user is presented with a program intro-
duction providing instructions for using the calculator The second worksheet tablsquoHow-To-Usersquo provides detailed step-by-step instructions for preparing the baselinedata for the program The third tab lsquoInput-Output rsquo contains the user input eldsand displays the results of the calculations The additional worksheets contain theactual calculations and program documentation Although users are welcome toreview these sheets an experienced user need only access the lsquoInput-Outputrsquo
spreadsheetTo begin a calculation the user enters the following information which is available
for each photo in the NASA Astronaut Photography Database (1) SN geographicalcoordinates of spacecraft nadir position at the time of photo (2) H spacecraftaltitude (3) PC the geographical coordinates of the centre of the photo (4) f nominal focal length and (5) d image format size The automatic implementationof the workbook on the web will automatically enter these values and completecalculations
For more accurate results the user may optionally enter the geographic co-ordinates and orientation for an auxiliary point on the photo which resolves thecamerarsquos rotation uncertainty about the optical axis The auxiliary point data must
Astronaut photography as digital data for remote sensing 4425
be computed by the user following the instruction contained in the lsquoHow-To-Usersquotab of the spreadsheet
After entering the input data the geographic coordinates of the photo footprint(ie four photo corner points four points at the bisector of each edge and the centreof the photo) are immediately displayed below the input elds along with any errormessages generated by the user input or by the calculations ( gure 7) Although
results are computed and displayed they should not be used when error messagesare produced by the program The program also computes the tilt angle for each ofthe photo vectors relative to the spacecraft nadir vector To the right of the photofootprint coordinates is displayed the arc distance along the surface of the sphere
between adjacent computed points
532 Calculation assumptions
The mathematical calculations implemented in the Low Oblique Space PhotoFootprint Calculator use the following assumptions
1 The SN location is used as exact even though the true value may vary by upto plusmn01 degree from the location provided with the photo
2 The spacecraft altitude is used as exact Although the determination of the
nadir point at the instant of a known spacecraft vector is relatively precise(plusmn115times10 Otilde 4 degrees) the propagator interpolates between sets of approxi-mately 10ndash40 known vectors per day and the time code recorded on the lmcan drift Thus the true value for SN may vary by up to plusmn01 degree fromthe value provided with the photo
3 The perspective centre of the camera is assumed to be at the given altitudeover the speci ed spacecraft nadir location at the time of photo exposure
4 The PC location is used as exact even though the true value may vary by upto plusmn05deg latitude and plusmn05deg longitude from the location provided with the
photo5 A spherical earth model is used with a nominal radius of 6 372 16154 m (a
common rst-order approximation for a spherical earth used in geodeticcomputations)
6 The nominal lens focal length of the camera lens is used in the computations(calibrated focal length values are not available)
7 The photo projection is based on the classic pin-hole camera model8 No correction for lens distortion or atmospheric re ection is made
9 If no auxiliary point data is provided the lsquoTop of the Imagersquo is orientedperpendicular to the vector from SN towards PC
533 T ransformation from Earth to photo coordinate systemsThe calculations begin by converting the geographic coordinates ( latitude and
longitude) of the SN and PC to a Rectangular Earth-Centred Coordinate System(R-Earth) de ned as shown in gure 9 (with the centre of the Earth at (0 0 0))
Using the vector from the Earthrsquos centre through SN and the spacecraft altitude thespacecraft location (SC) is also computed in R-Earth
For ease of computation a Rectangular Spacecraft-Centred Coordinate System(R-Spacecraft ) is de ned as shown in gure 9 The origin of R-Spacecraft is located
at SC with its +Z-axis aligned with vector from the centre of the Earth throughSN and its +X-axis aligned with the vector from SN to PC ( gure 9) The speci c
J A Robinson et al4426
Figure 9 Representation of the area included in a photograph as a geometric projectiononto the Earthrsquos surface Note that the focal length (the distance from the SpaceShuttle to the photo) is grossly overscale compared to the altitude (the distance fromthe Shuttle to the surface of the Earth
rotations and translation used to convert from the R-Earth to the R-Spacecraft arecomputed and documented in the spreadsheet
With the mathematical positions of the SC SN PC and the centre of the Earthcomputed in R-Spacecraft the program next computes the location of the camerarsquosprincipal point (PP) The principle point is the point of intersection of the opticalaxis of the lens with the image plane (ie the lm) It is nominally positioned at a
Astronaut photography as digital data for remote sensing 4427
distance equal to the focal length from the perspective centre of the camera (whichis assumed to be at SC) along the vector from PC through SC as shown in gure 9
A third coordinate system the Rectangular Photo Coordinate System (R-Photo)is created with its origin at PP its XndashY axial plane normal to the vector from PCthrough SC and its +X axis aligned with the +X axis of R-Spacecraft as shownin gure 9 The XndashY plane of this coordinate system represents the image plane ofthe photograph
534 Auxiliary point calculationsThe calculations above employ a critical assumption that all photos are taken
with the lsquotop of the imagersquo oriented perpendicular to the vector from the SN towardsPC as shown in gure 9 To avoid non-uniform solar heating of the external surfacemost orbiting spacecraft are slowly and continually rotated about one or more axesIn this condition a ight crew member taking a photo while oating in microgravitycould orient the photo with practically any orientation relative to the horizon (seealso gure 8(b)) Unfortunately since these photos are taken with conventionalhandheld cameras there is no other information available which can be used toresolve the photorsquos rotational ambiguity about the optical axis other then the photoitself This is why the above assumption is used and the footprint computed by thiscalculator is actually a lsquolocator ellipsersquo which estimates the area on the ground whatis likely to be included in a speci c photograph (see gure 6) This locator ellipse ismost accurate for square image formats and subject to additional distortion as thephotograph format becomes more rectangular
If the user wants a more precise calculation of the image footprint the photorsquosrotational ambiguity about the optical axis must be resolved This can be done inthe calculator by adding data for an auxiliary point Detailed instructions regardinghow to prepare and use auxiliary point data in the computations are included in thelsquoHow-To-Usersquo tab of the spreadsheet Basically the user determines which side ofthe photograph is top and then measures the angle between the line from PP to thetop of the photo and from PP to the auxiliary point on the photo ( gure 9)
If the user includes data for an auxiliary point a series of computations arecompleted to resolve the photo rotation ambiguity about the optical axis (ie the
+Z axis in R-Photo) A vector from the Auxiliary Point on the Earth (AE) throughthe photograph perspective centre ( located at SC) is intersected with the photo imageplane (XndashY plane of R-Photo) to compute the coordinates of the Auxiliary Point onthe photo (AP) in R-Photo A two-dimensional angle in the XndashY plane of R-Photofrom the shy X axis to a line from PP to AP is calculated as shown in gure 9 The
shy X axis is used as the origin of the angle since it represents the top of the photoonce it passes through the perspective centre The diVerence between the computedangle and the angle measured by the user on the photo resolves the ambiguity inthe rotation of the photo relative to the principal line ( gures 7 and 9) Thetransformations from R-Spacecraft and R-Photo are then modi ed to include anadditional rotation angle about the +Z axis in R-Photo
535 lsquoFootprint rsquo calculationsThe program next computes the coordinates of eight points about the perimeter
of the image format (ie located at the four photo corners plus a bisector pointalong each edge of the image) These points are identi ed in R-Photo based uponthe photograph format size and then converted to R-Spacecraft Since all computa-tions are done in orthogonal coordinate systems the R-Spacecraft to R-Photo
J A Robinson et al4428
rotation matrix is transposed to produce an R-Photo to R-Spacecraft rotation matrixOnce in R-Spacecraft a unit vector from each of the eight perimeter points throughthe photo perspective centre (the same point as SC) is computed This provides thecoordinates for points about the perimeter of the image format with their directionvectors in a common coordinate system with other key points needed to computethe photo footprint
The next step is to compute the point of intersection between the spherical earthmodel and each of the eight perimeter point vectors The scalar value for eachperimeter point unit vector is computed using two-dimensional planar trigonometryAn angle c is computed using the formula for the cosine of an angle between twoincluded vectors (the perimeter point unit vector and the vector from SC to thecentre of the Earth) Angle y is computed using the Law of Sines Angle e=180degrees shy c shy y The scalar value of the perimeter point vector is computed using eand the Law of Cosines The scalar value is then multiplied by the perimeter pointunit vector to produce the three-dimensional point of intersection of the vector withEarthrsquos surface in R-Spacecraft The process is repeated independently for each ofthe eight perimeter point vectors Aside from its mathematical simplicity the valueof arcsine (computed in step 2) will exceed the normal range for the sin of an anglewhen the perimeter point vectors fail to intersect with the surface of the earth Asimple test based on this principle allows the program to correctly handle obliquephotos which image a portion of the horizon (see results for high oblique photographsin table 5)
The nal step in the process converts the eight earth intersection points from theR-Spacecraft to R-Earth The results are then converted to the standard geographiccoordinate system and displayed on the lsquoInput-Outputrsquo page of the spreadsheet
54 ExamplesAll three formulations were applied to the photographs included in this paper
and the results are compared in table 5 Formulation 1 gives too small a value fordistance across the photograph (D) for all but the most nadir shots and thus servesas an indicator of the best theoretical case but is not a good measure for a speci cphotograph For example the photograph of Lake Eyre taken from 276 km altitude( gure 2(a)) and the photograph of Limmen Bight ( gure 7) were closest to beingnadir views (oVsetslt68 km or tlt15deg table 5) For these photographs D calculatedusing Formulation 1 was similar to the minimum D calculated using Formulations2 and 3 For almost all other more oblique photographs Formulation 1 gave asigni cant underestimate of the distance covered in the photograph For gure 5(A)(the picture of Houston taken with a 40 mm lens) Formulation 1 did not give anunderestimate for D This is because Formulation 1 does not account for curvatureof the Earth in any way With this large eld of view assuming a at Earth in atedthe value of D above the minimum from calculations that included Earth curvature
A major diVerence between Formulations 2 and 3 is the ability to estimate pixelsizes (P) in both directions (along the principal line and perpendicular to the principalline) For the more oblique photographs the vertical estimate of D and pixel sizes ismuch larger than in the horizontal direction (eg the low oblique and high obliquephotographs of Hawaii gure 4 table 5)
For the area of Limmen Bight ( gure 7) table 5 illustrates the improvement inthe estimate of distance and pixel size that can be obtained by re-estimating thelocation of the PC with greater accuracy Centre points in the catalogued data are
Astronaut photography as digital data for remote sensing 4429
plusmn05deg of latitude and longitude When the centre point was re-estimated to plusmn002degwe determined that the photograph was not taken as obliquely as rst thought(change in the estimate of the look angle t from 169 to 128deg table 5) When theauxiliary point was added to the calculations of Formula 3 the calculated look angleshrank further to 110deg indicating that this photograph was taken at very close toa nadir view Of course this improvement in accuracy could have also led to estimatesof greater obliquity and corresponding larger pixel sizes
We also tested the performance of the scale calculator with auxiliary point byestimating the corner and point locations on the photograph using a 11 000 000Operational Navigational Chart For this test we estimated our ability to readcoordinates from the map as plusmn002degand our error in nding the locations of thecorner points as plusmn015deg (this error varies among photographs depending on thedetail that can be matched between photo and map) For Limmen Bight ( gure 7)the mean diVerence between map estimates and calculator estimates for four pointswas 031deg (SD=018 n=8) For a photograph of San Francisco Bay (STS062-151-291) the mean diVerence between map estimates and calculator estimates forfour points was 0064deg (SD=018 n=8) For a photograph of San Francisco Bay(STS062-151-291) the mean diVerence between map estimates and calculator estim-ates for eight points was 0196deg (SD=0146 n=16) Thus in one case the calculatorestimates were better than our estimate of the error in locating corner points on themap It is reasonable to expect that for nadir-viewing photographs the calculatorused with an auxiliary point can estimate locations of the edges of a photograph towithin plusmn03deg
55 Empirical con rmation of spatial resolution estimatesAs stated previously a challenge to estimating system-AWAR for astronaut
photography of Earth is the lack of suitable targets Small features in an image cansometimes be used as a check on the size of objects that can be successfully resolvedgiving an approximate value for GRD Similarly the number of pixels that make upthose features in the digitized image can be used to make an independent calculationof pixel size We have successfully used features such as roads and airport runwaysto make estimates of spatial scale and resolution (eg Robinson et al 2000c) Whilerecognizing that the use of linear detail in an image is a poor approximation to abar target and that linear objects smaller than the resolving power can often bedetected (Charman 1965) few objects other than roads could be found to make anydirect estimates of GRD Thus roads and runways were used in the images ofHouston (where one can readily conduct ground veri cations and where a numberof higher-contrast concrete roadways were available) to obtain empirical estimatesof GRD and pixel size for comparison with table 5
In the all-digital ESC image of Houston ( gure 3(d )) we examined EllingtonField runway 4-22 (centre left of the image) which is 24384 mtimes457 m This runwayis approximately 6ndash7 pixels in width and 304ndash3094 pixels in length so pixelsrepresent an distance on the ground 7ndash8 m Using a lower contrast measure of astreet length between two intersections (2125 m=211 pixels) a pixel width of 101 mis estimated These results compare favourably with the minimum estimate of 81 mpixels using Formulation 1 (table 5) For an estimate of GRD that would be morecomparable to aerial photography of a line target the smallest street without treecover that could be clearly distinguished on the photograph was 792 m wide Thesmallest non-street object (a gap between stages of the Saturn rocket on display in
J A Robinson et al4430
Tab
le5
App
lica
tio
nof
met
hod
sfo
res
tim
ati
ng
spati
alre
solu
tio
no
fast
ron
aut
ph
oto
gra
ph
sin
clu
din
gF
orm
ula
tion
s12
and
3Im
pro
ved
cen
tre
po
int
esti
mat
esan
dauxilia
ryp
oin
tca
lcu
lati
ons
are
sho
wn
for
gure
7V
aria
ble
sas
use
din
the
text
fle
ns
foca
lle
ngt
hH
al
titu
de
PC
ph
oto
gra
ph
cen
tre
poin
tSN
sp
ace
craft
nadir
po
int
D
dis
tance
cover
edo
nth
egr
oun
d
Pd
ista
nce
on
the
grou
nd
repre
sen
ted
by
apix
el
OVse
td
ista
nce
bet
wee
nP
Cand
SNusi
ng
the
grea
tci
rcle
form
ula
t
look
angle
away
from
SN
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
2L
ake
Eyre
ST
S093
-717
-80
250
276
shy29
0137
0shy
28
413
71
607
11
767
413
7609
ndash64
212
013
760
9ndash64
7121
124
ST
S082
-754
-61
250
617
shy28
5137
0shy
28
113
50
135
726
1201
18
013
8ndash14
827
517
9138ndash15
3277
294
Fig
ure
3H
ou
sto
nS
TS
062
-97-
151
4029
829
5shy
95
530
5shy
95
4410
78
8112
20
5348ndash58
990
1205
351
ndash63
8951
10
36
ST
S094
-737
-28
100
294
295
shy
95
528
3shy
95
5162
31
1133
24
415
8ndash20
334
724
3158ndash20
7351
390
ST
S055
-71-
41250
302
295
shy
95
028
2shy
93
766
412
8192
32
5736
ndash84
715
232
273
9ndash85
7154
185
S73
E51
45300
2685
295
shy
95
024
78
0F
igu
re4
Haw
aii
ST
S069
-701
-65
250
370
195
shy
155
520
4shy
153
881
415
7204
28
9876
ndash98
917
928
688
0ndash10
9181
210
ST
S069
-701
-70
250
370
210
shy
157
519
2shy
150
781
415
7738
63
414
9ndash23
336
760
7148ndash55
6394
10
63
ST
S069
-729
-27
100
398
200
shy
156
620
5shy
148
2219
42
1867
65
432
8ndash13
10
158
62
4323+
311
N
A
Astronaut photography as digital data for remote sensing 4431
Tab
le5
(Cont
inued
)
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
7L
imm
enB
igh
tS
TS
093
-704
-50
250
283
shy15
0135
0shy
15
013
58
623
120
859
16
963
0ndash67
3125
16
863
0ndash67
612
613
2P
CR
e-es
tim
ated
shy14
733
1352
67
64
5128
623
ndash65
5123
128
623
ndash65
712
3127
Auxilia
ryP
oin
t611
062
3ndash65
112
312
5F
igu
re10
N
ear
Au
stin
T
XS
TS
095
-716
-46
250
543
30
5shy
975
28
6shy
1007
120
230
375
34
613
5ndash157
281
34
1136ndash16
128
636
0
1 All
pho
togr
aphs
exce
pt
on
eta
ken
wit
hH
ass
elbla
dca
mer
aso
dis
tan
ceacr
oss
the
image
d=
55m
m
for
S73
E51
45
taken
wit
hel
ectr
onic
still
cam
erad=
276
5m
mho
rizo
nta
l2 P
Cand
SN
are
giv
enin
the
form
(lati
tud
elo
ngit
ude)
deg
rees
w
ith
neg
ativ
en
um
ber
sre
pre
senti
ng
south
and
wes
tA
ccura
cyo
fP
Cis
plusmn0
5and
SN
isplusmn
01
as
reco
rded
inth
eA
stro
naut
Ph
oto
gra
phy
Data
base
ex
cep
tw
her
en
ote
dth
atce
ntr
ep
oin
tw
asre
-est
imate
dw
ith
grea
ter
accu
racy
3 C
alc
ula
ted
usi
ng
the
mea
nofth
em
inim
um
an
dm
axim
um
esti
mate
sof
DF
orm
ula
tion
2co
nsi
der
sD
inth
ed
irec
tion
per
pen
dic
ula
rto
the
Pri
nci
pal
Lin
eo
nly
4 C
alc
ula
ted
inho
rizo
nta
lan
dve
rtic
aldir
ecti
on
sb
ut
still
ass
um
ing
geom
etry
of
gure
6(B
)F
irst
nu
mb
eris
the
mea
no
fth
em
inim
um
Dat
the
bott
om
of
the
ph
oto
and
the
maxi
mum
Dat
the
top
of
the
ph
oto
(range
show
nin
the
tab
le)
and
isco
mpara
ble
toF
orm
ula
tio
n2
Sec
ond
num
ber
isth
em
ean
of
Des
tim
ated
alon
gth
eP
rin
cip
alline
and
alo
ng
the
ou
tsid
eed
ge
(range
not
show
n)
5 Alt
itu
de
isap
pro
xim
ate
for
mis
sion
as
exac
tti
me
pho
togr
ap
hw
asta
ken
and
corr
esp
on
din
gn
adir
data
are
no
tava
ilab
le
Lack
of
nad
irdata
pre
ven
tsap
plica
tion
of
Form
ula
tion
s2
an
d3
6 Au
xilliar
ypo
int
add
edw
as
shy14
80
lati
tud
e13
51
2lo
ngit
ude
shy46
5degro
tati
on
angle
in
stru
ctio
ns
for
esti
mati
ng
the
rota
tio
nan
gle
para
met
erare
conta
ined
as
par
tof
the
scal
eca
lcu
lato
rsp
read
shee
tdo
cum
enta
tio
n
J A Robinson et al4432
a park at Johnson Space Center) that could clearly be distinguished on the photo-graph was 853 m wide
For the photograph of Houston taken with a 250 mm lens ( gure 3(C)) anddigitized from second generation lm at 2400 ppi (106 mm pixel Otilde 1 ) Ellington Fieldrunway 4-22 is 3 pixels in width and 1613 pixels in length so pixels represent151ndash152 m on the ground These results compare favourably with the minimumestimate of 152 m using Formulation 2 and 154ndash185 m pixels using Formulation 3(table 5) For an estimate of GRD using an 8times8 inch print (1369 enlargement) and4times magni cation the smallest street that could clearly be distinguished was 822 mwide the same feature could barely be distinguished on the digitized image
We also made an empirical estimate of spatial resolution for lower contrastvegetation boundaries By clearing forest so that a pattern would be visible to landingaircraft a landowner outside Austin Texas (see also aerial photo in Lisheron 2000)created a target that is also useful for evaluating spatial resolution of astronautphotographs The forest was selectively cleared in order to spell the landownerrsquosname lsquoLUECKErsquo with the remaining trees ( gure 10) According to local surveyorswho planned the clearing the plan was to create letters that were 3100 fttimes1700 ft(9449 mtimes5182 m) Photographed at a high altitude relative to most Shuttle missions(543 km) with a 250 mm lens Formula 3 predicts that each pixel would represent anarea 286 mtimes360 m on the ground (table 5) When original lm was digitized at2400 ppi (106 mm pixel Otilde 1 ) letters correspond to 294times188 pixels for a comparablepixel size of 27ndash32 m
6 Summary and conclusionsAstronaut photographs can be an excellent source of data for remote sensing
applications Best-case resolutions are similar to that for Landsat or SPOT withpixels as small as lt10 m It was estimated that of images taken to date 504 hadlens and obliquity characteristics that make them potential candidates for remotesensing information Digitized astronaut photographs can be overlaid with othersatellite data using GIS (Eckardt et al 2000) or used to ll in gaps in time serieswhen other imagery is not available As a source of public-domain information theycan be very useful for scientists who do not have access to satellite imagery eitherbecause of the expense of image acquisition (to the end user) or the computersystems needed for processing satellite images These diVerences in image acquisitioncosts (to the user) are summarized in table 6
Searching of the complete database of NASA astronaut photography includinglow-spatial-resolution browse images is available via the Web (OYce of EarthSciences 2000) This provides nearly global access for identifying images that willcontribute to a speci c scienti c project For the most detailed studies digitalproducts posted to the web will not be of suYcient quality To date we have madeit a practice of digitizing small numbers of images when requested by scientists atno charge Such requests (including a description of the project involved) can bemade through the authors or using the contact information listed on the websiteInvestigators with needs for larger numbers of digital images have also been servedthrough collaborative agreements
In our experience scientists that nd the data most valuable are those in develop-ing countries those studying areas that have not usually been targets for the majorsatellites those having diYculty in nding low-cloud images those interested inconstructing time series or those interested in using a large number of images
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
J A Robinson et al4404
1971) the Earth-orbiting Apollo missions (Colwell 1971) the Apollo-Soyuz mission(El-Baz 1977 El-Baz and Warner 1979) Skylab (NASA 1974 Wilmarth et al 1977 )a few Shuttle missions (eg the two Space Radar Laboratory missions of 1994 Joneset al 1996) and the Shuttle-Mir missions (Evans et al 2000) Extensive training inphotography is available to members of all ight crews during their general trainingperiod and during intensive training for speci c missions (Jones et al 1996) Mostphotographs have been taken by astronauts on a time-available basis Astronautphotographs are thus a subset of the potential scenes selected both by opportunity(orbital parameters lighting and crew workloads and schedules) and by the trainingexperience and interest of the photographers
As remote sensing and geographic information systems have become more widelyavailable tools we have collaborated with a number of scientists interested in usingastronaut photography for quantitative remote sensing applications Digitized imagesfrom lm are suitable for geometric recti cation and image enhancement followed byclassi cation and other remote sensing techniques (Lulla and Helfert 1989 Mohler et al1989 Helfert et al 1990 Lulla et al 1991 Eckardt et al 2000 Robinson et al 2000a2000c and in press Webb et al in press) The photographs are in the public domainthus they provide a low-cost alternative data source for cases where commercial imagerycannot be acquired Such cases often include studies in developing countries in areasthat have not usually been targets for major satellites needing supplemental low-clouddata requiring a time series or requiring a large number of images
12 Extent of the datasetAs of 30 September 1999 378 461 frames were included in the photograph
database (OYce of Earth Sciences 2000) comprising 99 missions from Mercury 3(21 July 1961) through STS-96 (27 May to 6 June 1999) The database was reviewedremoving photographs that were deemed unsuitable for remote sensing analysisbecause they were not Earth-looking (no entry for tilt angle) had no estimated focallength were over- or underexposed or were taken at oblique tilt angles (furtherdiscussion of these characteristics follows) The remaining 190 911 frames (504 ofthe records present in the database) represent photographs that are potentiallysuitable for remote sensing data As the database is always having new photographyadded statistics can be updated on request using the web link lsquoSummary of DatabaseContentsrsquo at the Gateway to Astronaut Photography of Earth (OYce of EarthSciences 2000)
13 ObjectivesThe overall purpose of this paper is to provide understanding of the properties of
astronaut-acquired orbital photographs to help scientists evaluate their applicabilityfor digital remote sensing Previous general descriptions of astronaut photographyhave been much less extensive and have tended to focus on numbers of photographstaken and geographic coverage and emphasized interpretative applications (Helfertand Wood 1989 Lulla et al 1993 1994 1996) By providing a unique compilationof the background information necessary to extend the use of astronaut photographybeyond interpretation to rigorous analysis it is hoped to provide a resource forscientists who could use this data in a variety of disciplines In keeping with apotential interdisciplinary audience we have tried to provide suYcient backgroundinformation for scientists who do not have extensive training in remote sensing
We focus on spatial resolution because it is one of the most important factors
Astronaut photography as digital data for remote sensing 4405
determining the suitability of an image for a remote sensing objective In remotesensing literature spatial resolution for aerial photography is often treated verydiVerently from spatial resolution of multispectral scanning sensors To providesuYcient background for a discussion of spatial resolution for a data source thatshares properties with both aerial photography and satellite remote sensing we rst summarize the ways that spatial resolution is typically quanti ed for aerialphotography and satellite remote sensors
Given this background information we (1 ) describe and illustrate factors thatin uence the spatial resolution of astronaut photographs (2) show examples ofestimating spatial resolution for a variety of photographs in a way that willenable users to make these calculations whether or not they have a background inphotogrammetry and (3) compare spatial resolution of digital data extracted fromastronaut photographs to data obtained from other satellites
2 BackgroundmdashSpatial resolution of imaging systemsSpatial resolution is a fundamental property of any imaging system used to
collect remote sensing data and directly determines the spatial scale of the resultantinformation Resolution seems intuitively obvious but its technical de nition andprecise application in remote sensing have been complex For example Townshend(1980) summarized 13 diVerent ways to estimate the resolving power of the LandsatMultiSpectral Scanner (MSS) Simonett (1983 p 20) stated lsquoIn the simplest casespatial resolution may be de ned as the minimum distance between two objects thata sensor can record distinctly[but] it is the format of the sensor system thatdetermines how spatial resolution is measuredrsquo By focusing attention on the proper-ties of the system (and not the images acquired) this de nition can be applied to avariety of types of images provided by diVerent systems In order to provide ascomplete a treatment as possible background is provided on several views of spatialresolution that are relevant when considering the spatial resolution of astronautphotography
21 Ground resolved distancePhotogrammetrists were measuring spatial resolution of aerial photographs using
empirical methods long before the rst scanning sensor was placed in orbit Thesemethods integrated properties of the imaging system with the other external factorsthat determine spatial resolution Ground resolved distance (GRD) is the parameterof most interest because it measures the applicability of an image to a speci c taskThe GRD of an image is de ned as the dimensions of the smallest discernible objectThe GRD is a function of geometry (altitude focal length of optics) equipment(internal system spatial resolution of the camera or scanner) and also on re ectancecharacteristics of the object compared to its surroundings (contrast)
Performance of a lm lm and camera or deployed aerial photography systemis measured empirically using standard targets that consist of black-and-white barsof graduated widths and spacings ( gure 6ndash9 in Slater et al 1983) The area-weightedaverage resolution (AWAR) in lp mm Otilde 1 ( line pairs mm Otilde 1 ) at the lm plane is deter-mined by measuring the smallest set of line pairs that can be discriminated on anoriginal lm negative or transparency Line pairs are quoted because it is necessaryto discriminate between one object and another to detect it and measure it
Film resolving power (in lp mm Otilde 1 ) is measured by manufacturers under standardphotographi c conditions (Smith and Anson 1968 Eastman Kodak Company 1998)
J A Robinson et al4406
at high contrast (objectbackground ratio 10001) and low contrast (objectback-ground ratio 161) Most terrestrial surfaces recorded from orbit are low contrastfor the purpose of estimating resolving power of lm Kodak no longer measures lm resolving power for non-aerial photographic lms (Karen Teitelbaum EastmanKodak Company personal communication) including lms that NASA routinelyuses for Earth photography
AWAR can be measured for the static case of lm and camera or for thecamerandashaircraft system in motion For example AWAR for the National AerialPhotography Program includes eVects of the lens resolving power of original lmimage blur due to aircraft motion and spatial resolution of duplicating lm (Light1996) Given AWAR for a system in motion the GRD can be calculated bytrigonometry (see equation (3) in sect5 d=1AWAR and D=GRD)
An impediment to similar rigorous measurement of GRD for orbital remotesensing systems is the lack of a target of suitable scale on the ground thus spatialresolution for most orbiting sensors is described in terms of a less all-encompassingmeasure instantaneous eld of view (see below) Additional challenges to measuringAWAR for a complete astronaut photography system include the number of diVerentoptions for aspects of the system including diVerent cameras lms and orbitalaltitudes These elements that must be standardized to determine AWAR providea useful list of those characteristics of astronaut photography that will mostin uence GRD
22 Instantaneous eld of viewThe instantaneous eld of view (IFOV) is generally used to represent the spatial
resolution of automated satellite systems and is commonly used interchangeablywith the term spatial resolution when comparing diVerent sensors IFOV is a combina-tion of geometric mechanical and electronic properties of the imaging systemGeometric properties including satellite orbital altitude detector size and the focallength of the optical system (Simonett 1983) The sensitivity of each detector elementat the wavelength desired plus the signal-to-noise level desired are electronic proper-ties that determine a minimum time for energy absorption For a linear sensor array(pushbroom scanning) this minimum time is translated to areal coverage by theforward velocity of the platform (Campbell 1996 p 97) Usually each detectorelement in the array corresponds to a pixel in the image Thus for a given altitudethe width of the pixel is determined by the optics and sensor size and the height ofthe pixel is determined by the rate of forward motion When magni ed by the ratioof the sensor altitude to the focal length of the optics of the sensor system IFOV isthe size of the area on the ground represented by an individual detector element(pixel Slater 1980 p 27)
The equation of IFOV with spatial resolution can be misleading because IFOVdoes not include factors other than geometrymdashfactors that largely determine thelevel of detail that can be distinguished in an image For example IFOV does notinclude characteristics of the target (contrast with surroundings shape of an objectcolour) atmospheric conditions illumination and characteristics of the interpreter(machine or human) Factors in uencing spatial resolution that are included in IFOVcan be calculated from design speci cations before the system has been built whereasthe actual spatial resolution of an image captured by the sensor will be unique tothat image IFOV represents the best spatial resolution possible given optimalconditions
Astronaut photography as digital data for remote sensing 4407
23 Imagery from scanners versus lm from camerasLike aerial photography systems the GRD of early remote sensing scanners
(electro-optical imaging systems) was calibrated empirically ( gures 1ndash6 in Simonett1983) but such methods are impractical for routine applications Satellites are nowcommon platforms for collecting remote sensing data small-scale aerial photographsare widely available within the USA and Europe and astronaut photographshave become available worldwide Straightforward comparison of resolutions ofphotographs to IFOVs of scanner images is necessary for many applications
How can spatial resolutions of data from diVerent sources be compared usingcommon units Oft-quoted lsquoresolution numbersrsquo actually IFOV of digital scanningsystems should be at least doubled for comparison to the standard method ofexpressing photographi c spatial resolution as GRD simply because of the objectcontrast requirement The following rule of thumb has been used in geographicapplications to compare spatial resolution of scanners and aerial photography (Welch1982 Jensen 1983) a low-contrast target may be converted to an approximate IFOVby dividing the GRD (of the photograph) by 24 Such a rule of thumb might alsobe reasonably applied to astronaut photography making it possible to convertobserved GRD to IFOV and IFOV to an estimated GRD
In this paper we discuss the spatial resolution of astronaut photographs in termsof system properties and compute the area on the ground represented by a singlepixel the equivalent to IFOV To complete our treatment of spatial resolution wealso provide estimates of the sizes of small features identi able in images for severalcases where this can be readily done
3 Factors that determine the footprint (area covered by the photograph )The most fundamental metric that forms the basis for estimating spatial resolution
of astronaut photographs is the size of the footprint or area on the ground capturedin a photograph (see review of satellite photogrammetry by Light 1980) The basicgeometric variables that in uence the area covered by an astronaut photograph are(1) the altitude of the orbit H (2) the focal length of the lens f (3) the actual sizeof the image on the lm d and (4) the orientation of the camera axis relative to theground (the obliquity or look angle t ) The relationships among these parametersare illustrated in gure 1 (also see equation (3) in sect5)
31 AltitudeHuman space ight missions have had a variety of primary objectives that required
diVerent orbital altitudes The higher the altitude the larger the footprint of thephotographs The diVering scale of photographs taken at diVerent altitudes is illus-trated in gure 2 The two photographs of Lake Eyre Australia were taken with thesame camera and lens but on diVerent dates and from diVerent altitudes In (a) thelake is relatively ooded while in (b) it is dry A 23timesdiVerence in altitude leads toa corresponding diVerence in the scales of the resulting photographs
32 L ensesThe longer the lens focal length the more magni cation the greater detail and
the smaller footprint A variety of lenses with diVerent focal lengths are own oneach space mission (table 1) The eVect of lens length on spatial coverage and imagedetail is shown in gure 3 In these views of Houston (a)ndash(c) taken from approxi-mately similar altitudes with the same camera most of the diVerence in scale of the
J A Robinson et al4408
Figure 1 Geometric relationship between look angle (t) spacecraft nadir point (SN) photo-graph centre point (PC) and photograph principal point (PP) The dark shaded arearepresents Earthrsquos surface The distance covered on the ground D corresponds totable 5 Original image size d is listed for various lms in table 1 Variables correspondto equations (2)ndash(9) in sect5
Astronaut photography as digital data for remote sensing 4409
(a) (b)
Figure 2 Example of the eVect of altitude on area covered in a photograph Both photographsof Lake Eyre Australia were taken using a Hasselblad camera with 100 mm lens(a) altitude 283 km STS093-702-062 (b) altitude 617 km STS082-754-61
photographs is due to the diVerent magni cations of the lenses Although taken froma diVerent altitude and using a diVerent camera with a diVerent original image size(see table 2) an electronic still camera image taken with a 300 mm lens is alsoincluded for comparison
For remote sensing longer focal length lenses are generally preferred (250 mm or350 mm for Hasselblad 300 mm or 400 mm lenses for 35 mm format cameras andESCs) Unfortunately longer-focal-length lenses exhibit poorer performance towardthe edge of a frame For example a 250 mm lens (Distagon CF 56) used on theHasselblad camera has spatial resolution of 57 lp mm Otilde 1 at the centre but only51 lp mm Otilde 1 at the edge and 46 lp mm Otilde 1 at the corner (tested with Ektachrome 5017f8 aperture and high contrast) A longer 300 mm lens used on the Nikon camerahas a greater spatial resolution diVerence between centre (82 lp mm Otilde 1 ) and corner(49 lp mm Otilde 1 ) using Kodak 5017 Ektachrome f4 aperture and high contrast FredPearce (unpublished data) There are tradeoVs among lens optics and speed Forexample the lenses for the Linhof system and the 250 mm lens for the Hasselblad(Distagon CF 56 see footnotes to table 1) are limited to apertures smaller thanf56 This becomes an important constraint in selecting shutter speeds (see discussionof shutter speed in sect42)
33 Cameras and actual image sizesAfter passing through the lens the photographic image is projected onto lm
inside the camera The size of this original image is another important propertydetermining spatial resolution and is determined by the camera used Cameraformats include 35 mm and 70 mm (Lowman 1980 Amsbury 1989) and occasionally5times4 inch (127times100 mm) and larger (table 1) Cameras own on each mission arenot metricmdashthey lack vacuum platens or reseau grids image-motion compensationor gyro-stabilized mounts The workhorse for engineering and Earth photographyon NASA missions has been a series of 70 mm Hasselblad cameras (table 1) chosenfor their reliability The modi ed magazine databack imprints a unique number and
J A Robinson et al4410
Tab
le1
Det
ails
on
the
cam
eras
use
dfo
rast
ron
aut
ph
oto
gra
phy
of
Eart
h
Data
incl
ud
ep
ho
togr
aphy
from
all
NA
SA
hu
man
spac
eig
ht
mis
sion
sth
rough
ST
S-9
6(J
un
e19
99
378
461
tota
lfr
am
esin
the
dat
abas
e)1
Imag
esi
zeT
ypic
al
lense
sN
um
ber
of
Cam
era
make
(mm
timesm
m)
use
d(m
m)
ph
oto
gra
ph
sP
erio
dof
use
No
tes
Has
selb
lad
55times
55
40
50100
2502
288
494
1963ndashp
rese
nt
Lin
ho
f10
5times120
90
250
29
373
1984ndashp
rese
nt
No
won
lyo
wn
occ
asio
nal
lyM
ult
ispec
tral
Cam
era
(S19
0A
)57times
57152
32
913
1973ndash19
74S
kyla
b
xed
-mou
nt3
cam
era
(NA
SA
1974
)E
art
hT
erra
inC
amer
a(S
190B
)11
5times11
5457
5478
1973ndash19
74S
kyla
b
xed
-mou
nt3
cam
era
(NA
SA
1974
)R
ole
iex
55times
55
50
100250
4352
1990ndash19
92L
arg
eF
orm
at
Cam
era4
229
times457
305
2143
1984
Mis
sio
nST
Sndash41G
only
(Mer
kel
etal
198
5)
Maure
r55
times55
76
80818
1961ndash19
68
1 Sm
all-f
orm
atca
mer
as
(35
mm
lm
and
ES
C)
are
no
tin
clud
edin
this
table
bec
ause
of
the
larg
enu
mb
erof
len
ses
and
con
gu
rati
on
sth
at
have
bee
nu
sed
and
bec
au
seth
ese
data
are
no
tcu
rren
tly
incl
uded
inth
edata
bas
efo
rall
mis
sion
s2 L
ense
suse
dfo
rH
ass
elbla
dm
anufa
cture
dby
Carl
Zei
ss
Dis
tago
nC
F4
(40
mm
)D
ista
gon
CF
4(5
0m
m)
Dis
tagon
CF
35
(100
mm
)D
ista
go
nC
F5
6(2
50
mm
)S
tand
ard
Has
selb
lad
len
ses
ow
nw
ill
chan
ge
inth
eyea
r20
00
toD
ista
gon
FE
28
(50
mm
)D
ista
go
nF
E2
(110
mm
)T
ele-
Tes
sar
FE
4(2
50m
m)
and
Tel
e-T
essa
rF
E4
(350
mm
)3 T
hes
eca
mer
as
wer
ein
xed
mou
nti
ngs
on
the
outs
ide
of
the
Skyla
bsp
ace
stati
on
A
ltho
ugh
not
hand
hel
d
the
lm
isca
talo
gued
wit
ho
ther
ast
ron
aut
ph
oto
gra
ph
yan
dis
incl
uded
her
efo
rco
mp
lete
nes
s4 A
NA
SA
-des
igned
cam
era
wit
hatt
itud
ere
fere
nce
syst
emfo
rd
eter
min
ing
pre
cise
nadir
poin
ts
Dat
ais
no
tcu
rren
tly
inco
rpora
ted
into
on
lin
ed
atab
ase
bu
tis
cata
loged
by
Mer
kel
etal
(1
985
)
Astronaut photography as digital data for remote sensing 4411
(a) (b)
(c) (d)
Figure 3 Example of the eVect of lens focal length on resolving power Hasselblad imagesof Houston were all taken from an altitude of 294 kmndash302 km with diVerent lenses(a) 40 mm STS062-97-151 (b) 100 mm STS094-737-28 (c) 250 mm STS055-71-41 Forcomparison an Electronic Still Camera image with 300 mm lens at 269 km altitude isalso shown (d ) S73E5145
timestamp on each frame at the time of exposure Cameras are serviced between ights Occasionally there is enough volume and mass allowance so that a Linhof5times4 inch (127times101 mm) format camera can be own Nikon 35 mm cameras are own routinely also because of proven reliability Electronic still cameras (ESC)were tested for Earth photography beginning in 1992 (Lulla and Holland 1993) Inan ESC a CCD (charge-coupled device) is used as a digital replacement for lmrecording the image projected inside the camera An ESC (consisting of a KodakDCS 460c CCD in a Nikon N-90S body) was added as routine equipment forhandheld photographs in 1995 ESCs have also been operated remotely to captureand downlink Earth images through a NASA-sponsored educational programme(EarthKAM) Discussion of CCD spatial array and radiometric sensitivity arebeyond the scope of this paper but are summarized by Robinson et al (2000b) The
J A Robinson et al4412
format of the lm (or CCD) and image size projected onto the lm (or CCD) aresummarized for all the diVerent cameras own in table 2
34 L ook angle or obliquityNo handheld photographs can be considered perfectly nadir they are taken at a
variety of look angles ranging from near vertical ( looking down at approximatelythe nadir position of the spacecraft ) to high oblique (images that include the curvatureof Earth) Imaging at oblique look angles leads to an image where scale degradesaway from nadir A set of views of the same area from diVerent look angles is shownin gure 4 The rst two shots of the island of Hawaii were taken only a few secondsapart and with the same lens The third photograph was taken on a subsequentorbit and with a shorter lens The curvature of the Earth can be seen in the upperleft corner
Obliquity can be described qualitatively ( gure 4) or quantitatively as the lookangle (t gure 1 calculations described in formulation 2 (sect52)) Because obliquityand look angle have such a dramatic in uence on the footprint we summarize thedatabase characteristics relative to these two parameters Figure 5 is a breakdownof the spatial resolution characteristics of low oblique and near vertical photographsin the NASA Astronaut Photography database Number of photographs are grouped(1) by calculated values for look angle (t ) and (2) by altitude After observing theoverlap between near vertical and low oblique classes we are currently restructuringthis variable (lsquotiltrsquo) in the database to provide a measure of t when available Userswill still be able to do searches based on the qualitative measures but these measureswill be more closely tied to actual look angle
341 Obliquity and georeferencing digitised photographsOften the rst step in a remote sensing analysis of a digitized astronaut photo-
graph is to georeference the data and resample it to conform to a known mapprojection Details and recommendations for resampling astronaut photographydata are provided by Robinson et al (2000a 2000c) and a tutorial is also available(McRay et al 2000) Slightly oblique photographs can be geometrically correctedfor remote sensing purposes but extremely oblique photographs are not suited forgeometric correction When obliquity is too great the spatial scale far away fromnadir is much larger than the spatial scale closer to nadir resampling results areunsuitable because pixels near nadir are lost as the image is resampled while manypixels far away from nadir are excessively replicated by resampling
To avoid the generation of excess pixels during georeferencing the pixel sizes ofthe original digitized image should be smaller than the pixels in the nal resampledimage Calculations of original pixel size using methods presented below can beuseful in ensuring meaningful resampling For slightly oblique images formulation 3(sect53) can be used to estimate pixel sizes at various locations in a photograph (nearnadir and away from nadir) and these pixel sizes then used to determine a reasonablepixel scale following resampling
4 Other characteristics that in uence spatial resolutionBeyond basic geometry many other factors internal to the astronaut photography
system impact the observed ground resolved distance in a photograph The lensesand cameras already discussed are imperfect and introduce radiometric degradations(Moik 1980)Vignetting (slight darkening around the edges of an image) occurs in
Astronaut photography as digital data for remote sensing 4413
Tab
le2
Max
imu
mpo
ssib
led
igit
alsp
atia
lre
solu
tio
n(p
ixel
size
inm
)fo
rca
mer
asy
stem
scu
rren
tly
use
dby
ast
ron
auts
top
ho
togr
aph
eart
hb
ase
do
nth
ege
om
etry
of
alt
itud
ele
ns
and
lm
size
C
alc
ula
tion
sas
sum
eth
atco
lour
tran
spar
ency
lm
isdig
itiz
edat
2400
ppi
(pix
els
per
inch
10
6m
mpix
elOtilde
1 )
Min
imu
mgro
un
ddis
tance
repre
sente
dby
1p
ixel
(m)
As
afu
nct
ion
of
alti
tude1
Imag
esi
zeM
inim
um
alt
itud
eM
edia
nalt
itu
de
Maxi
mum
alt
itud
eC
amer
asy
stem
mm
timesm
mp
ixel
s2(2
22k
m)
(326
km
)(6
11
km
)
Lin
ho
f125
mm
90
mm
lens
105
times120
8978times
11
434
261
383
718
250
mm
lens
94
138
259
Has
selb
lad
70
mm
100
mm
lens
55times
55
5198
times519
823
534
564
725
0m
mle
ns
94
138
259
Nik
on
35m
m30
0m
mle
ns
36times
24
3308
times217
478
115
216
400
mm
lens
59
86
162
ESC
35m
m18
0m
mle
ns3
27
6times
18
530
69times
204
311
116
330
630
0m
mle
ns
67
98
184
400
mm
lens
50
74
138
1 Alt
itud
essh
ow
nar
em
inim
um
m
edia
nan
dm
axim
um
thro
ugh
ST
S-8
9(J
anuary
199
8N
=84
mis
sion
s)
2 Act
ual
pix
els
for
ES
C
oth
erw
ise
assu
med
dig
itiz
ati
on
at240
0pp
i(p
ixel
sp
erin
ch)
equ
alto
10
6m
mpix
elOtilde
1 3 T
he
mo
stco
mm
on
con
gura
tion
for
Eart
hph
oto
gra
ph
yas
part
of
the
Eart
hK
AM
pro
ject
J A Robinson et al4414
Figure 4 De nitions of look angle for photographs taken from low orbit around EarthThe example photographs of Hawaii were all taken from approximately the samealtitude (370ndash398 km) and with the same (250 mm) lens on a Hasselblad camera(STS069-701-69 STS069-701-70 and STS069-729-27 [100 mm lens])
astronaut photography due to light path properties of the lenses used A lack of atness of the lm at the moment of exposure (because cameras used in orbit donot incorporate a vacuum platen) also introduces slight degradations A numberof additional sources of image degradation that would aVect GRD of astronautphotographs are listed
41 Films and processing411 Colour reversal lms
Most lms used in NASA handheld cameras in orbit have been E-6 processcolour reversal lms (Ektachromes) that have a nominal speed of 64 to 100 ASAalthough many other lms have been own (table 3) Reversal lms are used becausedamage from radiation exposure while outside the atmospheric protection of Earthis more noticeable in negative lms than in positive lms (Slater 1996) The choiceof ASA has been to balance the coarser grains ( lower lm resolving power) of high-speed lms with the fact that slower lms require longer exposures and thus areaVected more by vehicle motion (approximately 73 km s Otilde 1 relative to Earthrsquos surfacefor the Space Shuttle) Extremely fast lms (gt400 ASA) are also more susceptibleto radiation damage in orbit (particularly during long-duration or high-altitudemissions Slater 1996) and have not traditionally been used for Earth photography The manufacturer stock numbers identifying the lm used are available for eachimage in the Astronaut Photography Database (OYce of Earth Sciences 2000)
412 Colour infrared lmsColour infrared lm (CIR) was used during an Earth-orbiting Apollo mission
(Colwell 1971) in the multispectral S-190A and the high-resolution S-190B camera
Astronaut photography as digital data for remote sensing 4415
Figure 5 (a) Distributions of astronaut photographs by look angle oV-nadir (calculated fromcentre point and nadir point using Formulation 2) Data included for all photographsfor which centre and nadir points are known with high oblique look angles (n=55 155) excluded (Total sample shown is 108 293 of 382 563 total records in thedatabase when compiled For records where nadir data was not available number ofobservations for qualitative look angles are near vertical 14 086 low oblique 115 834undetermined 89 195) (b) Altitude distribution for astronaut photographs of Earthtaken with the Hasselblad camera Data included for Hasselblad camera only focallength determined with high oblique look angles excluded (Total sample shown is159 046 of 206 971 Hasselblad records with known altitude and of 382 563 total recordsin the database when compiled For Hasselblad records with known altitude data notshown includes high oblique photographs using 100 mm and 250 mm lenses n=20 262and all look angles with other lenses n=27 663)
systems on Skylab (NASA 1974 Wilmarth et al 1977) and occasionally on Shuttlemissions and Shuttle-Mir missions (table 3) The CIR lm used is a three-layerAerochrome having one layer that is sensitive to re ected solar infrared energy toapproximately 900 nm (Eastman Kodak Company 1998) protective coatings onmost spacecraft windows also limit IR transmittance between 800 mm and 1000 nmThis layer is extremely sensitive to the temperature of the lm which createsunpredictable degradation of the IR signature under some space ight conditions
J A Robinson et al4416
Tab
le3
Det
ails
on
lm
suse
dfo
rast
ron
aut
pho
togr
aphy
of
Eart
h
Dat
ain
clud
ep
ho
togr
aphy
fro
mall
NA
SA
hum
an
spac
eig
ht
mis
sio
ns
thro
ugh
ST
S-9
6(J
un
e19
99
378
461
tota
lfr
ames
inth
edata
base
)F
ilm
sw
ith
lt50
00fr
am
esex
po
sed
and
lm
suse
dfo
r35
mm
cam
eras
only
are
no
tin
clud
ed
Res
olv
ing
Nu
mb
erof
Film
man
ufa
ctu
rer
AS
Ap
ow
er1
fram
esP
erio
dof
use
Cam
eras
Colo
ur
posi
tive
Kod
akA
eroch
rom
eII
(2447
2448
SO
-131S
O-3
68
)32
40ndash50
513
8196
8ndash19
85
Hass
elbla
dL
inh
of
Kod
akE
kta
chro
me
(5017
502
56
017
6118
SO
-121
64
50
113
932
196
5ndash19
68
Hass
elbla
dL
inh
of
Role
iex
SO
-217
Q
X-8
24
QX
-868
)19
81ndash1993
Mau
rer
Nik
on
Kod
akL
um
iere
(5046
5048
)100
105
743
199
4ndash19
98
Hass
elbla
dL
inho
fK
od
akE
lite
100
S(5
069
)100
41
486
199
6ndashp
rese
nt
Hass
elbla
d2
Fu
jiV
elvia
50C
S135
-36
32
13
882
1992ndash19
97H
ass
elbla
dN
iko
nK
od
akH
i-R
esolu
tio
nC
olo
r(S
O-2
42S
O-3
56
)10
096
74
1973ndash19
75H
ass
elbla
d
S19
0A
S190
B
Infr
are
dand
bla
ck-a
nd-w
hit
e
lms
Kod
akA
eroch
rom
eC
olo
rIn
frar
ed40
32
40
704
196
5ndashp
rese
nt2
Hass
elbla
dL
inh
of
Nik
on
S190
A
(8443
34
432443
)S
190B
Kod
akB
ampW
Infr
are
d(2
424
)32
10
553
197
3ndash19
74
S190
AK
od
akP
anato
mic
-XB
ampW
(SO
-022
)63
10
657
197
3ndash19
74
S190
A
1A
tlo
wco
ntr
ast
(ob
ject
back
grou
nd
rati
o16
1)
aspro
vid
edby
Ko
dak
and
list
edby
Sla
ter
etal
(1
983
256
)R
esolv
ing
pow
erw
as
dis
con
tin
ued
fro
mth
ete
chn
ical
test
sper
form
edb
yK
odak
inapp
roxim
atel
y19
93
an
dit
isn
ot
kn
ow
nif
curr
ent
lm
sh
ave
sim
ilar
reso
lvin
gpo
wer
too
lder
ver
sion
so
fsi
milar
lm
s(K
T
eite
lbaum
K
od
akp
erso
nal
com
mu
nic
atio
n)
Th
egra
nu
lari
tym
easu
reno
wpro
vid
edb
yK
od
ak
can
no
tb
ere
adily
con
ver
ted
toa
spat
ial
reso
luti
on
2 The
Lin
hof
was
use
dag
ain
on
ST
S-1
03in
Dec
emb
er19
99(d
ata
not
cata
logued
asof
the
pre
para
tion
of
this
tab
le)
usi
ng
Ko
dak
Elite
100S
lm
3 B
egin
nin
gin
July
1999
2443
colo
ur-
infr
are
d
lmpro
cess
was
chan
ged
from
EA
-5to
E-6
Astronaut photography as digital data for remote sensing 4417
413 Film duplicationThe original lm is archived permanently after producing about 20 duplicate
printing masters (second generation) Duplicate printing masters are disseminatedto regional archives and used to produce products (thirdndashfourth generation) for themedia the public and for scienti c use When prints slides transparencies anddigital products are produced for public distribution they are often colour adjustedto correspond to a more realistic look Digital products from second generationcopies can be requested by scienti c users (contact the authors for information onobtaining such products for a particular project) and these are recommended forremote sensing applications Care should be taken that the digital product acquiredfor remote sensing analysis has not been colour adjusted for presentation purposesBased on qualitative observations there is little visible increase in GRD in the thirdor fourth generation products There is a signi cant degradation in delity of colourin third and fourth generation duplicates because an increase in contrast occurswith every copy
42 Shutter speedsThe impact of camera motion (both due to the photographer and to the motion
of the spacecraft ) on image quality is determined by shutter speedmdash1250 to 1500second have been used because slower speeds record obvious blurring caused bythe rapid motion of the spacecraft relative to the surface of the Earth Generally1250 was used for ISO 64 lms (and slower) and after the switch to ISO 100 lmsa 1500 setting became standard New Hasselblad cameras that have been ownbeginning with STS-92 in October 2000 vary shutter speed using a focal plane shutterfor exposure bracketing (rather than varying aperture) and 1250 1500 and 11000second are used
Across an altitude range of 120ndash300 nautical miles (222ndash555 km) and orbitalinclinations of 285deg and 570deg the median relative ground velocity of orbitingspacecraft is 73 km s Otilde 1 At a nominal shutter speed of 1500 the expected blur dueto motion relative to the ground would be 146 m When cameras are handheld (asopposed to mounted in a bracket) blur can be reduced by physical compensationfor the motion (tracking) by the photographer many photographs show a level ofdetail indicating that blur at this level did not occur (eg gure 3(d )) Thus motionrelative to the ground is not an absolute barrier to ground resolution for handheldphotographs
43 Spacecraft windowsMost of the photographs in the NASA Astronaut Photography Database were
taken through window ports on the spacecraft used The transmittance of the windowport may be aVected by material inhomogeniety of the glass the number of layersof panes coatings used the quality of the surface polish environmentally inducedchanges in window materials (pressure loads thermal gradients) or deposited con-tamination Such degradation of the window cannot be corrected is diVerent foreach window and changes over time See Eppler et al (1996) and Scott (2000) fordiscussion of the spectral transmittance of the fused quartz window that is part ofthe US Laboratory Module (Destiny) of the International Space Station
44 Digitized images from lmWhen lm is scanned digitally the amount of information retained depends on
the spectral information extracted from the lm at the spatial limits of the scanner
J A Robinson et al4418
To date standard digitizing methodologies for astronaut photographs have not beenestablished and lm is digitized on a case-by-case basis using the equipment available
441 Digitizing and spatial resolutionLight (1993 1996) provided equations for determining the digitizing spatial
resolution needed to preserve spatial resolution of aerial photography based on thestatic resolving power (AWAR) for the system For lms where the manufacturer hasprovided data (Eastman Kodak 1998 K Teitelbaum personal communication) the resolving power of lms used for astronaut photography ranges from 32 lp mm Otilde 1to 100 lp mm Otilde 1 at low contrast (objectbackground ratio 161 table 3) The AWARfor the static case of the Hasselblad camera has been measured at high and lowcontrast (using lenses shown in table 1) with a maximum of approximately55 lp mm Otilde 1 (Fred Pearce unpublished data)
Based on the method of Light (1993 1996) the dimension of one spatial resolutionelement for a photograph with maximum static AWAR of 55 lp mm Otilde 1 would be18 mm lp Otilde 1 The acceptable range of spot size to preserve spatial information wouldthen be
18 mm
2 atilde 2aring scan spot size aring
18 mm
2(1)
and 6 mm aring scan spot size aring 9 mm Similarly for a more typical static AWAR of30 lp mm Otilde 1 (33 mm lpOtilde 1 low contrast Fred Pearce unpublished data) 11 mm aring scanspot sizearing 17 mm These scan spot sizes correspond to digitizing resolutions rangingfrom 4233ndash2822 ppi (pixels inch Otilde 1 ) for AWAR of 55 lp mm Otilde 1 and 2309ndash1494 ppi forAWAR of 30 lp mm Otilde 1 Scan spot sizes calculated for astronaut photography arecomparable to those calculated for the National Aerial Photography Program(9ndash13 mm Light 1996)
Widely available scanners that can digitize colour transparency lm currentlyhave a maximum spatial resolution of approximately 2400 ppi (106 mm pixel Otilde 1) Forexample we routinely use an Agfa Arcus II desktop scanner (see also Baltsavias1996) with 2400 ppi digitizing spatial resolution (2400 ppi optical resolution in onedirection and 1200 ppi interpolated to 2400 ppi in the other direction) and 2400 ppiwas used to calculate IFOV equivalents (tables 2 and 4) For some combinations oflens camera and contrast 2400 ppi will capture nearly all of the spatial informationcontained in the lm However for lm with higher resolving power thanEktachrome-64 for better lenses and for higher contrast targets digitizing at 2400 ppiwill not capture all of the spatial information in the lm
Improvements in digitizing technology will only produce real increases in IFOVto the limit of the AWAR of the photography system The incremental increase inspatial information above 3000 ppi (85 mm pixel Otilde 1 ) is not likely to outweigh thecosts of storing large images (Luman et al 1997) At 2400 ppi a Hasselblad frameis approximately 5200times5200 or 27 million pixels (table 2) while the same imagedigitized at 4000 ppi would contain 75 million pixels
Initial studies using astronaut photographs digitized at 2400 ppi (106 mm pixel Otilde 1 Webb et al 2000 Robinson et al 2000c) indicate that some GRD is lost comparedto photographic products Nevertheless the spatial resolution is still comparablewith other widely used data sources (Webb et al 2000) Robinson et al (2000c)found that digitizing at 21 mm pixel Otilde 1 provided information equivalent to 106 mmpixel Otilde 1 for identifying general urban area boundaries for six cities except for a single
Astronaut photography as digital data for remote sensing 4419
photo that required higher spatial resolution digitizing (it had been taken with ashorter lens and thus had less spatial resolution) Part of the equivalence observedin the urban areas study may be attributable to the fact that the atbed scannerused interpolates from 1200 ppi to 2400 ppi in one direction The appropriate digitiz-ing spatial resolution will in part depend on the scale of features of interest to theresearchers with a maximum upper limit set of a scan spot size of approximately6 mm (4233 ppi)
442 Digitizing and spectral resolutionWhen colour lm is digitized there will be loss of spectral resolution The three
lm emulsion layers (red green blue) have relatively distinct spectral responses butare fused together so that digitizers detect one colour signal and must convert itback into three (red green blue) digital channels Digital scanning is also subject tospectral calibration and reproduction errors Studies using digital astronaut photo-graphs to date have used 8 bits per channel However this is another parameterthat can be controlled when the lm is digitized We do not further address spectralresolution of digitally scanned images in this paper
45 External factors that in uence GRDAlthough not discussed in detail in this paper factors external to the spacecraft
and camera system (as listed by Moik (1980) for remote sensing in general) alsoimpact GRD These include atmospheric interference (due to scattering attenuationhaze) variable surface illumination (diVerences in terrain slope and orientation) andchange of terrain re ectance with viewing angle (bidirectional re ectance) For astro-naut photographs variable illumination is particularly important because orbits arenot sun-synchronous Photographs are illuminated by diVerent sun angles and imagesof a given location will have colour intensities that vary widely In addition theviewing angle has an eVect on the degree of object-to-backgroun d contrast andatmospheric interference
5 Estimating spatial resolution of astronaut photographsIn order to use astronaut photographs for digital remote sensing it is important
to be able to calculate the equivalent to an IFOVmdashthe ground area represented bya single pixel in a digitized orbital photograph The obliquity of most photographsmeans that pixel lsquosizesrsquo vary at diVerent places in an image Given a ground distanceD represented by a photograph in each direction (horizontal vertical ) an approxi-mate average pixel width (P the equivalent of IFOV) for the entire image can becalculated as follows
P=D
dS(2)
where D is the projected distance on the ground covered by the image in the samedirection as the pixel is measured d is the actual width of the image on the original lm (table 1) and S is the digitizing spatial resolution
Here we present three mathematical formulations for estimating the size of thefootprint or area on the ground covered by the image Example results from theapplication of all three formulations are given in table 5 The rst and simplestcalculation (formulation 1) gives an idea of the maximum spatial resolution attainableat a given altitude of orbit with a given lm format and a perfectly vertical (nadir)
J A Robinson et al4420
view downward Formulation 2 takes into account obliquity by calculating lookangle from the diVerence between the location on the ground represented at thecentre of the photograph and the nadir location of the spacecraft at the time thephotograph was taken ( gure 1) Formulation 3 describes an alternate solution tothe oblique look angle problem using coordinate-system transformations Thisformulation has been implemented in a documented spreadsheet and is availablefor download (httpeoljscnasagovsseopFootprintCalculatorxls) and is in theprocess of being implemented on a Web-based user interface to the AstronautPhotography Database (OYce of Earth Sciences 2000)
Although formulations 2 and 3 account for obliquity for the purposes of calcula-tion they treat the position of the spacecraft and position of the camera as one Inactuality astronauts are generally holding the cameras by hand (although camerasbracketed in the window are also used) and the selection of window position of theastronaut in the window and rotation of the camera relative to the movement ofthe spacecraft are not known Thus calculations using only the photo centre point(PC) and spacecraft nadir point (SN) give a locator ellipse and not the locations ofthe corners of the photograph A locator ellipse describes an estimated area on theground that is likely to be included in a speci c photograph regardless of the rotationof the lm plane about the camerarsquos optical axis ( gure 6)
Estimating the corner positions of the photo requires additional user input of asingle auxiliary pointmdasha location on the image that has a known location on theground Addition of this auxiliary point is an option available to users of thespreadsheet An example of the results of adding an auxiliary point is shown in gure 7 with comparisons of the various calculations in table 5
51 Formulation 1 Footprint for a nadir viewThe simplest way to estimate footprint size is to use the geometry of camera lens
and spacecraft altitude to calculate the scaling relationship between the image in the
Figure 6 Sketch representation of the locator ellipse an estimated area on the ground thatis likely to be included in a speci c photograph regardless of the rotation of the lmplane about the camerarsquos optical axis
Astronaut photography as digital data for remote sensing 4421
Figure 7 An example of the application of the Low Oblique Space Photo FootprintCalculator The photograph (STS093-704-50) of Limmen Bight Australia was takenfrom an altitude of 283 km using the Hasselblad camera with a 250 mm lens Mapused is 11 000 000 Operational Navigational Chart The distances shown equate toan average pixel size of 124 m for this photograph
lm and the area covered on the ground For a perfect nadir view the scalerelationship from the geometry of similar triangles is
d
D=
f
H(3)
where d=original image size D=distance of footprint on the ground f =focallength of lens and H=altitude Once D is known pixel size ( length=width) can becalculated from equation (2) These calculations represent the minimum footprintand minimum pixel size possible for a given camera system altitude and digitisingspatial resolution (table 2) Formulation 1 was used for calculating minimum pixelsizes shown in tables 2 and 4
By assuming digitizing at 2400 ppi (106 mm pixel Otilde 1 ) currently a spatial resolutioncommonly attainable from multipurpose colour transparency scanners (see sect441)this formulation was used to convert area covered to an IFOV equivalent for missionsof diVerent altitudes (table 2) Table 4 provides a comparison of IFOV of imagesfrom various satellites including the equivalent for astronaut photography Valuesin this table were derived by using formulation 1 to estimate the area covered becausea perfect nadir view represents the best possible spatial resolution and smallest eldof view that could be obtained
52 Formulation 2 Footprint for oblique views using simpli ed geometry and thegreat circle distance
A more realistic approach to determining the footprint of the photographaccounts for the fact that the camera is not usually pointing perfectly down at the
J A Robinson et al4422
Table 4 Comparison of pixel sizes (instantaneous elds of view) for satellite remote sensorswith pixel sizes for near vertical viewing photography from spacecraft Note that alldata returned from the automated satellites have the same eld of view while indicatedpixel sizes for astronaut photography are best-case for near vertical look angles See gure 5 for distributions of astronaut photographs of various look angles High-altitude aerial photography is also included for reference
Altitude PixelInstrument (km) width (mtimesm) Bands1
Landsat Multipectral Scanner2 (MSS 1974-) 880ndash940 79times56 4ndash5Landsat Thematic Mapper (TM 1982-) ~705 30times30 7Satellite pour lrsquoObservation de la Terre (SPOT 1986-) ~832
SPOT 1-3 Panchromatic 10times10 1SPOT 1-3 Multispectral (XS) 20times20 3
Astronaut Photography (1961-)Skylab3 (1973-1974 ) ~435
Multispectral Camera (6 camera stations) 17ndash19times17ndash19 3ndash8Earth Terrain Camera (hi-res colour) 875times875 3
Space Shuttle5 (1981-) 222ndash611Hasselblad 70 mm (250mm lens colour) vertical view 9ndash26times9ndash26 3Hasselblad 70 mm (100mm lens colour) vertical view 23ndash65times23ndash65 3Nikon 35 mm (400mm lens colour lm or ESC) vertical 5ndash16times5ndash16 3
High Altitude Aerial Photograph (1100 000) ~12 2times2 3
1Three bands listed for aerial photography obtainable by high-resolution colour digitizingFor high-resolution colour lm at 65 lp mmOtilde 1 (K Teitelbaum Eastman Kodak Co personalcommunication) the maximum information contained would be 3302 ppi
2Data from Simonett (1983Table 1-2)3Calculated as recommended by Jensen (19831596) the dividend of estimated ground
resolution at low contrast given in NASA (1974179-180) and 244Six lters plus colour and colour infrared lm (NASA 19747) 5Extracted from table 2
nadir point The look angle (the angle oV nadir that the camera is pointing) can becalculated trigonometrically by assuming a spherical earth and calculating the dis-tance between the coordinates of SN and PC ( gure 1) using the Great Circledistance haversine solution (Sinnott 1984 Snyder 198730ndash32 Chamberlain 1996)The diVerence between the spacecraft centre and nadir point latitudes Dlat=lat 2 shy lat 1 and the diVerence between the spacecraft centre and nadir pointlongitudes Dlon=lon 2 shy lon 1 enter the following equations
a=sin2ADlat
2 B+cos(lat 1) cos(lat 2) sin2ADlon
2 B (4)
c=2 arcsin(min[1 atilde a]) (5)
and oVset=R c with R the radius of a spherical Earth=6370 kmThe look angle (t ) is then given by
t=arctanAoffset
H B (6)
Assuming that the camera was positioned so that the imaginary line between thecentre and nadir points (the principal line) runs vertically through the centre of thephotograph the distance between the geometric centre of the photograph (principalpoint PP) and the top of the photograph is d2 ( gure 8(a)) The scale at any point
Astronaut photography as digital data for remote sensing 4423
(a) (b)
Figure 8 The geometric relationship between look angle lens focal length and the centrepoint and nadir points of a photograph (see also Estes and Simonett 1975) Note thatthe Earthrsquos surface and footprint represented in gure 1 are not shown here only arepresentation of the image area of the photograph Variables shown in the gurelook angle t focal length f PP centre of the image on the lm SN spacecraft nadird original image size (a) The special case where the camera is aligned squarely withthe isometric parallel In this case tilt occurs along a plane parallel with the centre ofthe photograph When the conditions are met calculations using Formulation 3 willgive the ground coordinates of the corner points of the photograph (b) The generalrelationship between these parameters and key photogrammetric points on the photo-graph For this more typical case coordinates of an additional point in the photographmust be input in order to adjust for twist of the camera and to nd the corner pointcoordinates
in the photograph varies as a function of the distance y along the principal linebetween the isocentre and the point according to the relationship
d
D=
f shy ysint
H(7)
(Wong 1980 equation (214) H amp the ground elevation h) Using gure 8(a) at thetop of the photo
y= f tanAt
2B+d
2(8)
and at the bottom of the photo
y= f tanAt
2Bshyd
2(9)
Thus for given PC and SN coordinates and assuming a photo orientation as in gure 8(a) and not gure 8(b) we can estimate a minimum D (using equations (7)and (8)) and a maximum D (using equations (7) and (9)) and then average the twoto determine the pixel size (P ) via equation (2)
J A Robinson et al4424
53 Formulation 3 T he L ow Oblique Space Photo Footprint CalculatorThe Low Oblique Space Photo Footprint Calculator was developed to provide
a more accurate estimation of the geographic coordinates for the footprint of a lowoblique photo of the Earthrsquos surface taken from a human-occupied spacecraft inorbit The calculator performs a series of three-dimensional coordinate transforma-tions to compute the location and orientation of the centre of the photo exposureplane relative to an earth referenced coordinate system The nominal camera focallength is then used to create a vector from the photorsquos perspective point througheach of eight points around the parameter of the image as de ned by the formatsize The geographical coordinates for the photo footprint are then computed byintersecting these photo vectors with a spherical earth model Although more sophist-icated projection algorithms are available no signi cant increase in the accuracy ofthe results would be produced by these algorithms due to inherent uncertainties inthe available input data (ie the spacecraft altitude photo centre location etc)
The calculations were initially implemented within a Microsoft Excel workbookwhich allowed one to embed the mathematical and graphical documentation nextto the actual calculations Thus interested users can inspect the mathematical pro-cessing A set of error traps was also built into the calculations to detect erroneousresults A summary of any errors generated is reported to the user with the resultsof the calculations Interested individuals are invited to download the Excel work-book from httpeoljscnasagovsseopFootprintCalculatorxls The calculations arecurrently being encoded in a high-level programming language and should soon beavailable alongside other background data provided for each photograph at OYceof Earth Sciences (2000)
For the purposes of these calculations a low oblique photograph was de ned as
one with the centre within 10 degrees of latitude and longitude of the spacecraftnadir point For the typical range of spacecraft altitudes to date this restricted thecalculations to photographs in which Earthrsquos horizon does not appear (the generalde nition of a low oblique photograph eg Campbell 199671 and gure 4)
531 Input data and resultsUpon opening the Excel workbook the user is presented with a program intro-
duction providing instructions for using the calculator The second worksheet tablsquoHow-To-Usersquo provides detailed step-by-step instructions for preparing the baselinedata for the program The third tab lsquoInput-Output rsquo contains the user input eldsand displays the results of the calculations The additional worksheets contain theactual calculations and program documentation Although users are welcome toreview these sheets an experienced user need only access the lsquoInput-Outputrsquo
spreadsheetTo begin a calculation the user enters the following information which is available
for each photo in the NASA Astronaut Photography Database (1) SN geographicalcoordinates of spacecraft nadir position at the time of photo (2) H spacecraftaltitude (3) PC the geographical coordinates of the centre of the photo (4) f nominal focal length and (5) d image format size The automatic implementationof the workbook on the web will automatically enter these values and completecalculations
For more accurate results the user may optionally enter the geographic co-ordinates and orientation for an auxiliary point on the photo which resolves thecamerarsquos rotation uncertainty about the optical axis The auxiliary point data must
Astronaut photography as digital data for remote sensing 4425
be computed by the user following the instruction contained in the lsquoHow-To-Usersquotab of the spreadsheet
After entering the input data the geographic coordinates of the photo footprint(ie four photo corner points four points at the bisector of each edge and the centreof the photo) are immediately displayed below the input elds along with any errormessages generated by the user input or by the calculations ( gure 7) Although
results are computed and displayed they should not be used when error messagesare produced by the program The program also computes the tilt angle for each ofthe photo vectors relative to the spacecraft nadir vector To the right of the photofootprint coordinates is displayed the arc distance along the surface of the sphere
between adjacent computed points
532 Calculation assumptions
The mathematical calculations implemented in the Low Oblique Space PhotoFootprint Calculator use the following assumptions
1 The SN location is used as exact even though the true value may vary by upto plusmn01 degree from the location provided with the photo
2 The spacecraft altitude is used as exact Although the determination of the
nadir point at the instant of a known spacecraft vector is relatively precise(plusmn115times10 Otilde 4 degrees) the propagator interpolates between sets of approxi-mately 10ndash40 known vectors per day and the time code recorded on the lmcan drift Thus the true value for SN may vary by up to plusmn01 degree fromthe value provided with the photo
3 The perspective centre of the camera is assumed to be at the given altitudeover the speci ed spacecraft nadir location at the time of photo exposure
4 The PC location is used as exact even though the true value may vary by upto plusmn05deg latitude and plusmn05deg longitude from the location provided with the
photo5 A spherical earth model is used with a nominal radius of 6 372 16154 m (a
common rst-order approximation for a spherical earth used in geodeticcomputations)
6 The nominal lens focal length of the camera lens is used in the computations(calibrated focal length values are not available)
7 The photo projection is based on the classic pin-hole camera model8 No correction for lens distortion or atmospheric re ection is made
9 If no auxiliary point data is provided the lsquoTop of the Imagersquo is orientedperpendicular to the vector from SN towards PC
533 T ransformation from Earth to photo coordinate systemsThe calculations begin by converting the geographic coordinates ( latitude and
longitude) of the SN and PC to a Rectangular Earth-Centred Coordinate System(R-Earth) de ned as shown in gure 9 (with the centre of the Earth at (0 0 0))
Using the vector from the Earthrsquos centre through SN and the spacecraft altitude thespacecraft location (SC) is also computed in R-Earth
For ease of computation a Rectangular Spacecraft-Centred Coordinate System(R-Spacecraft ) is de ned as shown in gure 9 The origin of R-Spacecraft is located
at SC with its +Z-axis aligned with vector from the centre of the Earth throughSN and its +X-axis aligned with the vector from SN to PC ( gure 9) The speci c
J A Robinson et al4426
Figure 9 Representation of the area included in a photograph as a geometric projectiononto the Earthrsquos surface Note that the focal length (the distance from the SpaceShuttle to the photo) is grossly overscale compared to the altitude (the distance fromthe Shuttle to the surface of the Earth
rotations and translation used to convert from the R-Earth to the R-Spacecraft arecomputed and documented in the spreadsheet
With the mathematical positions of the SC SN PC and the centre of the Earthcomputed in R-Spacecraft the program next computes the location of the camerarsquosprincipal point (PP) The principle point is the point of intersection of the opticalaxis of the lens with the image plane (ie the lm) It is nominally positioned at a
Astronaut photography as digital data for remote sensing 4427
distance equal to the focal length from the perspective centre of the camera (whichis assumed to be at SC) along the vector from PC through SC as shown in gure 9
A third coordinate system the Rectangular Photo Coordinate System (R-Photo)is created with its origin at PP its XndashY axial plane normal to the vector from PCthrough SC and its +X axis aligned with the +X axis of R-Spacecraft as shownin gure 9 The XndashY plane of this coordinate system represents the image plane ofthe photograph
534 Auxiliary point calculationsThe calculations above employ a critical assumption that all photos are taken
with the lsquotop of the imagersquo oriented perpendicular to the vector from the SN towardsPC as shown in gure 9 To avoid non-uniform solar heating of the external surfacemost orbiting spacecraft are slowly and continually rotated about one or more axesIn this condition a ight crew member taking a photo while oating in microgravitycould orient the photo with practically any orientation relative to the horizon (seealso gure 8(b)) Unfortunately since these photos are taken with conventionalhandheld cameras there is no other information available which can be used toresolve the photorsquos rotational ambiguity about the optical axis other then the photoitself This is why the above assumption is used and the footprint computed by thiscalculator is actually a lsquolocator ellipsersquo which estimates the area on the ground whatis likely to be included in a speci c photograph (see gure 6) This locator ellipse ismost accurate for square image formats and subject to additional distortion as thephotograph format becomes more rectangular
If the user wants a more precise calculation of the image footprint the photorsquosrotational ambiguity about the optical axis must be resolved This can be done inthe calculator by adding data for an auxiliary point Detailed instructions regardinghow to prepare and use auxiliary point data in the computations are included in thelsquoHow-To-Usersquo tab of the spreadsheet Basically the user determines which side ofthe photograph is top and then measures the angle between the line from PP to thetop of the photo and from PP to the auxiliary point on the photo ( gure 9)
If the user includes data for an auxiliary point a series of computations arecompleted to resolve the photo rotation ambiguity about the optical axis (ie the
+Z axis in R-Photo) A vector from the Auxiliary Point on the Earth (AE) throughthe photograph perspective centre ( located at SC) is intersected with the photo imageplane (XndashY plane of R-Photo) to compute the coordinates of the Auxiliary Point onthe photo (AP) in R-Photo A two-dimensional angle in the XndashY plane of R-Photofrom the shy X axis to a line from PP to AP is calculated as shown in gure 9 The
shy X axis is used as the origin of the angle since it represents the top of the photoonce it passes through the perspective centre The diVerence between the computedangle and the angle measured by the user on the photo resolves the ambiguity inthe rotation of the photo relative to the principal line ( gures 7 and 9) Thetransformations from R-Spacecraft and R-Photo are then modi ed to include anadditional rotation angle about the +Z axis in R-Photo
535 lsquoFootprint rsquo calculationsThe program next computes the coordinates of eight points about the perimeter
of the image format (ie located at the four photo corners plus a bisector pointalong each edge of the image) These points are identi ed in R-Photo based uponthe photograph format size and then converted to R-Spacecraft Since all computa-tions are done in orthogonal coordinate systems the R-Spacecraft to R-Photo
J A Robinson et al4428
rotation matrix is transposed to produce an R-Photo to R-Spacecraft rotation matrixOnce in R-Spacecraft a unit vector from each of the eight perimeter points throughthe photo perspective centre (the same point as SC) is computed This provides thecoordinates for points about the perimeter of the image format with their directionvectors in a common coordinate system with other key points needed to computethe photo footprint
The next step is to compute the point of intersection between the spherical earthmodel and each of the eight perimeter point vectors The scalar value for eachperimeter point unit vector is computed using two-dimensional planar trigonometryAn angle c is computed using the formula for the cosine of an angle between twoincluded vectors (the perimeter point unit vector and the vector from SC to thecentre of the Earth) Angle y is computed using the Law of Sines Angle e=180degrees shy c shy y The scalar value of the perimeter point vector is computed using eand the Law of Cosines The scalar value is then multiplied by the perimeter pointunit vector to produce the three-dimensional point of intersection of the vector withEarthrsquos surface in R-Spacecraft The process is repeated independently for each ofthe eight perimeter point vectors Aside from its mathematical simplicity the valueof arcsine (computed in step 2) will exceed the normal range for the sin of an anglewhen the perimeter point vectors fail to intersect with the surface of the earth Asimple test based on this principle allows the program to correctly handle obliquephotos which image a portion of the horizon (see results for high oblique photographsin table 5)
The nal step in the process converts the eight earth intersection points from theR-Spacecraft to R-Earth The results are then converted to the standard geographiccoordinate system and displayed on the lsquoInput-Outputrsquo page of the spreadsheet
54 ExamplesAll three formulations were applied to the photographs included in this paper
and the results are compared in table 5 Formulation 1 gives too small a value fordistance across the photograph (D) for all but the most nadir shots and thus servesas an indicator of the best theoretical case but is not a good measure for a speci cphotograph For example the photograph of Lake Eyre taken from 276 km altitude( gure 2(a)) and the photograph of Limmen Bight ( gure 7) were closest to beingnadir views (oVsetslt68 km or tlt15deg table 5) For these photographs D calculatedusing Formulation 1 was similar to the minimum D calculated using Formulations2 and 3 For almost all other more oblique photographs Formulation 1 gave asigni cant underestimate of the distance covered in the photograph For gure 5(A)(the picture of Houston taken with a 40 mm lens) Formulation 1 did not give anunderestimate for D This is because Formulation 1 does not account for curvatureof the Earth in any way With this large eld of view assuming a at Earth in atedthe value of D above the minimum from calculations that included Earth curvature
A major diVerence between Formulations 2 and 3 is the ability to estimate pixelsizes (P) in both directions (along the principal line and perpendicular to the principalline) For the more oblique photographs the vertical estimate of D and pixel sizes ismuch larger than in the horizontal direction (eg the low oblique and high obliquephotographs of Hawaii gure 4 table 5)
For the area of Limmen Bight ( gure 7) table 5 illustrates the improvement inthe estimate of distance and pixel size that can be obtained by re-estimating thelocation of the PC with greater accuracy Centre points in the catalogued data are
Astronaut photography as digital data for remote sensing 4429
plusmn05deg of latitude and longitude When the centre point was re-estimated to plusmn002degwe determined that the photograph was not taken as obliquely as rst thought(change in the estimate of the look angle t from 169 to 128deg table 5) When theauxiliary point was added to the calculations of Formula 3 the calculated look angleshrank further to 110deg indicating that this photograph was taken at very close toa nadir view Of course this improvement in accuracy could have also led to estimatesof greater obliquity and corresponding larger pixel sizes
We also tested the performance of the scale calculator with auxiliary point byestimating the corner and point locations on the photograph using a 11 000 000Operational Navigational Chart For this test we estimated our ability to readcoordinates from the map as plusmn002degand our error in nding the locations of thecorner points as plusmn015deg (this error varies among photographs depending on thedetail that can be matched between photo and map) For Limmen Bight ( gure 7)the mean diVerence between map estimates and calculator estimates for four pointswas 031deg (SD=018 n=8) For a photograph of San Francisco Bay (STS062-151-291) the mean diVerence between map estimates and calculator estimates forfour points was 0064deg (SD=018 n=8) For a photograph of San Francisco Bay(STS062-151-291) the mean diVerence between map estimates and calculator estim-ates for eight points was 0196deg (SD=0146 n=16) Thus in one case the calculatorestimates were better than our estimate of the error in locating corner points on themap It is reasonable to expect that for nadir-viewing photographs the calculatorused with an auxiliary point can estimate locations of the edges of a photograph towithin plusmn03deg
55 Empirical con rmation of spatial resolution estimatesAs stated previously a challenge to estimating system-AWAR for astronaut
photography of Earth is the lack of suitable targets Small features in an image cansometimes be used as a check on the size of objects that can be successfully resolvedgiving an approximate value for GRD Similarly the number of pixels that make upthose features in the digitized image can be used to make an independent calculationof pixel size We have successfully used features such as roads and airport runwaysto make estimates of spatial scale and resolution (eg Robinson et al 2000c) Whilerecognizing that the use of linear detail in an image is a poor approximation to abar target and that linear objects smaller than the resolving power can often bedetected (Charman 1965) few objects other than roads could be found to make anydirect estimates of GRD Thus roads and runways were used in the images ofHouston (where one can readily conduct ground veri cations and where a numberof higher-contrast concrete roadways were available) to obtain empirical estimatesof GRD and pixel size for comparison with table 5
In the all-digital ESC image of Houston ( gure 3(d )) we examined EllingtonField runway 4-22 (centre left of the image) which is 24384 mtimes457 m This runwayis approximately 6ndash7 pixels in width and 304ndash3094 pixels in length so pixelsrepresent an distance on the ground 7ndash8 m Using a lower contrast measure of astreet length between two intersections (2125 m=211 pixels) a pixel width of 101 mis estimated These results compare favourably with the minimum estimate of 81 mpixels using Formulation 1 (table 5) For an estimate of GRD that would be morecomparable to aerial photography of a line target the smallest street without treecover that could be clearly distinguished on the photograph was 792 m wide Thesmallest non-street object (a gap between stages of the Saturn rocket on display in
J A Robinson et al4430
Tab
le5
App
lica
tio
nof
met
hod
sfo
res
tim
ati
ng
spati
alre
solu
tio
no
fast
ron
aut
ph
oto
gra
ph
sin
clu
din
gF
orm
ula
tion
s12
and
3Im
pro
ved
cen
tre
po
int
esti
mat
esan
dauxilia
ryp
oin
tca
lcu
lati
ons
are
sho
wn
for
gure
7V
aria
ble
sas
use
din
the
text
fle
ns
foca
lle
ngt
hH
al
titu
de
PC
ph
oto
gra
ph
cen
tre
poin
tSN
sp
ace
craft
nadir
po
int
D
dis
tance
cover
edo
nth
egr
oun
d
Pd
ista
nce
on
the
grou
nd
repre
sen
ted
by
apix
el
OVse
td
ista
nce
bet
wee
nP
Cand
SNusi
ng
the
grea
tci
rcle
form
ula
t
look
angle
away
from
SN
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
2L
ake
Eyre
ST
S093
-717
-80
250
276
shy29
0137
0shy
28
413
71
607
11
767
413
7609
ndash64
212
013
760
9ndash64
7121
124
ST
S082
-754
-61
250
617
shy28
5137
0shy
28
113
50
135
726
1201
18
013
8ndash14
827
517
9138ndash15
3277
294
Fig
ure
3H
ou
sto
nS
TS
062
-97-
151
4029
829
5shy
95
530
5shy
95
4410
78
8112
20
5348ndash58
990
1205
351
ndash63
8951
10
36
ST
S094
-737
-28
100
294
295
shy
95
528
3shy
95
5162
31
1133
24
415
8ndash20
334
724
3158ndash20
7351
390
ST
S055
-71-
41250
302
295
shy
95
028
2shy
93
766
412
8192
32
5736
ndash84
715
232
273
9ndash85
7154
185
S73
E51
45300
2685
295
shy
95
024
78
0F
igu
re4
Haw
aii
ST
S069
-701
-65
250
370
195
shy
155
520
4shy
153
881
415
7204
28
9876
ndash98
917
928
688
0ndash10
9181
210
ST
S069
-701
-70
250
370
210
shy
157
519
2shy
150
781
415
7738
63
414
9ndash23
336
760
7148ndash55
6394
10
63
ST
S069
-729
-27
100
398
200
shy
156
620
5shy
148
2219
42
1867
65
432
8ndash13
10
158
62
4323+
311
N
A
Astronaut photography as digital data for remote sensing 4431
Tab
le5
(Cont
inued
)
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
7L
imm
enB
igh
tS
TS
093
-704
-50
250
283
shy15
0135
0shy
15
013
58
623
120
859
16
963
0ndash67
3125
16
863
0ndash67
612
613
2P
CR
e-es
tim
ated
shy14
733
1352
67
64
5128
623
ndash65
5123
128
623
ndash65
712
3127
Auxilia
ryP
oin
t611
062
3ndash65
112
312
5F
igu
re10
N
ear
Au
stin
T
XS
TS
095
-716
-46
250
543
30
5shy
975
28
6shy
1007
120
230
375
34
613
5ndash157
281
34
1136ndash16
128
636
0
1 All
pho
togr
aphs
exce
pt
on
eta
ken
wit
hH
ass
elbla
dca
mer
aso
dis
tan
ceacr
oss
the
image
d=
55m
m
for
S73
E51
45
taken
wit
hel
ectr
onic
still
cam
erad=
276
5m
mho
rizo
nta
l2 P
Cand
SN
are
giv
enin
the
form
(lati
tud
elo
ngit
ude)
deg
rees
w
ith
neg
ativ
en
um
ber
sre
pre
senti
ng
south
and
wes
tA
ccura
cyo
fP
Cis
plusmn0
5and
SN
isplusmn
01
as
reco
rded
inth
eA
stro
naut
Ph
oto
gra
phy
Data
base
ex
cep
tw
her
en
ote
dth
atce
ntr
ep
oin
tw
asre
-est
imate
dw
ith
grea
ter
accu
racy
3 C
alc
ula
ted
usi
ng
the
mea
nofth
em
inim
um
an
dm
axim
um
esti
mate
sof
DF
orm
ula
tion
2co
nsi
der
sD
inth
ed
irec
tion
per
pen
dic
ula
rto
the
Pri
nci
pal
Lin
eo
nly
4 C
alc
ula
ted
inho
rizo
nta
lan
dve
rtic
aldir
ecti
on
sb
ut
still
ass
um
ing
geom
etry
of
gure
6(B
)F
irst
nu
mb
eris
the
mea
no
fth
em
inim
um
Dat
the
bott
om
of
the
ph
oto
and
the
maxi
mum
Dat
the
top
of
the
ph
oto
(range
show
nin
the
tab
le)
and
isco
mpara
ble
toF
orm
ula
tio
n2
Sec
ond
num
ber
isth
em
ean
of
Des
tim
ated
alon
gth
eP
rin
cip
alline
and
alo
ng
the
ou
tsid
eed
ge
(range
not
show
n)
5 Alt
itu
de
isap
pro
xim
ate
for
mis
sion
as
exac
tti
me
pho
togr
ap
hw
asta
ken
and
corr
esp
on
din
gn
adir
data
are
no
tava
ilab
le
Lack
of
nad
irdata
pre
ven
tsap
plica
tion
of
Form
ula
tion
s2
an
d3
6 Au
xilliar
ypo
int
add
edw
as
shy14
80
lati
tud
e13
51
2lo
ngit
ude
shy46
5degro
tati
on
angle
in
stru
ctio
ns
for
esti
mati
ng
the
rota
tio
nan
gle
para
met
erare
conta
ined
as
par
tof
the
scal
eca
lcu
lato
rsp
read
shee
tdo
cum
enta
tio
n
J A Robinson et al4432
a park at Johnson Space Center) that could clearly be distinguished on the photo-graph was 853 m wide
For the photograph of Houston taken with a 250 mm lens ( gure 3(C)) anddigitized from second generation lm at 2400 ppi (106 mm pixel Otilde 1 ) Ellington Fieldrunway 4-22 is 3 pixels in width and 1613 pixels in length so pixels represent151ndash152 m on the ground These results compare favourably with the minimumestimate of 152 m using Formulation 2 and 154ndash185 m pixels using Formulation 3(table 5) For an estimate of GRD using an 8times8 inch print (1369 enlargement) and4times magni cation the smallest street that could clearly be distinguished was 822 mwide the same feature could barely be distinguished on the digitized image
We also made an empirical estimate of spatial resolution for lower contrastvegetation boundaries By clearing forest so that a pattern would be visible to landingaircraft a landowner outside Austin Texas (see also aerial photo in Lisheron 2000)created a target that is also useful for evaluating spatial resolution of astronautphotographs The forest was selectively cleared in order to spell the landownerrsquosname lsquoLUECKErsquo with the remaining trees ( gure 10) According to local surveyorswho planned the clearing the plan was to create letters that were 3100 fttimes1700 ft(9449 mtimes5182 m) Photographed at a high altitude relative to most Shuttle missions(543 km) with a 250 mm lens Formula 3 predicts that each pixel would represent anarea 286 mtimes360 m on the ground (table 5) When original lm was digitized at2400 ppi (106 mm pixel Otilde 1 ) letters correspond to 294times188 pixels for a comparablepixel size of 27ndash32 m
6 Summary and conclusionsAstronaut photographs can be an excellent source of data for remote sensing
applications Best-case resolutions are similar to that for Landsat or SPOT withpixels as small as lt10 m It was estimated that of images taken to date 504 hadlens and obliquity characteristics that make them potential candidates for remotesensing information Digitized astronaut photographs can be overlaid with othersatellite data using GIS (Eckardt et al 2000) or used to ll in gaps in time serieswhen other imagery is not available As a source of public-domain information theycan be very useful for scientists who do not have access to satellite imagery eitherbecause of the expense of image acquisition (to the end user) or the computersystems needed for processing satellite images These diVerences in image acquisitioncosts (to the user) are summarized in table 6
Searching of the complete database of NASA astronaut photography includinglow-spatial-resolution browse images is available via the Web (OYce of EarthSciences 2000) This provides nearly global access for identifying images that willcontribute to a speci c scienti c project For the most detailed studies digitalproducts posted to the web will not be of suYcient quality To date we have madeit a practice of digitizing small numbers of images when requested by scientists atno charge Such requests (including a description of the project involved) can bemade through the authors or using the contact information listed on the websiteInvestigators with needs for larger numbers of digital images have also been servedthrough collaborative agreements
In our experience scientists that nd the data most valuable are those in develop-ing countries those studying areas that have not usually been targets for the majorsatellites those having diYculty in nding low-cloud images those interested inconstructing time series or those interested in using a large number of images
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
Astronaut photography as digital data for remote sensing 4405
determining the suitability of an image for a remote sensing objective In remotesensing literature spatial resolution for aerial photography is often treated verydiVerently from spatial resolution of multispectral scanning sensors To providesuYcient background for a discussion of spatial resolution for a data source thatshares properties with both aerial photography and satellite remote sensing we rst summarize the ways that spatial resolution is typically quanti ed for aerialphotography and satellite remote sensors
Given this background information we (1 ) describe and illustrate factors thatin uence the spatial resolution of astronaut photographs (2) show examples ofestimating spatial resolution for a variety of photographs in a way that willenable users to make these calculations whether or not they have a background inphotogrammetry and (3) compare spatial resolution of digital data extracted fromastronaut photographs to data obtained from other satellites
2 BackgroundmdashSpatial resolution of imaging systemsSpatial resolution is a fundamental property of any imaging system used to
collect remote sensing data and directly determines the spatial scale of the resultantinformation Resolution seems intuitively obvious but its technical de nition andprecise application in remote sensing have been complex For example Townshend(1980) summarized 13 diVerent ways to estimate the resolving power of the LandsatMultiSpectral Scanner (MSS) Simonett (1983 p 20) stated lsquoIn the simplest casespatial resolution may be de ned as the minimum distance between two objects thata sensor can record distinctly[but] it is the format of the sensor system thatdetermines how spatial resolution is measuredrsquo By focusing attention on the proper-ties of the system (and not the images acquired) this de nition can be applied to avariety of types of images provided by diVerent systems In order to provide ascomplete a treatment as possible background is provided on several views of spatialresolution that are relevant when considering the spatial resolution of astronautphotography
21 Ground resolved distancePhotogrammetrists were measuring spatial resolution of aerial photographs using
empirical methods long before the rst scanning sensor was placed in orbit Thesemethods integrated properties of the imaging system with the other external factorsthat determine spatial resolution Ground resolved distance (GRD) is the parameterof most interest because it measures the applicability of an image to a speci c taskThe GRD of an image is de ned as the dimensions of the smallest discernible objectThe GRD is a function of geometry (altitude focal length of optics) equipment(internal system spatial resolution of the camera or scanner) and also on re ectancecharacteristics of the object compared to its surroundings (contrast)
Performance of a lm lm and camera or deployed aerial photography systemis measured empirically using standard targets that consist of black-and-white barsof graduated widths and spacings ( gure 6ndash9 in Slater et al 1983) The area-weightedaverage resolution (AWAR) in lp mm Otilde 1 ( line pairs mm Otilde 1 ) at the lm plane is deter-mined by measuring the smallest set of line pairs that can be discriminated on anoriginal lm negative or transparency Line pairs are quoted because it is necessaryto discriminate between one object and another to detect it and measure it
Film resolving power (in lp mm Otilde 1 ) is measured by manufacturers under standardphotographi c conditions (Smith and Anson 1968 Eastman Kodak Company 1998)
J A Robinson et al4406
at high contrast (objectbackground ratio 10001) and low contrast (objectback-ground ratio 161) Most terrestrial surfaces recorded from orbit are low contrastfor the purpose of estimating resolving power of lm Kodak no longer measures lm resolving power for non-aerial photographic lms (Karen Teitelbaum EastmanKodak Company personal communication) including lms that NASA routinelyuses for Earth photography
AWAR can be measured for the static case of lm and camera or for thecamerandashaircraft system in motion For example AWAR for the National AerialPhotography Program includes eVects of the lens resolving power of original lmimage blur due to aircraft motion and spatial resolution of duplicating lm (Light1996) Given AWAR for a system in motion the GRD can be calculated bytrigonometry (see equation (3) in sect5 d=1AWAR and D=GRD)
An impediment to similar rigorous measurement of GRD for orbital remotesensing systems is the lack of a target of suitable scale on the ground thus spatialresolution for most orbiting sensors is described in terms of a less all-encompassingmeasure instantaneous eld of view (see below) Additional challenges to measuringAWAR for a complete astronaut photography system include the number of diVerentoptions for aspects of the system including diVerent cameras lms and orbitalaltitudes These elements that must be standardized to determine AWAR providea useful list of those characteristics of astronaut photography that will mostin uence GRD
22 Instantaneous eld of viewThe instantaneous eld of view (IFOV) is generally used to represent the spatial
resolution of automated satellite systems and is commonly used interchangeablywith the term spatial resolution when comparing diVerent sensors IFOV is a combina-tion of geometric mechanical and electronic properties of the imaging systemGeometric properties including satellite orbital altitude detector size and the focallength of the optical system (Simonett 1983) The sensitivity of each detector elementat the wavelength desired plus the signal-to-noise level desired are electronic proper-ties that determine a minimum time for energy absorption For a linear sensor array(pushbroom scanning) this minimum time is translated to areal coverage by theforward velocity of the platform (Campbell 1996 p 97) Usually each detectorelement in the array corresponds to a pixel in the image Thus for a given altitudethe width of the pixel is determined by the optics and sensor size and the height ofthe pixel is determined by the rate of forward motion When magni ed by the ratioof the sensor altitude to the focal length of the optics of the sensor system IFOV isthe size of the area on the ground represented by an individual detector element(pixel Slater 1980 p 27)
The equation of IFOV with spatial resolution can be misleading because IFOVdoes not include factors other than geometrymdashfactors that largely determine thelevel of detail that can be distinguished in an image For example IFOV does notinclude characteristics of the target (contrast with surroundings shape of an objectcolour) atmospheric conditions illumination and characteristics of the interpreter(machine or human) Factors in uencing spatial resolution that are included in IFOVcan be calculated from design speci cations before the system has been built whereasthe actual spatial resolution of an image captured by the sensor will be unique tothat image IFOV represents the best spatial resolution possible given optimalconditions
Astronaut photography as digital data for remote sensing 4407
23 Imagery from scanners versus lm from camerasLike aerial photography systems the GRD of early remote sensing scanners
(electro-optical imaging systems) was calibrated empirically ( gures 1ndash6 in Simonett1983) but such methods are impractical for routine applications Satellites are nowcommon platforms for collecting remote sensing data small-scale aerial photographsare widely available within the USA and Europe and astronaut photographshave become available worldwide Straightforward comparison of resolutions ofphotographs to IFOVs of scanner images is necessary for many applications
How can spatial resolutions of data from diVerent sources be compared usingcommon units Oft-quoted lsquoresolution numbersrsquo actually IFOV of digital scanningsystems should be at least doubled for comparison to the standard method ofexpressing photographi c spatial resolution as GRD simply because of the objectcontrast requirement The following rule of thumb has been used in geographicapplications to compare spatial resolution of scanners and aerial photography (Welch1982 Jensen 1983) a low-contrast target may be converted to an approximate IFOVby dividing the GRD (of the photograph) by 24 Such a rule of thumb might alsobe reasonably applied to astronaut photography making it possible to convertobserved GRD to IFOV and IFOV to an estimated GRD
In this paper we discuss the spatial resolution of astronaut photographs in termsof system properties and compute the area on the ground represented by a singlepixel the equivalent to IFOV To complete our treatment of spatial resolution wealso provide estimates of the sizes of small features identi able in images for severalcases where this can be readily done
3 Factors that determine the footprint (area covered by the photograph )The most fundamental metric that forms the basis for estimating spatial resolution
of astronaut photographs is the size of the footprint or area on the ground capturedin a photograph (see review of satellite photogrammetry by Light 1980) The basicgeometric variables that in uence the area covered by an astronaut photograph are(1) the altitude of the orbit H (2) the focal length of the lens f (3) the actual sizeof the image on the lm d and (4) the orientation of the camera axis relative to theground (the obliquity or look angle t ) The relationships among these parametersare illustrated in gure 1 (also see equation (3) in sect5)
31 AltitudeHuman space ight missions have had a variety of primary objectives that required
diVerent orbital altitudes The higher the altitude the larger the footprint of thephotographs The diVering scale of photographs taken at diVerent altitudes is illus-trated in gure 2 The two photographs of Lake Eyre Australia were taken with thesame camera and lens but on diVerent dates and from diVerent altitudes In (a) thelake is relatively ooded while in (b) it is dry A 23timesdiVerence in altitude leads toa corresponding diVerence in the scales of the resulting photographs
32 L ensesThe longer the lens focal length the more magni cation the greater detail and
the smaller footprint A variety of lenses with diVerent focal lengths are own oneach space mission (table 1) The eVect of lens length on spatial coverage and imagedetail is shown in gure 3 In these views of Houston (a)ndash(c) taken from approxi-mately similar altitudes with the same camera most of the diVerence in scale of the
J A Robinson et al4408
Figure 1 Geometric relationship between look angle (t) spacecraft nadir point (SN) photo-graph centre point (PC) and photograph principal point (PP) The dark shaded arearepresents Earthrsquos surface The distance covered on the ground D corresponds totable 5 Original image size d is listed for various lms in table 1 Variables correspondto equations (2)ndash(9) in sect5
Astronaut photography as digital data for remote sensing 4409
(a) (b)
Figure 2 Example of the eVect of altitude on area covered in a photograph Both photographsof Lake Eyre Australia were taken using a Hasselblad camera with 100 mm lens(a) altitude 283 km STS093-702-062 (b) altitude 617 km STS082-754-61
photographs is due to the diVerent magni cations of the lenses Although taken froma diVerent altitude and using a diVerent camera with a diVerent original image size(see table 2) an electronic still camera image taken with a 300 mm lens is alsoincluded for comparison
For remote sensing longer focal length lenses are generally preferred (250 mm or350 mm for Hasselblad 300 mm or 400 mm lenses for 35 mm format cameras andESCs) Unfortunately longer-focal-length lenses exhibit poorer performance towardthe edge of a frame For example a 250 mm lens (Distagon CF 56) used on theHasselblad camera has spatial resolution of 57 lp mm Otilde 1 at the centre but only51 lp mm Otilde 1 at the edge and 46 lp mm Otilde 1 at the corner (tested with Ektachrome 5017f8 aperture and high contrast) A longer 300 mm lens used on the Nikon camerahas a greater spatial resolution diVerence between centre (82 lp mm Otilde 1 ) and corner(49 lp mm Otilde 1 ) using Kodak 5017 Ektachrome f4 aperture and high contrast FredPearce (unpublished data) There are tradeoVs among lens optics and speed Forexample the lenses for the Linhof system and the 250 mm lens for the Hasselblad(Distagon CF 56 see footnotes to table 1) are limited to apertures smaller thanf56 This becomes an important constraint in selecting shutter speeds (see discussionof shutter speed in sect42)
33 Cameras and actual image sizesAfter passing through the lens the photographic image is projected onto lm
inside the camera The size of this original image is another important propertydetermining spatial resolution and is determined by the camera used Cameraformats include 35 mm and 70 mm (Lowman 1980 Amsbury 1989) and occasionally5times4 inch (127times100 mm) and larger (table 1) Cameras own on each mission arenot metricmdashthey lack vacuum platens or reseau grids image-motion compensationor gyro-stabilized mounts The workhorse for engineering and Earth photographyon NASA missions has been a series of 70 mm Hasselblad cameras (table 1) chosenfor their reliability The modi ed magazine databack imprints a unique number and
J A Robinson et al4410
Tab
le1
Det
ails
on
the
cam
eras
use
dfo
rast
ron
aut
ph
oto
gra
phy
of
Eart
h
Data
incl
ud
ep
ho
togr
aphy
from
all
NA
SA
hu
man
spac
eig
ht
mis
sion
sth
rough
ST
S-9
6(J
un
e19
99
378
461
tota
lfr
am
esin
the
dat
abas
e)1
Imag
esi
zeT
ypic
al
lense
sN
um
ber
of
Cam
era
make
(mm
timesm
m)
use
d(m
m)
ph
oto
gra
ph
sP
erio
dof
use
No
tes
Has
selb
lad
55times
55
40
50100
2502
288
494
1963ndashp
rese
nt
Lin
ho
f10
5times120
90
250
29
373
1984ndashp
rese
nt
No
won
lyo
wn
occ
asio
nal
lyM
ult
ispec
tral
Cam
era
(S19
0A
)57times
57152
32
913
1973ndash19
74S
kyla
b
xed
-mou
nt3
cam
era
(NA
SA
1974
)E
art
hT
erra
inC
amer
a(S
190B
)11
5times11
5457
5478
1973ndash19
74S
kyla
b
xed
-mou
nt3
cam
era
(NA
SA
1974
)R
ole
iex
55times
55
50
100250
4352
1990ndash19
92L
arg
eF
orm
at
Cam
era4
229
times457
305
2143
1984
Mis
sio
nST
Sndash41G
only
(Mer
kel
etal
198
5)
Maure
r55
times55
76
80818
1961ndash19
68
1 Sm
all-f
orm
atca
mer
as
(35
mm
lm
and
ES
C)
are
no
tin
clud
edin
this
table
bec
ause
of
the
larg
enu
mb
erof
len
ses
and
con
gu
rati
on
sth
at
have
bee
nu
sed
and
bec
au
seth
ese
data
are
no
tcu
rren
tly
incl
uded
inth
edata
bas
efo
rall
mis
sion
s2 L
ense
suse
dfo
rH
ass
elbla
dm
anufa
cture
dby
Carl
Zei
ss
Dis
tago
nC
F4
(40
mm
)D
ista
gon
CF
4(5
0m
m)
Dis
tagon
CF
35
(100
mm
)D
ista
go
nC
F5
6(2
50
mm
)S
tand
ard
Has
selb
lad
len
ses
ow
nw
ill
chan
ge
inth
eyea
r20
00
toD
ista
gon
FE
28
(50
mm
)D
ista
go
nF
E2
(110
mm
)T
ele-
Tes
sar
FE
4(2
50m
m)
and
Tel
e-T
essa
rF
E4
(350
mm
)3 T
hes
eca
mer
as
wer
ein
xed
mou
nti
ngs
on
the
outs
ide
of
the
Skyla
bsp
ace
stati
on
A
ltho
ugh
not
hand
hel
d
the
lm
isca
talo
gued
wit
ho
ther
ast
ron
aut
ph
oto
gra
ph
yan
dis
incl
uded
her
efo
rco
mp
lete
nes
s4 A
NA
SA
-des
igned
cam
era
wit
hatt
itud
ere
fere
nce
syst
emfo
rd
eter
min
ing
pre
cise
nadir
poin
ts
Dat
ais
no
tcu
rren
tly
inco
rpora
ted
into
on
lin
ed
atab
ase
bu
tis
cata
loged
by
Mer
kel
etal
(1
985
)
Astronaut photography as digital data for remote sensing 4411
(a) (b)
(c) (d)
Figure 3 Example of the eVect of lens focal length on resolving power Hasselblad imagesof Houston were all taken from an altitude of 294 kmndash302 km with diVerent lenses(a) 40 mm STS062-97-151 (b) 100 mm STS094-737-28 (c) 250 mm STS055-71-41 Forcomparison an Electronic Still Camera image with 300 mm lens at 269 km altitude isalso shown (d ) S73E5145
timestamp on each frame at the time of exposure Cameras are serviced between ights Occasionally there is enough volume and mass allowance so that a Linhof5times4 inch (127times101 mm) format camera can be own Nikon 35 mm cameras are own routinely also because of proven reliability Electronic still cameras (ESC)were tested for Earth photography beginning in 1992 (Lulla and Holland 1993) Inan ESC a CCD (charge-coupled device) is used as a digital replacement for lmrecording the image projected inside the camera An ESC (consisting of a KodakDCS 460c CCD in a Nikon N-90S body) was added as routine equipment forhandheld photographs in 1995 ESCs have also been operated remotely to captureand downlink Earth images through a NASA-sponsored educational programme(EarthKAM) Discussion of CCD spatial array and radiometric sensitivity arebeyond the scope of this paper but are summarized by Robinson et al (2000b) The
J A Robinson et al4412
format of the lm (or CCD) and image size projected onto the lm (or CCD) aresummarized for all the diVerent cameras own in table 2
34 L ook angle or obliquityNo handheld photographs can be considered perfectly nadir they are taken at a
variety of look angles ranging from near vertical ( looking down at approximatelythe nadir position of the spacecraft ) to high oblique (images that include the curvatureof Earth) Imaging at oblique look angles leads to an image where scale degradesaway from nadir A set of views of the same area from diVerent look angles is shownin gure 4 The rst two shots of the island of Hawaii were taken only a few secondsapart and with the same lens The third photograph was taken on a subsequentorbit and with a shorter lens The curvature of the Earth can be seen in the upperleft corner
Obliquity can be described qualitatively ( gure 4) or quantitatively as the lookangle (t gure 1 calculations described in formulation 2 (sect52)) Because obliquityand look angle have such a dramatic in uence on the footprint we summarize thedatabase characteristics relative to these two parameters Figure 5 is a breakdownof the spatial resolution characteristics of low oblique and near vertical photographsin the NASA Astronaut Photography database Number of photographs are grouped(1) by calculated values for look angle (t ) and (2) by altitude After observing theoverlap between near vertical and low oblique classes we are currently restructuringthis variable (lsquotiltrsquo) in the database to provide a measure of t when available Userswill still be able to do searches based on the qualitative measures but these measureswill be more closely tied to actual look angle
341 Obliquity and georeferencing digitised photographsOften the rst step in a remote sensing analysis of a digitized astronaut photo-
graph is to georeference the data and resample it to conform to a known mapprojection Details and recommendations for resampling astronaut photographydata are provided by Robinson et al (2000a 2000c) and a tutorial is also available(McRay et al 2000) Slightly oblique photographs can be geometrically correctedfor remote sensing purposes but extremely oblique photographs are not suited forgeometric correction When obliquity is too great the spatial scale far away fromnadir is much larger than the spatial scale closer to nadir resampling results areunsuitable because pixels near nadir are lost as the image is resampled while manypixels far away from nadir are excessively replicated by resampling
To avoid the generation of excess pixels during georeferencing the pixel sizes ofthe original digitized image should be smaller than the pixels in the nal resampledimage Calculations of original pixel size using methods presented below can beuseful in ensuring meaningful resampling For slightly oblique images formulation 3(sect53) can be used to estimate pixel sizes at various locations in a photograph (nearnadir and away from nadir) and these pixel sizes then used to determine a reasonablepixel scale following resampling
4 Other characteristics that in uence spatial resolutionBeyond basic geometry many other factors internal to the astronaut photography
system impact the observed ground resolved distance in a photograph The lensesand cameras already discussed are imperfect and introduce radiometric degradations(Moik 1980)Vignetting (slight darkening around the edges of an image) occurs in
Astronaut photography as digital data for remote sensing 4413
Tab
le2
Max
imu
mpo
ssib
led
igit
alsp
atia
lre
solu
tio
n(p
ixel
size
inm
)fo
rca
mer
asy
stem
scu
rren
tly
use
dby
ast
ron
auts
top
ho
togr
aph
eart
hb
ase
do
nth
ege
om
etry
of
alt
itud
ele
ns
and
lm
size
C
alc
ula
tion
sas
sum
eth
atco
lour
tran
spar
ency
lm
isdig
itiz
edat
2400
ppi
(pix
els
per
inch
10
6m
mpix
elOtilde
1 )
Min
imu
mgro
un
ddis
tance
repre
sente
dby
1p
ixel
(m)
As
afu
nct
ion
of
alti
tude1
Imag
esi
zeM
inim
um
alt
itud
eM
edia
nalt
itu
de
Maxi
mum
alt
itud
eC
amer
asy
stem
mm
timesm
mp
ixel
s2(2
22k
m)
(326
km
)(6
11
km
)
Lin
ho
f125
mm
90
mm
lens
105
times120
8978times
11
434
261
383
718
250
mm
lens
94
138
259
Has
selb
lad
70
mm
100
mm
lens
55times
55
5198
times519
823
534
564
725
0m
mle
ns
94
138
259
Nik
on
35m
m30
0m
mle
ns
36times
24
3308
times217
478
115
216
400
mm
lens
59
86
162
ESC
35m
m18
0m
mle
ns3
27
6times
18
530
69times
204
311
116
330
630
0m
mle
ns
67
98
184
400
mm
lens
50
74
138
1 Alt
itud
essh
ow
nar
em
inim
um
m
edia
nan
dm
axim
um
thro
ugh
ST
S-8
9(J
anuary
199
8N
=84
mis
sion
s)
2 Act
ual
pix
els
for
ES
C
oth
erw
ise
assu
med
dig
itiz
ati
on
at240
0pp
i(p
ixel
sp
erin
ch)
equ
alto
10
6m
mpix
elOtilde
1 3 T
he
mo
stco
mm
on
con
gura
tion
for
Eart
hph
oto
gra
ph
yas
part
of
the
Eart
hK
AM
pro
ject
J A Robinson et al4414
Figure 4 De nitions of look angle for photographs taken from low orbit around EarthThe example photographs of Hawaii were all taken from approximately the samealtitude (370ndash398 km) and with the same (250 mm) lens on a Hasselblad camera(STS069-701-69 STS069-701-70 and STS069-729-27 [100 mm lens])
astronaut photography due to light path properties of the lenses used A lack of atness of the lm at the moment of exposure (because cameras used in orbit donot incorporate a vacuum platen) also introduces slight degradations A numberof additional sources of image degradation that would aVect GRD of astronautphotographs are listed
41 Films and processing411 Colour reversal lms
Most lms used in NASA handheld cameras in orbit have been E-6 processcolour reversal lms (Ektachromes) that have a nominal speed of 64 to 100 ASAalthough many other lms have been own (table 3) Reversal lms are used becausedamage from radiation exposure while outside the atmospheric protection of Earthis more noticeable in negative lms than in positive lms (Slater 1996) The choiceof ASA has been to balance the coarser grains ( lower lm resolving power) of high-speed lms with the fact that slower lms require longer exposures and thus areaVected more by vehicle motion (approximately 73 km s Otilde 1 relative to Earthrsquos surfacefor the Space Shuttle) Extremely fast lms (gt400 ASA) are also more susceptibleto radiation damage in orbit (particularly during long-duration or high-altitudemissions Slater 1996) and have not traditionally been used for Earth photography The manufacturer stock numbers identifying the lm used are available for eachimage in the Astronaut Photography Database (OYce of Earth Sciences 2000)
412 Colour infrared lmsColour infrared lm (CIR) was used during an Earth-orbiting Apollo mission
(Colwell 1971) in the multispectral S-190A and the high-resolution S-190B camera
Astronaut photography as digital data for remote sensing 4415
Figure 5 (a) Distributions of astronaut photographs by look angle oV-nadir (calculated fromcentre point and nadir point using Formulation 2) Data included for all photographsfor which centre and nadir points are known with high oblique look angles (n=55 155) excluded (Total sample shown is 108 293 of 382 563 total records in thedatabase when compiled For records where nadir data was not available number ofobservations for qualitative look angles are near vertical 14 086 low oblique 115 834undetermined 89 195) (b) Altitude distribution for astronaut photographs of Earthtaken with the Hasselblad camera Data included for Hasselblad camera only focallength determined with high oblique look angles excluded (Total sample shown is159 046 of 206 971 Hasselblad records with known altitude and of 382 563 total recordsin the database when compiled For Hasselblad records with known altitude data notshown includes high oblique photographs using 100 mm and 250 mm lenses n=20 262and all look angles with other lenses n=27 663)
systems on Skylab (NASA 1974 Wilmarth et al 1977) and occasionally on Shuttlemissions and Shuttle-Mir missions (table 3) The CIR lm used is a three-layerAerochrome having one layer that is sensitive to re ected solar infrared energy toapproximately 900 nm (Eastman Kodak Company 1998) protective coatings onmost spacecraft windows also limit IR transmittance between 800 mm and 1000 nmThis layer is extremely sensitive to the temperature of the lm which createsunpredictable degradation of the IR signature under some space ight conditions
J A Robinson et al4416
Tab
le3
Det
ails
on
lm
suse
dfo
rast
ron
aut
pho
togr
aphy
of
Eart
h
Dat
ain
clud
ep
ho
togr
aphy
fro
mall
NA
SA
hum
an
spac
eig
ht
mis
sio
ns
thro
ugh
ST
S-9
6(J
un
e19
99
378
461
tota
lfr
ames
inth
edata
base
)F
ilm
sw
ith
lt50
00fr
am
esex
po
sed
and
lm
suse
dfo
r35
mm
cam
eras
only
are
no
tin
clud
ed
Res
olv
ing
Nu
mb
erof
Film
man
ufa
ctu
rer
AS
Ap
ow
er1
fram
esP
erio
dof
use
Cam
eras
Colo
ur
posi
tive
Kod
akA
eroch
rom
eII
(2447
2448
SO
-131S
O-3
68
)32
40ndash50
513
8196
8ndash19
85
Hass
elbla
dL
inh
of
Kod
akE
kta
chro
me
(5017
502
56
017
6118
SO
-121
64
50
113
932
196
5ndash19
68
Hass
elbla
dL
inh
of
Role
iex
SO
-217
Q
X-8
24
QX
-868
)19
81ndash1993
Mau
rer
Nik
on
Kod
akL
um
iere
(5046
5048
)100
105
743
199
4ndash19
98
Hass
elbla
dL
inho
fK
od
akE
lite
100
S(5
069
)100
41
486
199
6ndashp
rese
nt
Hass
elbla
d2
Fu
jiV
elvia
50C
S135
-36
32
13
882
1992ndash19
97H
ass
elbla
dN
iko
nK
od
akH
i-R
esolu
tio
nC
olo
r(S
O-2
42S
O-3
56
)10
096
74
1973ndash19
75H
ass
elbla
d
S19
0A
S190
B
Infr
are
dand
bla
ck-a
nd-w
hit
e
lms
Kod
akA
eroch
rom
eC
olo
rIn
frar
ed40
32
40
704
196
5ndashp
rese
nt2
Hass
elbla
dL
inh
of
Nik
on
S190
A
(8443
34
432443
)S
190B
Kod
akB
ampW
Infr
are
d(2
424
)32
10
553
197
3ndash19
74
S190
AK
od
akP
anato
mic
-XB
ampW
(SO
-022
)63
10
657
197
3ndash19
74
S190
A
1A
tlo
wco
ntr
ast
(ob
ject
back
grou
nd
rati
o16
1)
aspro
vid
edby
Ko
dak
and
list
edby
Sla
ter
etal
(1
983
256
)R
esolv
ing
pow
erw
as
dis
con
tin
ued
fro
mth
ete
chn
ical
test
sper
form
edb
yK
odak
inapp
roxim
atel
y19
93
an
dit
isn
ot
kn
ow
nif
curr
ent
lm
sh
ave
sim
ilar
reso
lvin
gpo
wer
too
lder
ver
sion
so
fsi
milar
lm
s(K
T
eite
lbaum
K
od
akp
erso
nal
com
mu
nic
atio
n)
Th
egra
nu
lari
tym
easu
reno
wpro
vid
edb
yK
od
ak
can
no
tb
ere
adily
con
ver
ted
toa
spat
ial
reso
luti
on
2 The
Lin
hof
was
use
dag
ain
on
ST
S-1
03in
Dec
emb
er19
99(d
ata
not
cata
logued
asof
the
pre
para
tion
of
this
tab
le)
usi
ng
Ko
dak
Elite
100S
lm
3 B
egin
nin
gin
July
1999
2443
colo
ur-
infr
are
d
lmpro
cess
was
chan
ged
from
EA
-5to
E-6
Astronaut photography as digital data for remote sensing 4417
413 Film duplicationThe original lm is archived permanently after producing about 20 duplicate
printing masters (second generation) Duplicate printing masters are disseminatedto regional archives and used to produce products (thirdndashfourth generation) for themedia the public and for scienti c use When prints slides transparencies anddigital products are produced for public distribution they are often colour adjustedto correspond to a more realistic look Digital products from second generationcopies can be requested by scienti c users (contact the authors for information onobtaining such products for a particular project) and these are recommended forremote sensing applications Care should be taken that the digital product acquiredfor remote sensing analysis has not been colour adjusted for presentation purposesBased on qualitative observations there is little visible increase in GRD in the thirdor fourth generation products There is a signi cant degradation in delity of colourin third and fourth generation duplicates because an increase in contrast occurswith every copy
42 Shutter speedsThe impact of camera motion (both due to the photographer and to the motion
of the spacecraft ) on image quality is determined by shutter speedmdash1250 to 1500second have been used because slower speeds record obvious blurring caused bythe rapid motion of the spacecraft relative to the surface of the Earth Generally1250 was used for ISO 64 lms (and slower) and after the switch to ISO 100 lmsa 1500 setting became standard New Hasselblad cameras that have been ownbeginning with STS-92 in October 2000 vary shutter speed using a focal plane shutterfor exposure bracketing (rather than varying aperture) and 1250 1500 and 11000second are used
Across an altitude range of 120ndash300 nautical miles (222ndash555 km) and orbitalinclinations of 285deg and 570deg the median relative ground velocity of orbitingspacecraft is 73 km s Otilde 1 At a nominal shutter speed of 1500 the expected blur dueto motion relative to the ground would be 146 m When cameras are handheld (asopposed to mounted in a bracket) blur can be reduced by physical compensationfor the motion (tracking) by the photographer many photographs show a level ofdetail indicating that blur at this level did not occur (eg gure 3(d )) Thus motionrelative to the ground is not an absolute barrier to ground resolution for handheldphotographs
43 Spacecraft windowsMost of the photographs in the NASA Astronaut Photography Database were
taken through window ports on the spacecraft used The transmittance of the windowport may be aVected by material inhomogeniety of the glass the number of layersof panes coatings used the quality of the surface polish environmentally inducedchanges in window materials (pressure loads thermal gradients) or deposited con-tamination Such degradation of the window cannot be corrected is diVerent foreach window and changes over time See Eppler et al (1996) and Scott (2000) fordiscussion of the spectral transmittance of the fused quartz window that is part ofthe US Laboratory Module (Destiny) of the International Space Station
44 Digitized images from lmWhen lm is scanned digitally the amount of information retained depends on
the spectral information extracted from the lm at the spatial limits of the scanner
J A Robinson et al4418
To date standard digitizing methodologies for astronaut photographs have not beenestablished and lm is digitized on a case-by-case basis using the equipment available
441 Digitizing and spatial resolutionLight (1993 1996) provided equations for determining the digitizing spatial
resolution needed to preserve spatial resolution of aerial photography based on thestatic resolving power (AWAR) for the system For lms where the manufacturer hasprovided data (Eastman Kodak 1998 K Teitelbaum personal communication) the resolving power of lms used for astronaut photography ranges from 32 lp mm Otilde 1to 100 lp mm Otilde 1 at low contrast (objectbackground ratio 161 table 3) The AWARfor the static case of the Hasselblad camera has been measured at high and lowcontrast (using lenses shown in table 1) with a maximum of approximately55 lp mm Otilde 1 (Fred Pearce unpublished data)
Based on the method of Light (1993 1996) the dimension of one spatial resolutionelement for a photograph with maximum static AWAR of 55 lp mm Otilde 1 would be18 mm lp Otilde 1 The acceptable range of spot size to preserve spatial information wouldthen be
18 mm
2 atilde 2aring scan spot size aring
18 mm
2(1)
and 6 mm aring scan spot size aring 9 mm Similarly for a more typical static AWAR of30 lp mm Otilde 1 (33 mm lpOtilde 1 low contrast Fred Pearce unpublished data) 11 mm aring scanspot sizearing 17 mm These scan spot sizes correspond to digitizing resolutions rangingfrom 4233ndash2822 ppi (pixels inch Otilde 1 ) for AWAR of 55 lp mm Otilde 1 and 2309ndash1494 ppi forAWAR of 30 lp mm Otilde 1 Scan spot sizes calculated for astronaut photography arecomparable to those calculated for the National Aerial Photography Program(9ndash13 mm Light 1996)
Widely available scanners that can digitize colour transparency lm currentlyhave a maximum spatial resolution of approximately 2400 ppi (106 mm pixel Otilde 1) Forexample we routinely use an Agfa Arcus II desktop scanner (see also Baltsavias1996) with 2400 ppi digitizing spatial resolution (2400 ppi optical resolution in onedirection and 1200 ppi interpolated to 2400 ppi in the other direction) and 2400 ppiwas used to calculate IFOV equivalents (tables 2 and 4) For some combinations oflens camera and contrast 2400 ppi will capture nearly all of the spatial informationcontained in the lm However for lm with higher resolving power thanEktachrome-64 for better lenses and for higher contrast targets digitizing at 2400 ppiwill not capture all of the spatial information in the lm
Improvements in digitizing technology will only produce real increases in IFOVto the limit of the AWAR of the photography system The incremental increase inspatial information above 3000 ppi (85 mm pixel Otilde 1 ) is not likely to outweigh thecosts of storing large images (Luman et al 1997) At 2400 ppi a Hasselblad frameis approximately 5200times5200 or 27 million pixels (table 2) while the same imagedigitized at 4000 ppi would contain 75 million pixels
Initial studies using astronaut photographs digitized at 2400 ppi (106 mm pixel Otilde 1 Webb et al 2000 Robinson et al 2000c) indicate that some GRD is lost comparedto photographic products Nevertheless the spatial resolution is still comparablewith other widely used data sources (Webb et al 2000) Robinson et al (2000c)found that digitizing at 21 mm pixel Otilde 1 provided information equivalent to 106 mmpixel Otilde 1 for identifying general urban area boundaries for six cities except for a single
Astronaut photography as digital data for remote sensing 4419
photo that required higher spatial resolution digitizing (it had been taken with ashorter lens and thus had less spatial resolution) Part of the equivalence observedin the urban areas study may be attributable to the fact that the atbed scannerused interpolates from 1200 ppi to 2400 ppi in one direction The appropriate digitiz-ing spatial resolution will in part depend on the scale of features of interest to theresearchers with a maximum upper limit set of a scan spot size of approximately6 mm (4233 ppi)
442 Digitizing and spectral resolutionWhen colour lm is digitized there will be loss of spectral resolution The three
lm emulsion layers (red green blue) have relatively distinct spectral responses butare fused together so that digitizers detect one colour signal and must convert itback into three (red green blue) digital channels Digital scanning is also subject tospectral calibration and reproduction errors Studies using digital astronaut photo-graphs to date have used 8 bits per channel However this is another parameterthat can be controlled when the lm is digitized We do not further address spectralresolution of digitally scanned images in this paper
45 External factors that in uence GRDAlthough not discussed in detail in this paper factors external to the spacecraft
and camera system (as listed by Moik (1980) for remote sensing in general) alsoimpact GRD These include atmospheric interference (due to scattering attenuationhaze) variable surface illumination (diVerences in terrain slope and orientation) andchange of terrain re ectance with viewing angle (bidirectional re ectance) For astro-naut photographs variable illumination is particularly important because orbits arenot sun-synchronous Photographs are illuminated by diVerent sun angles and imagesof a given location will have colour intensities that vary widely In addition theviewing angle has an eVect on the degree of object-to-backgroun d contrast andatmospheric interference
5 Estimating spatial resolution of astronaut photographsIn order to use astronaut photographs for digital remote sensing it is important
to be able to calculate the equivalent to an IFOVmdashthe ground area represented bya single pixel in a digitized orbital photograph The obliquity of most photographsmeans that pixel lsquosizesrsquo vary at diVerent places in an image Given a ground distanceD represented by a photograph in each direction (horizontal vertical ) an approxi-mate average pixel width (P the equivalent of IFOV) for the entire image can becalculated as follows
P=D
dS(2)
where D is the projected distance on the ground covered by the image in the samedirection as the pixel is measured d is the actual width of the image on the original lm (table 1) and S is the digitizing spatial resolution
Here we present three mathematical formulations for estimating the size of thefootprint or area on the ground covered by the image Example results from theapplication of all three formulations are given in table 5 The rst and simplestcalculation (formulation 1) gives an idea of the maximum spatial resolution attainableat a given altitude of orbit with a given lm format and a perfectly vertical (nadir)
J A Robinson et al4420
view downward Formulation 2 takes into account obliquity by calculating lookangle from the diVerence between the location on the ground represented at thecentre of the photograph and the nadir location of the spacecraft at the time thephotograph was taken ( gure 1) Formulation 3 describes an alternate solution tothe oblique look angle problem using coordinate-system transformations Thisformulation has been implemented in a documented spreadsheet and is availablefor download (httpeoljscnasagovsseopFootprintCalculatorxls) and is in theprocess of being implemented on a Web-based user interface to the AstronautPhotography Database (OYce of Earth Sciences 2000)
Although formulations 2 and 3 account for obliquity for the purposes of calcula-tion they treat the position of the spacecraft and position of the camera as one Inactuality astronauts are generally holding the cameras by hand (although camerasbracketed in the window are also used) and the selection of window position of theastronaut in the window and rotation of the camera relative to the movement ofthe spacecraft are not known Thus calculations using only the photo centre point(PC) and spacecraft nadir point (SN) give a locator ellipse and not the locations ofthe corners of the photograph A locator ellipse describes an estimated area on theground that is likely to be included in a speci c photograph regardless of the rotationof the lm plane about the camerarsquos optical axis ( gure 6)
Estimating the corner positions of the photo requires additional user input of asingle auxiliary pointmdasha location on the image that has a known location on theground Addition of this auxiliary point is an option available to users of thespreadsheet An example of the results of adding an auxiliary point is shown in gure 7 with comparisons of the various calculations in table 5
51 Formulation 1 Footprint for a nadir viewThe simplest way to estimate footprint size is to use the geometry of camera lens
and spacecraft altitude to calculate the scaling relationship between the image in the
Figure 6 Sketch representation of the locator ellipse an estimated area on the ground thatis likely to be included in a speci c photograph regardless of the rotation of the lmplane about the camerarsquos optical axis
Astronaut photography as digital data for remote sensing 4421
Figure 7 An example of the application of the Low Oblique Space Photo FootprintCalculator The photograph (STS093-704-50) of Limmen Bight Australia was takenfrom an altitude of 283 km using the Hasselblad camera with a 250 mm lens Mapused is 11 000 000 Operational Navigational Chart The distances shown equate toan average pixel size of 124 m for this photograph
lm and the area covered on the ground For a perfect nadir view the scalerelationship from the geometry of similar triangles is
d
D=
f
H(3)
where d=original image size D=distance of footprint on the ground f =focallength of lens and H=altitude Once D is known pixel size ( length=width) can becalculated from equation (2) These calculations represent the minimum footprintand minimum pixel size possible for a given camera system altitude and digitisingspatial resolution (table 2) Formulation 1 was used for calculating minimum pixelsizes shown in tables 2 and 4
By assuming digitizing at 2400 ppi (106 mm pixel Otilde 1 ) currently a spatial resolutioncommonly attainable from multipurpose colour transparency scanners (see sect441)this formulation was used to convert area covered to an IFOV equivalent for missionsof diVerent altitudes (table 2) Table 4 provides a comparison of IFOV of imagesfrom various satellites including the equivalent for astronaut photography Valuesin this table were derived by using formulation 1 to estimate the area covered becausea perfect nadir view represents the best possible spatial resolution and smallest eldof view that could be obtained
52 Formulation 2 Footprint for oblique views using simpli ed geometry and thegreat circle distance
A more realistic approach to determining the footprint of the photographaccounts for the fact that the camera is not usually pointing perfectly down at the
J A Robinson et al4422
Table 4 Comparison of pixel sizes (instantaneous elds of view) for satellite remote sensorswith pixel sizes for near vertical viewing photography from spacecraft Note that alldata returned from the automated satellites have the same eld of view while indicatedpixel sizes for astronaut photography are best-case for near vertical look angles See gure 5 for distributions of astronaut photographs of various look angles High-altitude aerial photography is also included for reference
Altitude PixelInstrument (km) width (mtimesm) Bands1
Landsat Multipectral Scanner2 (MSS 1974-) 880ndash940 79times56 4ndash5Landsat Thematic Mapper (TM 1982-) ~705 30times30 7Satellite pour lrsquoObservation de la Terre (SPOT 1986-) ~832
SPOT 1-3 Panchromatic 10times10 1SPOT 1-3 Multispectral (XS) 20times20 3
Astronaut Photography (1961-)Skylab3 (1973-1974 ) ~435
Multispectral Camera (6 camera stations) 17ndash19times17ndash19 3ndash8Earth Terrain Camera (hi-res colour) 875times875 3
Space Shuttle5 (1981-) 222ndash611Hasselblad 70 mm (250mm lens colour) vertical view 9ndash26times9ndash26 3Hasselblad 70 mm (100mm lens colour) vertical view 23ndash65times23ndash65 3Nikon 35 mm (400mm lens colour lm or ESC) vertical 5ndash16times5ndash16 3
High Altitude Aerial Photograph (1100 000) ~12 2times2 3
1Three bands listed for aerial photography obtainable by high-resolution colour digitizingFor high-resolution colour lm at 65 lp mmOtilde 1 (K Teitelbaum Eastman Kodak Co personalcommunication) the maximum information contained would be 3302 ppi
2Data from Simonett (1983Table 1-2)3Calculated as recommended by Jensen (19831596) the dividend of estimated ground
resolution at low contrast given in NASA (1974179-180) and 244Six lters plus colour and colour infrared lm (NASA 19747) 5Extracted from table 2
nadir point The look angle (the angle oV nadir that the camera is pointing) can becalculated trigonometrically by assuming a spherical earth and calculating the dis-tance between the coordinates of SN and PC ( gure 1) using the Great Circledistance haversine solution (Sinnott 1984 Snyder 198730ndash32 Chamberlain 1996)The diVerence between the spacecraft centre and nadir point latitudes Dlat=lat 2 shy lat 1 and the diVerence between the spacecraft centre and nadir pointlongitudes Dlon=lon 2 shy lon 1 enter the following equations
a=sin2ADlat
2 B+cos(lat 1) cos(lat 2) sin2ADlon
2 B (4)
c=2 arcsin(min[1 atilde a]) (5)
and oVset=R c with R the radius of a spherical Earth=6370 kmThe look angle (t ) is then given by
t=arctanAoffset
H B (6)
Assuming that the camera was positioned so that the imaginary line between thecentre and nadir points (the principal line) runs vertically through the centre of thephotograph the distance between the geometric centre of the photograph (principalpoint PP) and the top of the photograph is d2 ( gure 8(a)) The scale at any point
Astronaut photography as digital data for remote sensing 4423
(a) (b)
Figure 8 The geometric relationship between look angle lens focal length and the centrepoint and nadir points of a photograph (see also Estes and Simonett 1975) Note thatthe Earthrsquos surface and footprint represented in gure 1 are not shown here only arepresentation of the image area of the photograph Variables shown in the gurelook angle t focal length f PP centre of the image on the lm SN spacecraft nadird original image size (a) The special case where the camera is aligned squarely withthe isometric parallel In this case tilt occurs along a plane parallel with the centre ofthe photograph When the conditions are met calculations using Formulation 3 willgive the ground coordinates of the corner points of the photograph (b) The generalrelationship between these parameters and key photogrammetric points on the photo-graph For this more typical case coordinates of an additional point in the photographmust be input in order to adjust for twist of the camera and to nd the corner pointcoordinates
in the photograph varies as a function of the distance y along the principal linebetween the isocentre and the point according to the relationship
d
D=
f shy ysint
H(7)
(Wong 1980 equation (214) H amp the ground elevation h) Using gure 8(a) at thetop of the photo
y= f tanAt
2B+d
2(8)
and at the bottom of the photo
y= f tanAt
2Bshyd
2(9)
Thus for given PC and SN coordinates and assuming a photo orientation as in gure 8(a) and not gure 8(b) we can estimate a minimum D (using equations (7)and (8)) and a maximum D (using equations (7) and (9)) and then average the twoto determine the pixel size (P ) via equation (2)
J A Robinson et al4424
53 Formulation 3 T he L ow Oblique Space Photo Footprint CalculatorThe Low Oblique Space Photo Footprint Calculator was developed to provide
a more accurate estimation of the geographic coordinates for the footprint of a lowoblique photo of the Earthrsquos surface taken from a human-occupied spacecraft inorbit The calculator performs a series of three-dimensional coordinate transforma-tions to compute the location and orientation of the centre of the photo exposureplane relative to an earth referenced coordinate system The nominal camera focallength is then used to create a vector from the photorsquos perspective point througheach of eight points around the parameter of the image as de ned by the formatsize The geographical coordinates for the photo footprint are then computed byintersecting these photo vectors with a spherical earth model Although more sophist-icated projection algorithms are available no signi cant increase in the accuracy ofthe results would be produced by these algorithms due to inherent uncertainties inthe available input data (ie the spacecraft altitude photo centre location etc)
The calculations were initially implemented within a Microsoft Excel workbookwhich allowed one to embed the mathematical and graphical documentation nextto the actual calculations Thus interested users can inspect the mathematical pro-cessing A set of error traps was also built into the calculations to detect erroneousresults A summary of any errors generated is reported to the user with the resultsof the calculations Interested individuals are invited to download the Excel work-book from httpeoljscnasagovsseopFootprintCalculatorxls The calculations arecurrently being encoded in a high-level programming language and should soon beavailable alongside other background data provided for each photograph at OYceof Earth Sciences (2000)
For the purposes of these calculations a low oblique photograph was de ned as
one with the centre within 10 degrees of latitude and longitude of the spacecraftnadir point For the typical range of spacecraft altitudes to date this restricted thecalculations to photographs in which Earthrsquos horizon does not appear (the generalde nition of a low oblique photograph eg Campbell 199671 and gure 4)
531 Input data and resultsUpon opening the Excel workbook the user is presented with a program intro-
duction providing instructions for using the calculator The second worksheet tablsquoHow-To-Usersquo provides detailed step-by-step instructions for preparing the baselinedata for the program The third tab lsquoInput-Output rsquo contains the user input eldsand displays the results of the calculations The additional worksheets contain theactual calculations and program documentation Although users are welcome toreview these sheets an experienced user need only access the lsquoInput-Outputrsquo
spreadsheetTo begin a calculation the user enters the following information which is available
for each photo in the NASA Astronaut Photography Database (1) SN geographicalcoordinates of spacecraft nadir position at the time of photo (2) H spacecraftaltitude (3) PC the geographical coordinates of the centre of the photo (4) f nominal focal length and (5) d image format size The automatic implementationof the workbook on the web will automatically enter these values and completecalculations
For more accurate results the user may optionally enter the geographic co-ordinates and orientation for an auxiliary point on the photo which resolves thecamerarsquos rotation uncertainty about the optical axis The auxiliary point data must
Astronaut photography as digital data for remote sensing 4425
be computed by the user following the instruction contained in the lsquoHow-To-Usersquotab of the spreadsheet
After entering the input data the geographic coordinates of the photo footprint(ie four photo corner points four points at the bisector of each edge and the centreof the photo) are immediately displayed below the input elds along with any errormessages generated by the user input or by the calculations ( gure 7) Although
results are computed and displayed they should not be used when error messagesare produced by the program The program also computes the tilt angle for each ofthe photo vectors relative to the spacecraft nadir vector To the right of the photofootprint coordinates is displayed the arc distance along the surface of the sphere
between adjacent computed points
532 Calculation assumptions
The mathematical calculations implemented in the Low Oblique Space PhotoFootprint Calculator use the following assumptions
1 The SN location is used as exact even though the true value may vary by upto plusmn01 degree from the location provided with the photo
2 The spacecraft altitude is used as exact Although the determination of the
nadir point at the instant of a known spacecraft vector is relatively precise(plusmn115times10 Otilde 4 degrees) the propagator interpolates between sets of approxi-mately 10ndash40 known vectors per day and the time code recorded on the lmcan drift Thus the true value for SN may vary by up to plusmn01 degree fromthe value provided with the photo
3 The perspective centre of the camera is assumed to be at the given altitudeover the speci ed spacecraft nadir location at the time of photo exposure
4 The PC location is used as exact even though the true value may vary by upto plusmn05deg latitude and plusmn05deg longitude from the location provided with the
photo5 A spherical earth model is used with a nominal radius of 6 372 16154 m (a
common rst-order approximation for a spherical earth used in geodeticcomputations)
6 The nominal lens focal length of the camera lens is used in the computations(calibrated focal length values are not available)
7 The photo projection is based on the classic pin-hole camera model8 No correction for lens distortion or atmospheric re ection is made
9 If no auxiliary point data is provided the lsquoTop of the Imagersquo is orientedperpendicular to the vector from SN towards PC
533 T ransformation from Earth to photo coordinate systemsThe calculations begin by converting the geographic coordinates ( latitude and
longitude) of the SN and PC to a Rectangular Earth-Centred Coordinate System(R-Earth) de ned as shown in gure 9 (with the centre of the Earth at (0 0 0))
Using the vector from the Earthrsquos centre through SN and the spacecraft altitude thespacecraft location (SC) is also computed in R-Earth
For ease of computation a Rectangular Spacecraft-Centred Coordinate System(R-Spacecraft ) is de ned as shown in gure 9 The origin of R-Spacecraft is located
at SC with its +Z-axis aligned with vector from the centre of the Earth throughSN and its +X-axis aligned with the vector from SN to PC ( gure 9) The speci c
J A Robinson et al4426
Figure 9 Representation of the area included in a photograph as a geometric projectiononto the Earthrsquos surface Note that the focal length (the distance from the SpaceShuttle to the photo) is grossly overscale compared to the altitude (the distance fromthe Shuttle to the surface of the Earth
rotations and translation used to convert from the R-Earth to the R-Spacecraft arecomputed and documented in the spreadsheet
With the mathematical positions of the SC SN PC and the centre of the Earthcomputed in R-Spacecraft the program next computes the location of the camerarsquosprincipal point (PP) The principle point is the point of intersection of the opticalaxis of the lens with the image plane (ie the lm) It is nominally positioned at a
Astronaut photography as digital data for remote sensing 4427
distance equal to the focal length from the perspective centre of the camera (whichis assumed to be at SC) along the vector from PC through SC as shown in gure 9
A third coordinate system the Rectangular Photo Coordinate System (R-Photo)is created with its origin at PP its XndashY axial plane normal to the vector from PCthrough SC and its +X axis aligned with the +X axis of R-Spacecraft as shownin gure 9 The XndashY plane of this coordinate system represents the image plane ofthe photograph
534 Auxiliary point calculationsThe calculations above employ a critical assumption that all photos are taken
with the lsquotop of the imagersquo oriented perpendicular to the vector from the SN towardsPC as shown in gure 9 To avoid non-uniform solar heating of the external surfacemost orbiting spacecraft are slowly and continually rotated about one or more axesIn this condition a ight crew member taking a photo while oating in microgravitycould orient the photo with practically any orientation relative to the horizon (seealso gure 8(b)) Unfortunately since these photos are taken with conventionalhandheld cameras there is no other information available which can be used toresolve the photorsquos rotational ambiguity about the optical axis other then the photoitself This is why the above assumption is used and the footprint computed by thiscalculator is actually a lsquolocator ellipsersquo which estimates the area on the ground whatis likely to be included in a speci c photograph (see gure 6) This locator ellipse ismost accurate for square image formats and subject to additional distortion as thephotograph format becomes more rectangular
If the user wants a more precise calculation of the image footprint the photorsquosrotational ambiguity about the optical axis must be resolved This can be done inthe calculator by adding data for an auxiliary point Detailed instructions regardinghow to prepare and use auxiliary point data in the computations are included in thelsquoHow-To-Usersquo tab of the spreadsheet Basically the user determines which side ofthe photograph is top and then measures the angle between the line from PP to thetop of the photo and from PP to the auxiliary point on the photo ( gure 9)
If the user includes data for an auxiliary point a series of computations arecompleted to resolve the photo rotation ambiguity about the optical axis (ie the
+Z axis in R-Photo) A vector from the Auxiliary Point on the Earth (AE) throughthe photograph perspective centre ( located at SC) is intersected with the photo imageplane (XndashY plane of R-Photo) to compute the coordinates of the Auxiliary Point onthe photo (AP) in R-Photo A two-dimensional angle in the XndashY plane of R-Photofrom the shy X axis to a line from PP to AP is calculated as shown in gure 9 The
shy X axis is used as the origin of the angle since it represents the top of the photoonce it passes through the perspective centre The diVerence between the computedangle and the angle measured by the user on the photo resolves the ambiguity inthe rotation of the photo relative to the principal line ( gures 7 and 9) Thetransformations from R-Spacecraft and R-Photo are then modi ed to include anadditional rotation angle about the +Z axis in R-Photo
535 lsquoFootprint rsquo calculationsThe program next computes the coordinates of eight points about the perimeter
of the image format (ie located at the four photo corners plus a bisector pointalong each edge of the image) These points are identi ed in R-Photo based uponthe photograph format size and then converted to R-Spacecraft Since all computa-tions are done in orthogonal coordinate systems the R-Spacecraft to R-Photo
J A Robinson et al4428
rotation matrix is transposed to produce an R-Photo to R-Spacecraft rotation matrixOnce in R-Spacecraft a unit vector from each of the eight perimeter points throughthe photo perspective centre (the same point as SC) is computed This provides thecoordinates for points about the perimeter of the image format with their directionvectors in a common coordinate system with other key points needed to computethe photo footprint
The next step is to compute the point of intersection between the spherical earthmodel and each of the eight perimeter point vectors The scalar value for eachperimeter point unit vector is computed using two-dimensional planar trigonometryAn angle c is computed using the formula for the cosine of an angle between twoincluded vectors (the perimeter point unit vector and the vector from SC to thecentre of the Earth) Angle y is computed using the Law of Sines Angle e=180degrees shy c shy y The scalar value of the perimeter point vector is computed using eand the Law of Cosines The scalar value is then multiplied by the perimeter pointunit vector to produce the three-dimensional point of intersection of the vector withEarthrsquos surface in R-Spacecraft The process is repeated independently for each ofthe eight perimeter point vectors Aside from its mathematical simplicity the valueof arcsine (computed in step 2) will exceed the normal range for the sin of an anglewhen the perimeter point vectors fail to intersect with the surface of the earth Asimple test based on this principle allows the program to correctly handle obliquephotos which image a portion of the horizon (see results for high oblique photographsin table 5)
The nal step in the process converts the eight earth intersection points from theR-Spacecraft to R-Earth The results are then converted to the standard geographiccoordinate system and displayed on the lsquoInput-Outputrsquo page of the spreadsheet
54 ExamplesAll three formulations were applied to the photographs included in this paper
and the results are compared in table 5 Formulation 1 gives too small a value fordistance across the photograph (D) for all but the most nadir shots and thus servesas an indicator of the best theoretical case but is not a good measure for a speci cphotograph For example the photograph of Lake Eyre taken from 276 km altitude( gure 2(a)) and the photograph of Limmen Bight ( gure 7) were closest to beingnadir views (oVsetslt68 km or tlt15deg table 5) For these photographs D calculatedusing Formulation 1 was similar to the minimum D calculated using Formulations2 and 3 For almost all other more oblique photographs Formulation 1 gave asigni cant underestimate of the distance covered in the photograph For gure 5(A)(the picture of Houston taken with a 40 mm lens) Formulation 1 did not give anunderestimate for D This is because Formulation 1 does not account for curvatureof the Earth in any way With this large eld of view assuming a at Earth in atedthe value of D above the minimum from calculations that included Earth curvature
A major diVerence between Formulations 2 and 3 is the ability to estimate pixelsizes (P) in both directions (along the principal line and perpendicular to the principalline) For the more oblique photographs the vertical estimate of D and pixel sizes ismuch larger than in the horizontal direction (eg the low oblique and high obliquephotographs of Hawaii gure 4 table 5)
For the area of Limmen Bight ( gure 7) table 5 illustrates the improvement inthe estimate of distance and pixel size that can be obtained by re-estimating thelocation of the PC with greater accuracy Centre points in the catalogued data are
Astronaut photography as digital data for remote sensing 4429
plusmn05deg of latitude and longitude When the centre point was re-estimated to plusmn002degwe determined that the photograph was not taken as obliquely as rst thought(change in the estimate of the look angle t from 169 to 128deg table 5) When theauxiliary point was added to the calculations of Formula 3 the calculated look angleshrank further to 110deg indicating that this photograph was taken at very close toa nadir view Of course this improvement in accuracy could have also led to estimatesof greater obliquity and corresponding larger pixel sizes
We also tested the performance of the scale calculator with auxiliary point byestimating the corner and point locations on the photograph using a 11 000 000Operational Navigational Chart For this test we estimated our ability to readcoordinates from the map as plusmn002degand our error in nding the locations of thecorner points as plusmn015deg (this error varies among photographs depending on thedetail that can be matched between photo and map) For Limmen Bight ( gure 7)the mean diVerence between map estimates and calculator estimates for four pointswas 031deg (SD=018 n=8) For a photograph of San Francisco Bay (STS062-151-291) the mean diVerence between map estimates and calculator estimates forfour points was 0064deg (SD=018 n=8) For a photograph of San Francisco Bay(STS062-151-291) the mean diVerence between map estimates and calculator estim-ates for eight points was 0196deg (SD=0146 n=16) Thus in one case the calculatorestimates were better than our estimate of the error in locating corner points on themap It is reasonable to expect that for nadir-viewing photographs the calculatorused with an auxiliary point can estimate locations of the edges of a photograph towithin plusmn03deg
55 Empirical con rmation of spatial resolution estimatesAs stated previously a challenge to estimating system-AWAR for astronaut
photography of Earth is the lack of suitable targets Small features in an image cansometimes be used as a check on the size of objects that can be successfully resolvedgiving an approximate value for GRD Similarly the number of pixels that make upthose features in the digitized image can be used to make an independent calculationof pixel size We have successfully used features such as roads and airport runwaysto make estimates of spatial scale and resolution (eg Robinson et al 2000c) Whilerecognizing that the use of linear detail in an image is a poor approximation to abar target and that linear objects smaller than the resolving power can often bedetected (Charman 1965) few objects other than roads could be found to make anydirect estimates of GRD Thus roads and runways were used in the images ofHouston (where one can readily conduct ground veri cations and where a numberof higher-contrast concrete roadways were available) to obtain empirical estimatesof GRD and pixel size for comparison with table 5
In the all-digital ESC image of Houston ( gure 3(d )) we examined EllingtonField runway 4-22 (centre left of the image) which is 24384 mtimes457 m This runwayis approximately 6ndash7 pixels in width and 304ndash3094 pixels in length so pixelsrepresent an distance on the ground 7ndash8 m Using a lower contrast measure of astreet length between two intersections (2125 m=211 pixels) a pixel width of 101 mis estimated These results compare favourably with the minimum estimate of 81 mpixels using Formulation 1 (table 5) For an estimate of GRD that would be morecomparable to aerial photography of a line target the smallest street without treecover that could be clearly distinguished on the photograph was 792 m wide Thesmallest non-street object (a gap between stages of the Saturn rocket on display in
J A Robinson et al4430
Tab
le5
App
lica
tio
nof
met
hod
sfo
res
tim
ati
ng
spati
alre
solu
tio
no
fast
ron
aut
ph
oto
gra
ph
sin
clu
din
gF
orm
ula
tion
s12
and
3Im
pro
ved
cen
tre
po
int
esti
mat
esan
dauxilia
ryp
oin
tca
lcu
lati
ons
are
sho
wn
for
gure
7V
aria
ble
sas
use
din
the
text
fle
ns
foca
lle
ngt
hH
al
titu
de
PC
ph
oto
gra
ph
cen
tre
poin
tSN
sp
ace
craft
nadir
po
int
D
dis
tance
cover
edo
nth
egr
oun
d
Pd
ista
nce
on
the
grou
nd
repre
sen
ted
by
apix
el
OVse
td
ista
nce
bet
wee
nP
Cand
SNusi
ng
the
grea
tci
rcle
form
ula
t
look
angle
away
from
SN
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
2L
ake
Eyre
ST
S093
-717
-80
250
276
shy29
0137
0shy
28
413
71
607
11
767
413
7609
ndash64
212
013
760
9ndash64
7121
124
ST
S082
-754
-61
250
617
shy28
5137
0shy
28
113
50
135
726
1201
18
013
8ndash14
827
517
9138ndash15
3277
294
Fig
ure
3H
ou
sto
nS
TS
062
-97-
151
4029
829
5shy
95
530
5shy
95
4410
78
8112
20
5348ndash58
990
1205
351
ndash63
8951
10
36
ST
S094
-737
-28
100
294
295
shy
95
528
3shy
95
5162
31
1133
24
415
8ndash20
334
724
3158ndash20
7351
390
ST
S055
-71-
41250
302
295
shy
95
028
2shy
93
766
412
8192
32
5736
ndash84
715
232
273
9ndash85
7154
185
S73
E51
45300
2685
295
shy
95
024
78
0F
igu
re4
Haw
aii
ST
S069
-701
-65
250
370
195
shy
155
520
4shy
153
881
415
7204
28
9876
ndash98
917
928
688
0ndash10
9181
210
ST
S069
-701
-70
250
370
210
shy
157
519
2shy
150
781
415
7738
63
414
9ndash23
336
760
7148ndash55
6394
10
63
ST
S069
-729
-27
100
398
200
shy
156
620
5shy
148
2219
42
1867
65
432
8ndash13
10
158
62
4323+
311
N
A
Astronaut photography as digital data for remote sensing 4431
Tab
le5
(Cont
inued
)
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
7L
imm
enB
igh
tS
TS
093
-704
-50
250
283
shy15
0135
0shy
15
013
58
623
120
859
16
963
0ndash67
3125
16
863
0ndash67
612
613
2P
CR
e-es
tim
ated
shy14
733
1352
67
64
5128
623
ndash65
5123
128
623
ndash65
712
3127
Auxilia
ryP
oin
t611
062
3ndash65
112
312
5F
igu
re10
N
ear
Au
stin
T
XS
TS
095
-716
-46
250
543
30
5shy
975
28
6shy
1007
120
230
375
34
613
5ndash157
281
34
1136ndash16
128
636
0
1 All
pho
togr
aphs
exce
pt
on
eta
ken
wit
hH
ass
elbla
dca
mer
aso
dis
tan
ceacr
oss
the
image
d=
55m
m
for
S73
E51
45
taken
wit
hel
ectr
onic
still
cam
erad=
276
5m
mho
rizo
nta
l2 P
Cand
SN
are
giv
enin
the
form
(lati
tud
elo
ngit
ude)
deg
rees
w
ith
neg
ativ
en
um
ber
sre
pre
senti
ng
south
and
wes
tA
ccura
cyo
fP
Cis
plusmn0
5and
SN
isplusmn
01
as
reco
rded
inth
eA
stro
naut
Ph
oto
gra
phy
Data
base
ex
cep
tw
her
en
ote
dth
atce
ntr
ep
oin
tw
asre
-est
imate
dw
ith
grea
ter
accu
racy
3 C
alc
ula
ted
usi
ng
the
mea
nofth
em
inim
um
an
dm
axim
um
esti
mate
sof
DF
orm
ula
tion
2co
nsi
der
sD
inth
ed
irec
tion
per
pen
dic
ula
rto
the
Pri
nci
pal
Lin
eo
nly
4 C
alc
ula
ted
inho
rizo
nta
lan
dve
rtic
aldir
ecti
on
sb
ut
still
ass
um
ing
geom
etry
of
gure
6(B
)F
irst
nu
mb
eris
the
mea
no
fth
em
inim
um
Dat
the
bott
om
of
the
ph
oto
and
the
maxi
mum
Dat
the
top
of
the
ph
oto
(range
show
nin
the
tab
le)
and
isco
mpara
ble
toF
orm
ula
tio
n2
Sec
ond
num
ber
isth
em
ean
of
Des
tim
ated
alon
gth
eP
rin
cip
alline
and
alo
ng
the
ou
tsid
eed
ge
(range
not
show
n)
5 Alt
itu
de
isap
pro
xim
ate
for
mis
sion
as
exac
tti
me
pho
togr
ap
hw
asta
ken
and
corr
esp
on
din
gn
adir
data
are
no
tava
ilab
le
Lack
of
nad
irdata
pre
ven
tsap
plica
tion
of
Form
ula
tion
s2
an
d3
6 Au
xilliar
ypo
int
add
edw
as
shy14
80
lati
tud
e13
51
2lo
ngit
ude
shy46
5degro
tati
on
angle
in
stru
ctio
ns
for
esti
mati
ng
the
rota
tio
nan
gle
para
met
erare
conta
ined
as
par
tof
the
scal
eca
lcu
lato
rsp
read
shee
tdo
cum
enta
tio
n
J A Robinson et al4432
a park at Johnson Space Center) that could clearly be distinguished on the photo-graph was 853 m wide
For the photograph of Houston taken with a 250 mm lens ( gure 3(C)) anddigitized from second generation lm at 2400 ppi (106 mm pixel Otilde 1 ) Ellington Fieldrunway 4-22 is 3 pixels in width and 1613 pixels in length so pixels represent151ndash152 m on the ground These results compare favourably with the minimumestimate of 152 m using Formulation 2 and 154ndash185 m pixels using Formulation 3(table 5) For an estimate of GRD using an 8times8 inch print (1369 enlargement) and4times magni cation the smallest street that could clearly be distinguished was 822 mwide the same feature could barely be distinguished on the digitized image
We also made an empirical estimate of spatial resolution for lower contrastvegetation boundaries By clearing forest so that a pattern would be visible to landingaircraft a landowner outside Austin Texas (see also aerial photo in Lisheron 2000)created a target that is also useful for evaluating spatial resolution of astronautphotographs The forest was selectively cleared in order to spell the landownerrsquosname lsquoLUECKErsquo with the remaining trees ( gure 10) According to local surveyorswho planned the clearing the plan was to create letters that were 3100 fttimes1700 ft(9449 mtimes5182 m) Photographed at a high altitude relative to most Shuttle missions(543 km) with a 250 mm lens Formula 3 predicts that each pixel would represent anarea 286 mtimes360 m on the ground (table 5) When original lm was digitized at2400 ppi (106 mm pixel Otilde 1 ) letters correspond to 294times188 pixels for a comparablepixel size of 27ndash32 m
6 Summary and conclusionsAstronaut photographs can be an excellent source of data for remote sensing
applications Best-case resolutions are similar to that for Landsat or SPOT withpixels as small as lt10 m It was estimated that of images taken to date 504 hadlens and obliquity characteristics that make them potential candidates for remotesensing information Digitized astronaut photographs can be overlaid with othersatellite data using GIS (Eckardt et al 2000) or used to ll in gaps in time serieswhen other imagery is not available As a source of public-domain information theycan be very useful for scientists who do not have access to satellite imagery eitherbecause of the expense of image acquisition (to the end user) or the computersystems needed for processing satellite images These diVerences in image acquisitioncosts (to the user) are summarized in table 6
Searching of the complete database of NASA astronaut photography includinglow-spatial-resolution browse images is available via the Web (OYce of EarthSciences 2000) This provides nearly global access for identifying images that willcontribute to a speci c scienti c project For the most detailed studies digitalproducts posted to the web will not be of suYcient quality To date we have madeit a practice of digitizing small numbers of images when requested by scientists atno charge Such requests (including a description of the project involved) can bemade through the authors or using the contact information listed on the websiteInvestigators with needs for larger numbers of digital images have also been servedthrough collaborative agreements
In our experience scientists that nd the data most valuable are those in develop-ing countries those studying areas that have not usually been targets for the majorsatellites those having diYculty in nding low-cloud images those interested inconstructing time series or those interested in using a large number of images
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
J A Robinson et al4406
at high contrast (objectbackground ratio 10001) and low contrast (objectback-ground ratio 161) Most terrestrial surfaces recorded from orbit are low contrastfor the purpose of estimating resolving power of lm Kodak no longer measures lm resolving power for non-aerial photographic lms (Karen Teitelbaum EastmanKodak Company personal communication) including lms that NASA routinelyuses for Earth photography
AWAR can be measured for the static case of lm and camera or for thecamerandashaircraft system in motion For example AWAR for the National AerialPhotography Program includes eVects of the lens resolving power of original lmimage blur due to aircraft motion and spatial resolution of duplicating lm (Light1996) Given AWAR for a system in motion the GRD can be calculated bytrigonometry (see equation (3) in sect5 d=1AWAR and D=GRD)
An impediment to similar rigorous measurement of GRD for orbital remotesensing systems is the lack of a target of suitable scale on the ground thus spatialresolution for most orbiting sensors is described in terms of a less all-encompassingmeasure instantaneous eld of view (see below) Additional challenges to measuringAWAR for a complete astronaut photography system include the number of diVerentoptions for aspects of the system including diVerent cameras lms and orbitalaltitudes These elements that must be standardized to determine AWAR providea useful list of those characteristics of astronaut photography that will mostin uence GRD
22 Instantaneous eld of viewThe instantaneous eld of view (IFOV) is generally used to represent the spatial
resolution of automated satellite systems and is commonly used interchangeablywith the term spatial resolution when comparing diVerent sensors IFOV is a combina-tion of geometric mechanical and electronic properties of the imaging systemGeometric properties including satellite orbital altitude detector size and the focallength of the optical system (Simonett 1983) The sensitivity of each detector elementat the wavelength desired plus the signal-to-noise level desired are electronic proper-ties that determine a minimum time for energy absorption For a linear sensor array(pushbroom scanning) this minimum time is translated to areal coverage by theforward velocity of the platform (Campbell 1996 p 97) Usually each detectorelement in the array corresponds to a pixel in the image Thus for a given altitudethe width of the pixel is determined by the optics and sensor size and the height ofthe pixel is determined by the rate of forward motion When magni ed by the ratioof the sensor altitude to the focal length of the optics of the sensor system IFOV isthe size of the area on the ground represented by an individual detector element(pixel Slater 1980 p 27)
The equation of IFOV with spatial resolution can be misleading because IFOVdoes not include factors other than geometrymdashfactors that largely determine thelevel of detail that can be distinguished in an image For example IFOV does notinclude characteristics of the target (contrast with surroundings shape of an objectcolour) atmospheric conditions illumination and characteristics of the interpreter(machine or human) Factors in uencing spatial resolution that are included in IFOVcan be calculated from design speci cations before the system has been built whereasthe actual spatial resolution of an image captured by the sensor will be unique tothat image IFOV represents the best spatial resolution possible given optimalconditions
Astronaut photography as digital data for remote sensing 4407
23 Imagery from scanners versus lm from camerasLike aerial photography systems the GRD of early remote sensing scanners
(electro-optical imaging systems) was calibrated empirically ( gures 1ndash6 in Simonett1983) but such methods are impractical for routine applications Satellites are nowcommon platforms for collecting remote sensing data small-scale aerial photographsare widely available within the USA and Europe and astronaut photographshave become available worldwide Straightforward comparison of resolutions ofphotographs to IFOVs of scanner images is necessary for many applications
How can spatial resolutions of data from diVerent sources be compared usingcommon units Oft-quoted lsquoresolution numbersrsquo actually IFOV of digital scanningsystems should be at least doubled for comparison to the standard method ofexpressing photographi c spatial resolution as GRD simply because of the objectcontrast requirement The following rule of thumb has been used in geographicapplications to compare spatial resolution of scanners and aerial photography (Welch1982 Jensen 1983) a low-contrast target may be converted to an approximate IFOVby dividing the GRD (of the photograph) by 24 Such a rule of thumb might alsobe reasonably applied to astronaut photography making it possible to convertobserved GRD to IFOV and IFOV to an estimated GRD
In this paper we discuss the spatial resolution of astronaut photographs in termsof system properties and compute the area on the ground represented by a singlepixel the equivalent to IFOV To complete our treatment of spatial resolution wealso provide estimates of the sizes of small features identi able in images for severalcases where this can be readily done
3 Factors that determine the footprint (area covered by the photograph )The most fundamental metric that forms the basis for estimating spatial resolution
of astronaut photographs is the size of the footprint or area on the ground capturedin a photograph (see review of satellite photogrammetry by Light 1980) The basicgeometric variables that in uence the area covered by an astronaut photograph are(1) the altitude of the orbit H (2) the focal length of the lens f (3) the actual sizeof the image on the lm d and (4) the orientation of the camera axis relative to theground (the obliquity or look angle t ) The relationships among these parametersare illustrated in gure 1 (also see equation (3) in sect5)
31 AltitudeHuman space ight missions have had a variety of primary objectives that required
diVerent orbital altitudes The higher the altitude the larger the footprint of thephotographs The diVering scale of photographs taken at diVerent altitudes is illus-trated in gure 2 The two photographs of Lake Eyre Australia were taken with thesame camera and lens but on diVerent dates and from diVerent altitudes In (a) thelake is relatively ooded while in (b) it is dry A 23timesdiVerence in altitude leads toa corresponding diVerence in the scales of the resulting photographs
32 L ensesThe longer the lens focal length the more magni cation the greater detail and
the smaller footprint A variety of lenses with diVerent focal lengths are own oneach space mission (table 1) The eVect of lens length on spatial coverage and imagedetail is shown in gure 3 In these views of Houston (a)ndash(c) taken from approxi-mately similar altitudes with the same camera most of the diVerence in scale of the
J A Robinson et al4408
Figure 1 Geometric relationship between look angle (t) spacecraft nadir point (SN) photo-graph centre point (PC) and photograph principal point (PP) The dark shaded arearepresents Earthrsquos surface The distance covered on the ground D corresponds totable 5 Original image size d is listed for various lms in table 1 Variables correspondto equations (2)ndash(9) in sect5
Astronaut photography as digital data for remote sensing 4409
(a) (b)
Figure 2 Example of the eVect of altitude on area covered in a photograph Both photographsof Lake Eyre Australia were taken using a Hasselblad camera with 100 mm lens(a) altitude 283 km STS093-702-062 (b) altitude 617 km STS082-754-61
photographs is due to the diVerent magni cations of the lenses Although taken froma diVerent altitude and using a diVerent camera with a diVerent original image size(see table 2) an electronic still camera image taken with a 300 mm lens is alsoincluded for comparison
For remote sensing longer focal length lenses are generally preferred (250 mm or350 mm for Hasselblad 300 mm or 400 mm lenses for 35 mm format cameras andESCs) Unfortunately longer-focal-length lenses exhibit poorer performance towardthe edge of a frame For example a 250 mm lens (Distagon CF 56) used on theHasselblad camera has spatial resolution of 57 lp mm Otilde 1 at the centre but only51 lp mm Otilde 1 at the edge and 46 lp mm Otilde 1 at the corner (tested with Ektachrome 5017f8 aperture and high contrast) A longer 300 mm lens used on the Nikon camerahas a greater spatial resolution diVerence between centre (82 lp mm Otilde 1 ) and corner(49 lp mm Otilde 1 ) using Kodak 5017 Ektachrome f4 aperture and high contrast FredPearce (unpublished data) There are tradeoVs among lens optics and speed Forexample the lenses for the Linhof system and the 250 mm lens for the Hasselblad(Distagon CF 56 see footnotes to table 1) are limited to apertures smaller thanf56 This becomes an important constraint in selecting shutter speeds (see discussionof shutter speed in sect42)
33 Cameras and actual image sizesAfter passing through the lens the photographic image is projected onto lm
inside the camera The size of this original image is another important propertydetermining spatial resolution and is determined by the camera used Cameraformats include 35 mm and 70 mm (Lowman 1980 Amsbury 1989) and occasionally5times4 inch (127times100 mm) and larger (table 1) Cameras own on each mission arenot metricmdashthey lack vacuum platens or reseau grids image-motion compensationor gyro-stabilized mounts The workhorse for engineering and Earth photographyon NASA missions has been a series of 70 mm Hasselblad cameras (table 1) chosenfor their reliability The modi ed magazine databack imprints a unique number and
J A Robinson et al4410
Tab
le1
Det
ails
on
the
cam
eras
use
dfo
rast
ron
aut
ph
oto
gra
phy
of
Eart
h
Data
incl
ud
ep
ho
togr
aphy
from
all
NA
SA
hu
man
spac
eig
ht
mis
sion
sth
rough
ST
S-9
6(J
un
e19
99
378
461
tota
lfr
am
esin
the
dat
abas
e)1
Imag
esi
zeT
ypic
al
lense
sN
um
ber
of
Cam
era
make
(mm
timesm
m)
use
d(m
m)
ph
oto
gra
ph
sP
erio
dof
use
No
tes
Has
selb
lad
55times
55
40
50100
2502
288
494
1963ndashp
rese
nt
Lin
ho
f10
5times120
90
250
29
373
1984ndashp
rese
nt
No
won
lyo
wn
occ
asio
nal
lyM
ult
ispec
tral
Cam
era
(S19
0A
)57times
57152
32
913
1973ndash19
74S
kyla
b
xed
-mou
nt3
cam
era
(NA
SA
1974
)E
art
hT
erra
inC
amer
a(S
190B
)11
5times11
5457
5478
1973ndash19
74S
kyla
b
xed
-mou
nt3
cam
era
(NA
SA
1974
)R
ole
iex
55times
55
50
100250
4352
1990ndash19
92L
arg
eF
orm
at
Cam
era4
229
times457
305
2143
1984
Mis
sio
nST
Sndash41G
only
(Mer
kel
etal
198
5)
Maure
r55
times55
76
80818
1961ndash19
68
1 Sm
all-f
orm
atca
mer
as
(35
mm
lm
and
ES
C)
are
no
tin
clud
edin
this
table
bec
ause
of
the
larg
enu
mb
erof
len
ses
and
con
gu
rati
on
sth
at
have
bee
nu
sed
and
bec
au
seth
ese
data
are
no
tcu
rren
tly
incl
uded
inth
edata
bas
efo
rall
mis
sion
s2 L
ense
suse
dfo
rH
ass
elbla
dm
anufa
cture
dby
Carl
Zei
ss
Dis
tago
nC
F4
(40
mm
)D
ista
gon
CF
4(5
0m
m)
Dis
tagon
CF
35
(100
mm
)D
ista
go
nC
F5
6(2
50
mm
)S
tand
ard
Has
selb
lad
len
ses
ow
nw
ill
chan
ge
inth
eyea
r20
00
toD
ista
gon
FE
28
(50
mm
)D
ista
go
nF
E2
(110
mm
)T
ele-
Tes
sar
FE
4(2
50m
m)
and
Tel
e-T
essa
rF
E4
(350
mm
)3 T
hes
eca
mer
as
wer
ein
xed
mou
nti
ngs
on
the
outs
ide
of
the
Skyla
bsp
ace
stati
on
A
ltho
ugh
not
hand
hel
d
the
lm
isca
talo
gued
wit
ho
ther
ast
ron
aut
ph
oto
gra
ph
yan
dis
incl
uded
her
efo
rco
mp
lete
nes
s4 A
NA
SA
-des
igned
cam
era
wit
hatt
itud
ere
fere
nce
syst
emfo
rd
eter
min
ing
pre
cise
nadir
poin
ts
Dat
ais
no
tcu
rren
tly
inco
rpora
ted
into
on
lin
ed
atab
ase
bu
tis
cata
loged
by
Mer
kel
etal
(1
985
)
Astronaut photography as digital data for remote sensing 4411
(a) (b)
(c) (d)
Figure 3 Example of the eVect of lens focal length on resolving power Hasselblad imagesof Houston were all taken from an altitude of 294 kmndash302 km with diVerent lenses(a) 40 mm STS062-97-151 (b) 100 mm STS094-737-28 (c) 250 mm STS055-71-41 Forcomparison an Electronic Still Camera image with 300 mm lens at 269 km altitude isalso shown (d ) S73E5145
timestamp on each frame at the time of exposure Cameras are serviced between ights Occasionally there is enough volume and mass allowance so that a Linhof5times4 inch (127times101 mm) format camera can be own Nikon 35 mm cameras are own routinely also because of proven reliability Electronic still cameras (ESC)were tested for Earth photography beginning in 1992 (Lulla and Holland 1993) Inan ESC a CCD (charge-coupled device) is used as a digital replacement for lmrecording the image projected inside the camera An ESC (consisting of a KodakDCS 460c CCD in a Nikon N-90S body) was added as routine equipment forhandheld photographs in 1995 ESCs have also been operated remotely to captureand downlink Earth images through a NASA-sponsored educational programme(EarthKAM) Discussion of CCD spatial array and radiometric sensitivity arebeyond the scope of this paper but are summarized by Robinson et al (2000b) The
J A Robinson et al4412
format of the lm (or CCD) and image size projected onto the lm (or CCD) aresummarized for all the diVerent cameras own in table 2
34 L ook angle or obliquityNo handheld photographs can be considered perfectly nadir they are taken at a
variety of look angles ranging from near vertical ( looking down at approximatelythe nadir position of the spacecraft ) to high oblique (images that include the curvatureof Earth) Imaging at oblique look angles leads to an image where scale degradesaway from nadir A set of views of the same area from diVerent look angles is shownin gure 4 The rst two shots of the island of Hawaii were taken only a few secondsapart and with the same lens The third photograph was taken on a subsequentorbit and with a shorter lens The curvature of the Earth can be seen in the upperleft corner
Obliquity can be described qualitatively ( gure 4) or quantitatively as the lookangle (t gure 1 calculations described in formulation 2 (sect52)) Because obliquityand look angle have such a dramatic in uence on the footprint we summarize thedatabase characteristics relative to these two parameters Figure 5 is a breakdownof the spatial resolution characteristics of low oblique and near vertical photographsin the NASA Astronaut Photography database Number of photographs are grouped(1) by calculated values for look angle (t ) and (2) by altitude After observing theoverlap between near vertical and low oblique classes we are currently restructuringthis variable (lsquotiltrsquo) in the database to provide a measure of t when available Userswill still be able to do searches based on the qualitative measures but these measureswill be more closely tied to actual look angle
341 Obliquity and georeferencing digitised photographsOften the rst step in a remote sensing analysis of a digitized astronaut photo-
graph is to georeference the data and resample it to conform to a known mapprojection Details and recommendations for resampling astronaut photographydata are provided by Robinson et al (2000a 2000c) and a tutorial is also available(McRay et al 2000) Slightly oblique photographs can be geometrically correctedfor remote sensing purposes but extremely oblique photographs are not suited forgeometric correction When obliquity is too great the spatial scale far away fromnadir is much larger than the spatial scale closer to nadir resampling results areunsuitable because pixels near nadir are lost as the image is resampled while manypixels far away from nadir are excessively replicated by resampling
To avoid the generation of excess pixels during georeferencing the pixel sizes ofthe original digitized image should be smaller than the pixels in the nal resampledimage Calculations of original pixel size using methods presented below can beuseful in ensuring meaningful resampling For slightly oblique images formulation 3(sect53) can be used to estimate pixel sizes at various locations in a photograph (nearnadir and away from nadir) and these pixel sizes then used to determine a reasonablepixel scale following resampling
4 Other characteristics that in uence spatial resolutionBeyond basic geometry many other factors internal to the astronaut photography
system impact the observed ground resolved distance in a photograph The lensesand cameras already discussed are imperfect and introduce radiometric degradations(Moik 1980)Vignetting (slight darkening around the edges of an image) occurs in
Astronaut photography as digital data for remote sensing 4413
Tab
le2
Max
imu
mpo
ssib
led
igit
alsp
atia
lre
solu
tio
n(p
ixel
size
inm
)fo
rca
mer
asy
stem
scu
rren
tly
use
dby
ast
ron
auts
top
ho
togr
aph
eart
hb
ase
do
nth
ege
om
etry
of
alt
itud
ele
ns
and
lm
size
C
alc
ula
tion
sas
sum
eth
atco
lour
tran
spar
ency
lm
isdig
itiz
edat
2400
ppi
(pix
els
per
inch
10
6m
mpix
elOtilde
1 )
Min
imu
mgro
un
ddis
tance
repre
sente
dby
1p
ixel
(m)
As
afu
nct
ion
of
alti
tude1
Imag
esi
zeM
inim
um
alt
itud
eM
edia
nalt
itu
de
Maxi
mum
alt
itud
eC
amer
asy
stem
mm
timesm
mp
ixel
s2(2
22k
m)
(326
km
)(6
11
km
)
Lin
ho
f125
mm
90
mm
lens
105
times120
8978times
11
434
261
383
718
250
mm
lens
94
138
259
Has
selb
lad
70
mm
100
mm
lens
55times
55
5198
times519
823
534
564
725
0m
mle
ns
94
138
259
Nik
on
35m
m30
0m
mle
ns
36times
24
3308
times217
478
115
216
400
mm
lens
59
86
162
ESC
35m
m18
0m
mle
ns3
27
6times
18
530
69times
204
311
116
330
630
0m
mle
ns
67
98
184
400
mm
lens
50
74
138
1 Alt
itud
essh
ow
nar
em
inim
um
m
edia
nan
dm
axim
um
thro
ugh
ST
S-8
9(J
anuary
199
8N
=84
mis
sion
s)
2 Act
ual
pix
els
for
ES
C
oth
erw
ise
assu
med
dig
itiz
ati
on
at240
0pp
i(p
ixel
sp
erin
ch)
equ
alto
10
6m
mpix
elOtilde
1 3 T
he
mo
stco
mm
on
con
gura
tion
for
Eart
hph
oto
gra
ph
yas
part
of
the
Eart
hK
AM
pro
ject
J A Robinson et al4414
Figure 4 De nitions of look angle for photographs taken from low orbit around EarthThe example photographs of Hawaii were all taken from approximately the samealtitude (370ndash398 km) and with the same (250 mm) lens on a Hasselblad camera(STS069-701-69 STS069-701-70 and STS069-729-27 [100 mm lens])
astronaut photography due to light path properties of the lenses used A lack of atness of the lm at the moment of exposure (because cameras used in orbit donot incorporate a vacuum platen) also introduces slight degradations A numberof additional sources of image degradation that would aVect GRD of astronautphotographs are listed
41 Films and processing411 Colour reversal lms
Most lms used in NASA handheld cameras in orbit have been E-6 processcolour reversal lms (Ektachromes) that have a nominal speed of 64 to 100 ASAalthough many other lms have been own (table 3) Reversal lms are used becausedamage from radiation exposure while outside the atmospheric protection of Earthis more noticeable in negative lms than in positive lms (Slater 1996) The choiceof ASA has been to balance the coarser grains ( lower lm resolving power) of high-speed lms with the fact that slower lms require longer exposures and thus areaVected more by vehicle motion (approximately 73 km s Otilde 1 relative to Earthrsquos surfacefor the Space Shuttle) Extremely fast lms (gt400 ASA) are also more susceptibleto radiation damage in orbit (particularly during long-duration or high-altitudemissions Slater 1996) and have not traditionally been used for Earth photography The manufacturer stock numbers identifying the lm used are available for eachimage in the Astronaut Photography Database (OYce of Earth Sciences 2000)
412 Colour infrared lmsColour infrared lm (CIR) was used during an Earth-orbiting Apollo mission
(Colwell 1971) in the multispectral S-190A and the high-resolution S-190B camera
Astronaut photography as digital data for remote sensing 4415
Figure 5 (a) Distributions of astronaut photographs by look angle oV-nadir (calculated fromcentre point and nadir point using Formulation 2) Data included for all photographsfor which centre and nadir points are known with high oblique look angles (n=55 155) excluded (Total sample shown is 108 293 of 382 563 total records in thedatabase when compiled For records where nadir data was not available number ofobservations for qualitative look angles are near vertical 14 086 low oblique 115 834undetermined 89 195) (b) Altitude distribution for astronaut photographs of Earthtaken with the Hasselblad camera Data included for Hasselblad camera only focallength determined with high oblique look angles excluded (Total sample shown is159 046 of 206 971 Hasselblad records with known altitude and of 382 563 total recordsin the database when compiled For Hasselblad records with known altitude data notshown includes high oblique photographs using 100 mm and 250 mm lenses n=20 262and all look angles with other lenses n=27 663)
systems on Skylab (NASA 1974 Wilmarth et al 1977) and occasionally on Shuttlemissions and Shuttle-Mir missions (table 3) The CIR lm used is a three-layerAerochrome having one layer that is sensitive to re ected solar infrared energy toapproximately 900 nm (Eastman Kodak Company 1998) protective coatings onmost spacecraft windows also limit IR transmittance between 800 mm and 1000 nmThis layer is extremely sensitive to the temperature of the lm which createsunpredictable degradation of the IR signature under some space ight conditions
J A Robinson et al4416
Tab
le3
Det
ails
on
lm
suse
dfo
rast
ron
aut
pho
togr
aphy
of
Eart
h
Dat
ain
clud
ep
ho
togr
aphy
fro
mall
NA
SA
hum
an
spac
eig
ht
mis
sio
ns
thro
ugh
ST
S-9
6(J
un
e19
99
378
461
tota
lfr
ames
inth
edata
base
)F
ilm
sw
ith
lt50
00fr
am
esex
po
sed
and
lm
suse
dfo
r35
mm
cam
eras
only
are
no
tin
clud
ed
Res
olv
ing
Nu
mb
erof
Film
man
ufa
ctu
rer
AS
Ap
ow
er1
fram
esP
erio
dof
use
Cam
eras
Colo
ur
posi
tive
Kod
akA
eroch
rom
eII
(2447
2448
SO
-131S
O-3
68
)32
40ndash50
513
8196
8ndash19
85
Hass
elbla
dL
inh
of
Kod
akE
kta
chro
me
(5017
502
56
017
6118
SO
-121
64
50
113
932
196
5ndash19
68
Hass
elbla
dL
inh
of
Role
iex
SO
-217
Q
X-8
24
QX
-868
)19
81ndash1993
Mau
rer
Nik
on
Kod
akL
um
iere
(5046
5048
)100
105
743
199
4ndash19
98
Hass
elbla
dL
inho
fK
od
akE
lite
100
S(5
069
)100
41
486
199
6ndashp
rese
nt
Hass
elbla
d2
Fu
jiV
elvia
50C
S135
-36
32
13
882
1992ndash19
97H
ass
elbla
dN
iko
nK
od
akH
i-R
esolu
tio
nC
olo
r(S
O-2
42S
O-3
56
)10
096
74
1973ndash19
75H
ass
elbla
d
S19
0A
S190
B
Infr
are
dand
bla
ck-a
nd-w
hit
e
lms
Kod
akA
eroch
rom
eC
olo
rIn
frar
ed40
32
40
704
196
5ndashp
rese
nt2
Hass
elbla
dL
inh
of
Nik
on
S190
A
(8443
34
432443
)S
190B
Kod
akB
ampW
Infr
are
d(2
424
)32
10
553
197
3ndash19
74
S190
AK
od
akP
anato
mic
-XB
ampW
(SO
-022
)63
10
657
197
3ndash19
74
S190
A
1A
tlo
wco
ntr
ast
(ob
ject
back
grou
nd
rati
o16
1)
aspro
vid
edby
Ko
dak
and
list
edby
Sla
ter
etal
(1
983
256
)R
esolv
ing
pow
erw
as
dis
con
tin
ued
fro
mth
ete
chn
ical
test
sper
form
edb
yK
odak
inapp
roxim
atel
y19
93
an
dit
isn
ot
kn
ow
nif
curr
ent
lm
sh
ave
sim
ilar
reso
lvin
gpo
wer
too
lder
ver
sion
so
fsi
milar
lm
s(K
T
eite
lbaum
K
od
akp
erso
nal
com
mu
nic
atio
n)
Th
egra
nu
lari
tym
easu
reno
wpro
vid
edb
yK
od
ak
can
no
tb
ere
adily
con
ver
ted
toa
spat
ial
reso
luti
on
2 The
Lin
hof
was
use
dag
ain
on
ST
S-1
03in
Dec
emb
er19
99(d
ata
not
cata
logued
asof
the
pre
para
tion
of
this
tab
le)
usi
ng
Ko
dak
Elite
100S
lm
3 B
egin
nin
gin
July
1999
2443
colo
ur-
infr
are
d
lmpro
cess
was
chan
ged
from
EA
-5to
E-6
Astronaut photography as digital data for remote sensing 4417
413 Film duplicationThe original lm is archived permanently after producing about 20 duplicate
printing masters (second generation) Duplicate printing masters are disseminatedto regional archives and used to produce products (thirdndashfourth generation) for themedia the public and for scienti c use When prints slides transparencies anddigital products are produced for public distribution they are often colour adjustedto correspond to a more realistic look Digital products from second generationcopies can be requested by scienti c users (contact the authors for information onobtaining such products for a particular project) and these are recommended forremote sensing applications Care should be taken that the digital product acquiredfor remote sensing analysis has not been colour adjusted for presentation purposesBased on qualitative observations there is little visible increase in GRD in the thirdor fourth generation products There is a signi cant degradation in delity of colourin third and fourth generation duplicates because an increase in contrast occurswith every copy
42 Shutter speedsThe impact of camera motion (both due to the photographer and to the motion
of the spacecraft ) on image quality is determined by shutter speedmdash1250 to 1500second have been used because slower speeds record obvious blurring caused bythe rapid motion of the spacecraft relative to the surface of the Earth Generally1250 was used for ISO 64 lms (and slower) and after the switch to ISO 100 lmsa 1500 setting became standard New Hasselblad cameras that have been ownbeginning with STS-92 in October 2000 vary shutter speed using a focal plane shutterfor exposure bracketing (rather than varying aperture) and 1250 1500 and 11000second are used
Across an altitude range of 120ndash300 nautical miles (222ndash555 km) and orbitalinclinations of 285deg and 570deg the median relative ground velocity of orbitingspacecraft is 73 km s Otilde 1 At a nominal shutter speed of 1500 the expected blur dueto motion relative to the ground would be 146 m When cameras are handheld (asopposed to mounted in a bracket) blur can be reduced by physical compensationfor the motion (tracking) by the photographer many photographs show a level ofdetail indicating that blur at this level did not occur (eg gure 3(d )) Thus motionrelative to the ground is not an absolute barrier to ground resolution for handheldphotographs
43 Spacecraft windowsMost of the photographs in the NASA Astronaut Photography Database were
taken through window ports on the spacecraft used The transmittance of the windowport may be aVected by material inhomogeniety of the glass the number of layersof panes coatings used the quality of the surface polish environmentally inducedchanges in window materials (pressure loads thermal gradients) or deposited con-tamination Such degradation of the window cannot be corrected is diVerent foreach window and changes over time See Eppler et al (1996) and Scott (2000) fordiscussion of the spectral transmittance of the fused quartz window that is part ofthe US Laboratory Module (Destiny) of the International Space Station
44 Digitized images from lmWhen lm is scanned digitally the amount of information retained depends on
the spectral information extracted from the lm at the spatial limits of the scanner
J A Robinson et al4418
To date standard digitizing methodologies for astronaut photographs have not beenestablished and lm is digitized on a case-by-case basis using the equipment available
441 Digitizing and spatial resolutionLight (1993 1996) provided equations for determining the digitizing spatial
resolution needed to preserve spatial resolution of aerial photography based on thestatic resolving power (AWAR) for the system For lms where the manufacturer hasprovided data (Eastman Kodak 1998 K Teitelbaum personal communication) the resolving power of lms used for astronaut photography ranges from 32 lp mm Otilde 1to 100 lp mm Otilde 1 at low contrast (objectbackground ratio 161 table 3) The AWARfor the static case of the Hasselblad camera has been measured at high and lowcontrast (using lenses shown in table 1) with a maximum of approximately55 lp mm Otilde 1 (Fred Pearce unpublished data)
Based on the method of Light (1993 1996) the dimension of one spatial resolutionelement for a photograph with maximum static AWAR of 55 lp mm Otilde 1 would be18 mm lp Otilde 1 The acceptable range of spot size to preserve spatial information wouldthen be
18 mm
2 atilde 2aring scan spot size aring
18 mm
2(1)
and 6 mm aring scan spot size aring 9 mm Similarly for a more typical static AWAR of30 lp mm Otilde 1 (33 mm lpOtilde 1 low contrast Fred Pearce unpublished data) 11 mm aring scanspot sizearing 17 mm These scan spot sizes correspond to digitizing resolutions rangingfrom 4233ndash2822 ppi (pixels inch Otilde 1 ) for AWAR of 55 lp mm Otilde 1 and 2309ndash1494 ppi forAWAR of 30 lp mm Otilde 1 Scan spot sizes calculated for astronaut photography arecomparable to those calculated for the National Aerial Photography Program(9ndash13 mm Light 1996)
Widely available scanners that can digitize colour transparency lm currentlyhave a maximum spatial resolution of approximately 2400 ppi (106 mm pixel Otilde 1) Forexample we routinely use an Agfa Arcus II desktop scanner (see also Baltsavias1996) with 2400 ppi digitizing spatial resolution (2400 ppi optical resolution in onedirection and 1200 ppi interpolated to 2400 ppi in the other direction) and 2400 ppiwas used to calculate IFOV equivalents (tables 2 and 4) For some combinations oflens camera and contrast 2400 ppi will capture nearly all of the spatial informationcontained in the lm However for lm with higher resolving power thanEktachrome-64 for better lenses and for higher contrast targets digitizing at 2400 ppiwill not capture all of the spatial information in the lm
Improvements in digitizing technology will only produce real increases in IFOVto the limit of the AWAR of the photography system The incremental increase inspatial information above 3000 ppi (85 mm pixel Otilde 1 ) is not likely to outweigh thecosts of storing large images (Luman et al 1997) At 2400 ppi a Hasselblad frameis approximately 5200times5200 or 27 million pixels (table 2) while the same imagedigitized at 4000 ppi would contain 75 million pixels
Initial studies using astronaut photographs digitized at 2400 ppi (106 mm pixel Otilde 1 Webb et al 2000 Robinson et al 2000c) indicate that some GRD is lost comparedto photographic products Nevertheless the spatial resolution is still comparablewith other widely used data sources (Webb et al 2000) Robinson et al (2000c)found that digitizing at 21 mm pixel Otilde 1 provided information equivalent to 106 mmpixel Otilde 1 for identifying general urban area boundaries for six cities except for a single
Astronaut photography as digital data for remote sensing 4419
photo that required higher spatial resolution digitizing (it had been taken with ashorter lens and thus had less spatial resolution) Part of the equivalence observedin the urban areas study may be attributable to the fact that the atbed scannerused interpolates from 1200 ppi to 2400 ppi in one direction The appropriate digitiz-ing spatial resolution will in part depend on the scale of features of interest to theresearchers with a maximum upper limit set of a scan spot size of approximately6 mm (4233 ppi)
442 Digitizing and spectral resolutionWhen colour lm is digitized there will be loss of spectral resolution The three
lm emulsion layers (red green blue) have relatively distinct spectral responses butare fused together so that digitizers detect one colour signal and must convert itback into three (red green blue) digital channels Digital scanning is also subject tospectral calibration and reproduction errors Studies using digital astronaut photo-graphs to date have used 8 bits per channel However this is another parameterthat can be controlled when the lm is digitized We do not further address spectralresolution of digitally scanned images in this paper
45 External factors that in uence GRDAlthough not discussed in detail in this paper factors external to the spacecraft
and camera system (as listed by Moik (1980) for remote sensing in general) alsoimpact GRD These include atmospheric interference (due to scattering attenuationhaze) variable surface illumination (diVerences in terrain slope and orientation) andchange of terrain re ectance with viewing angle (bidirectional re ectance) For astro-naut photographs variable illumination is particularly important because orbits arenot sun-synchronous Photographs are illuminated by diVerent sun angles and imagesof a given location will have colour intensities that vary widely In addition theviewing angle has an eVect on the degree of object-to-backgroun d contrast andatmospheric interference
5 Estimating spatial resolution of astronaut photographsIn order to use astronaut photographs for digital remote sensing it is important
to be able to calculate the equivalent to an IFOVmdashthe ground area represented bya single pixel in a digitized orbital photograph The obliquity of most photographsmeans that pixel lsquosizesrsquo vary at diVerent places in an image Given a ground distanceD represented by a photograph in each direction (horizontal vertical ) an approxi-mate average pixel width (P the equivalent of IFOV) for the entire image can becalculated as follows
P=D
dS(2)
where D is the projected distance on the ground covered by the image in the samedirection as the pixel is measured d is the actual width of the image on the original lm (table 1) and S is the digitizing spatial resolution
Here we present three mathematical formulations for estimating the size of thefootprint or area on the ground covered by the image Example results from theapplication of all three formulations are given in table 5 The rst and simplestcalculation (formulation 1) gives an idea of the maximum spatial resolution attainableat a given altitude of orbit with a given lm format and a perfectly vertical (nadir)
J A Robinson et al4420
view downward Formulation 2 takes into account obliquity by calculating lookangle from the diVerence between the location on the ground represented at thecentre of the photograph and the nadir location of the spacecraft at the time thephotograph was taken ( gure 1) Formulation 3 describes an alternate solution tothe oblique look angle problem using coordinate-system transformations Thisformulation has been implemented in a documented spreadsheet and is availablefor download (httpeoljscnasagovsseopFootprintCalculatorxls) and is in theprocess of being implemented on a Web-based user interface to the AstronautPhotography Database (OYce of Earth Sciences 2000)
Although formulations 2 and 3 account for obliquity for the purposes of calcula-tion they treat the position of the spacecraft and position of the camera as one Inactuality astronauts are generally holding the cameras by hand (although camerasbracketed in the window are also used) and the selection of window position of theastronaut in the window and rotation of the camera relative to the movement ofthe spacecraft are not known Thus calculations using only the photo centre point(PC) and spacecraft nadir point (SN) give a locator ellipse and not the locations ofthe corners of the photograph A locator ellipse describes an estimated area on theground that is likely to be included in a speci c photograph regardless of the rotationof the lm plane about the camerarsquos optical axis ( gure 6)
Estimating the corner positions of the photo requires additional user input of asingle auxiliary pointmdasha location on the image that has a known location on theground Addition of this auxiliary point is an option available to users of thespreadsheet An example of the results of adding an auxiliary point is shown in gure 7 with comparisons of the various calculations in table 5
51 Formulation 1 Footprint for a nadir viewThe simplest way to estimate footprint size is to use the geometry of camera lens
and spacecraft altitude to calculate the scaling relationship between the image in the
Figure 6 Sketch representation of the locator ellipse an estimated area on the ground thatis likely to be included in a speci c photograph regardless of the rotation of the lmplane about the camerarsquos optical axis
Astronaut photography as digital data for remote sensing 4421
Figure 7 An example of the application of the Low Oblique Space Photo FootprintCalculator The photograph (STS093-704-50) of Limmen Bight Australia was takenfrom an altitude of 283 km using the Hasselblad camera with a 250 mm lens Mapused is 11 000 000 Operational Navigational Chart The distances shown equate toan average pixel size of 124 m for this photograph
lm and the area covered on the ground For a perfect nadir view the scalerelationship from the geometry of similar triangles is
d
D=
f
H(3)
where d=original image size D=distance of footprint on the ground f =focallength of lens and H=altitude Once D is known pixel size ( length=width) can becalculated from equation (2) These calculations represent the minimum footprintand minimum pixel size possible for a given camera system altitude and digitisingspatial resolution (table 2) Formulation 1 was used for calculating minimum pixelsizes shown in tables 2 and 4
By assuming digitizing at 2400 ppi (106 mm pixel Otilde 1 ) currently a spatial resolutioncommonly attainable from multipurpose colour transparency scanners (see sect441)this formulation was used to convert area covered to an IFOV equivalent for missionsof diVerent altitudes (table 2) Table 4 provides a comparison of IFOV of imagesfrom various satellites including the equivalent for astronaut photography Valuesin this table were derived by using formulation 1 to estimate the area covered becausea perfect nadir view represents the best possible spatial resolution and smallest eldof view that could be obtained
52 Formulation 2 Footprint for oblique views using simpli ed geometry and thegreat circle distance
A more realistic approach to determining the footprint of the photographaccounts for the fact that the camera is not usually pointing perfectly down at the
J A Robinson et al4422
Table 4 Comparison of pixel sizes (instantaneous elds of view) for satellite remote sensorswith pixel sizes for near vertical viewing photography from spacecraft Note that alldata returned from the automated satellites have the same eld of view while indicatedpixel sizes for astronaut photography are best-case for near vertical look angles See gure 5 for distributions of astronaut photographs of various look angles High-altitude aerial photography is also included for reference
Altitude PixelInstrument (km) width (mtimesm) Bands1
Landsat Multipectral Scanner2 (MSS 1974-) 880ndash940 79times56 4ndash5Landsat Thematic Mapper (TM 1982-) ~705 30times30 7Satellite pour lrsquoObservation de la Terre (SPOT 1986-) ~832
SPOT 1-3 Panchromatic 10times10 1SPOT 1-3 Multispectral (XS) 20times20 3
Astronaut Photography (1961-)Skylab3 (1973-1974 ) ~435
Multispectral Camera (6 camera stations) 17ndash19times17ndash19 3ndash8Earth Terrain Camera (hi-res colour) 875times875 3
Space Shuttle5 (1981-) 222ndash611Hasselblad 70 mm (250mm lens colour) vertical view 9ndash26times9ndash26 3Hasselblad 70 mm (100mm lens colour) vertical view 23ndash65times23ndash65 3Nikon 35 mm (400mm lens colour lm or ESC) vertical 5ndash16times5ndash16 3
High Altitude Aerial Photograph (1100 000) ~12 2times2 3
1Three bands listed for aerial photography obtainable by high-resolution colour digitizingFor high-resolution colour lm at 65 lp mmOtilde 1 (K Teitelbaum Eastman Kodak Co personalcommunication) the maximum information contained would be 3302 ppi
2Data from Simonett (1983Table 1-2)3Calculated as recommended by Jensen (19831596) the dividend of estimated ground
resolution at low contrast given in NASA (1974179-180) and 244Six lters plus colour and colour infrared lm (NASA 19747) 5Extracted from table 2
nadir point The look angle (the angle oV nadir that the camera is pointing) can becalculated trigonometrically by assuming a spherical earth and calculating the dis-tance between the coordinates of SN and PC ( gure 1) using the Great Circledistance haversine solution (Sinnott 1984 Snyder 198730ndash32 Chamberlain 1996)The diVerence between the spacecraft centre and nadir point latitudes Dlat=lat 2 shy lat 1 and the diVerence between the spacecraft centre and nadir pointlongitudes Dlon=lon 2 shy lon 1 enter the following equations
a=sin2ADlat
2 B+cos(lat 1) cos(lat 2) sin2ADlon
2 B (4)
c=2 arcsin(min[1 atilde a]) (5)
and oVset=R c with R the radius of a spherical Earth=6370 kmThe look angle (t ) is then given by
t=arctanAoffset
H B (6)
Assuming that the camera was positioned so that the imaginary line between thecentre and nadir points (the principal line) runs vertically through the centre of thephotograph the distance between the geometric centre of the photograph (principalpoint PP) and the top of the photograph is d2 ( gure 8(a)) The scale at any point
Astronaut photography as digital data for remote sensing 4423
(a) (b)
Figure 8 The geometric relationship between look angle lens focal length and the centrepoint and nadir points of a photograph (see also Estes and Simonett 1975) Note thatthe Earthrsquos surface and footprint represented in gure 1 are not shown here only arepresentation of the image area of the photograph Variables shown in the gurelook angle t focal length f PP centre of the image on the lm SN spacecraft nadird original image size (a) The special case where the camera is aligned squarely withthe isometric parallel In this case tilt occurs along a plane parallel with the centre ofthe photograph When the conditions are met calculations using Formulation 3 willgive the ground coordinates of the corner points of the photograph (b) The generalrelationship between these parameters and key photogrammetric points on the photo-graph For this more typical case coordinates of an additional point in the photographmust be input in order to adjust for twist of the camera and to nd the corner pointcoordinates
in the photograph varies as a function of the distance y along the principal linebetween the isocentre and the point according to the relationship
d
D=
f shy ysint
H(7)
(Wong 1980 equation (214) H amp the ground elevation h) Using gure 8(a) at thetop of the photo
y= f tanAt
2B+d
2(8)
and at the bottom of the photo
y= f tanAt
2Bshyd
2(9)
Thus for given PC and SN coordinates and assuming a photo orientation as in gure 8(a) and not gure 8(b) we can estimate a minimum D (using equations (7)and (8)) and a maximum D (using equations (7) and (9)) and then average the twoto determine the pixel size (P ) via equation (2)
J A Robinson et al4424
53 Formulation 3 T he L ow Oblique Space Photo Footprint CalculatorThe Low Oblique Space Photo Footprint Calculator was developed to provide
a more accurate estimation of the geographic coordinates for the footprint of a lowoblique photo of the Earthrsquos surface taken from a human-occupied spacecraft inorbit The calculator performs a series of three-dimensional coordinate transforma-tions to compute the location and orientation of the centre of the photo exposureplane relative to an earth referenced coordinate system The nominal camera focallength is then used to create a vector from the photorsquos perspective point througheach of eight points around the parameter of the image as de ned by the formatsize The geographical coordinates for the photo footprint are then computed byintersecting these photo vectors with a spherical earth model Although more sophist-icated projection algorithms are available no signi cant increase in the accuracy ofthe results would be produced by these algorithms due to inherent uncertainties inthe available input data (ie the spacecraft altitude photo centre location etc)
The calculations were initially implemented within a Microsoft Excel workbookwhich allowed one to embed the mathematical and graphical documentation nextto the actual calculations Thus interested users can inspect the mathematical pro-cessing A set of error traps was also built into the calculations to detect erroneousresults A summary of any errors generated is reported to the user with the resultsof the calculations Interested individuals are invited to download the Excel work-book from httpeoljscnasagovsseopFootprintCalculatorxls The calculations arecurrently being encoded in a high-level programming language and should soon beavailable alongside other background data provided for each photograph at OYceof Earth Sciences (2000)
For the purposes of these calculations a low oblique photograph was de ned as
one with the centre within 10 degrees of latitude and longitude of the spacecraftnadir point For the typical range of spacecraft altitudes to date this restricted thecalculations to photographs in which Earthrsquos horizon does not appear (the generalde nition of a low oblique photograph eg Campbell 199671 and gure 4)
531 Input data and resultsUpon opening the Excel workbook the user is presented with a program intro-
duction providing instructions for using the calculator The second worksheet tablsquoHow-To-Usersquo provides detailed step-by-step instructions for preparing the baselinedata for the program The third tab lsquoInput-Output rsquo contains the user input eldsand displays the results of the calculations The additional worksheets contain theactual calculations and program documentation Although users are welcome toreview these sheets an experienced user need only access the lsquoInput-Outputrsquo
spreadsheetTo begin a calculation the user enters the following information which is available
for each photo in the NASA Astronaut Photography Database (1) SN geographicalcoordinates of spacecraft nadir position at the time of photo (2) H spacecraftaltitude (3) PC the geographical coordinates of the centre of the photo (4) f nominal focal length and (5) d image format size The automatic implementationof the workbook on the web will automatically enter these values and completecalculations
For more accurate results the user may optionally enter the geographic co-ordinates and orientation for an auxiliary point on the photo which resolves thecamerarsquos rotation uncertainty about the optical axis The auxiliary point data must
Astronaut photography as digital data for remote sensing 4425
be computed by the user following the instruction contained in the lsquoHow-To-Usersquotab of the spreadsheet
After entering the input data the geographic coordinates of the photo footprint(ie four photo corner points four points at the bisector of each edge and the centreof the photo) are immediately displayed below the input elds along with any errormessages generated by the user input or by the calculations ( gure 7) Although
results are computed and displayed they should not be used when error messagesare produced by the program The program also computes the tilt angle for each ofthe photo vectors relative to the spacecraft nadir vector To the right of the photofootprint coordinates is displayed the arc distance along the surface of the sphere
between adjacent computed points
532 Calculation assumptions
The mathematical calculations implemented in the Low Oblique Space PhotoFootprint Calculator use the following assumptions
1 The SN location is used as exact even though the true value may vary by upto plusmn01 degree from the location provided with the photo
2 The spacecraft altitude is used as exact Although the determination of the
nadir point at the instant of a known spacecraft vector is relatively precise(plusmn115times10 Otilde 4 degrees) the propagator interpolates between sets of approxi-mately 10ndash40 known vectors per day and the time code recorded on the lmcan drift Thus the true value for SN may vary by up to plusmn01 degree fromthe value provided with the photo
3 The perspective centre of the camera is assumed to be at the given altitudeover the speci ed spacecraft nadir location at the time of photo exposure
4 The PC location is used as exact even though the true value may vary by upto plusmn05deg latitude and plusmn05deg longitude from the location provided with the
photo5 A spherical earth model is used with a nominal radius of 6 372 16154 m (a
common rst-order approximation for a spherical earth used in geodeticcomputations)
6 The nominal lens focal length of the camera lens is used in the computations(calibrated focal length values are not available)
7 The photo projection is based on the classic pin-hole camera model8 No correction for lens distortion or atmospheric re ection is made
9 If no auxiliary point data is provided the lsquoTop of the Imagersquo is orientedperpendicular to the vector from SN towards PC
533 T ransformation from Earth to photo coordinate systemsThe calculations begin by converting the geographic coordinates ( latitude and
longitude) of the SN and PC to a Rectangular Earth-Centred Coordinate System(R-Earth) de ned as shown in gure 9 (with the centre of the Earth at (0 0 0))
Using the vector from the Earthrsquos centre through SN and the spacecraft altitude thespacecraft location (SC) is also computed in R-Earth
For ease of computation a Rectangular Spacecraft-Centred Coordinate System(R-Spacecraft ) is de ned as shown in gure 9 The origin of R-Spacecraft is located
at SC with its +Z-axis aligned with vector from the centre of the Earth throughSN and its +X-axis aligned with the vector from SN to PC ( gure 9) The speci c
J A Robinson et al4426
Figure 9 Representation of the area included in a photograph as a geometric projectiononto the Earthrsquos surface Note that the focal length (the distance from the SpaceShuttle to the photo) is grossly overscale compared to the altitude (the distance fromthe Shuttle to the surface of the Earth
rotations and translation used to convert from the R-Earth to the R-Spacecraft arecomputed and documented in the spreadsheet
With the mathematical positions of the SC SN PC and the centre of the Earthcomputed in R-Spacecraft the program next computes the location of the camerarsquosprincipal point (PP) The principle point is the point of intersection of the opticalaxis of the lens with the image plane (ie the lm) It is nominally positioned at a
Astronaut photography as digital data for remote sensing 4427
distance equal to the focal length from the perspective centre of the camera (whichis assumed to be at SC) along the vector from PC through SC as shown in gure 9
A third coordinate system the Rectangular Photo Coordinate System (R-Photo)is created with its origin at PP its XndashY axial plane normal to the vector from PCthrough SC and its +X axis aligned with the +X axis of R-Spacecraft as shownin gure 9 The XndashY plane of this coordinate system represents the image plane ofthe photograph
534 Auxiliary point calculationsThe calculations above employ a critical assumption that all photos are taken
with the lsquotop of the imagersquo oriented perpendicular to the vector from the SN towardsPC as shown in gure 9 To avoid non-uniform solar heating of the external surfacemost orbiting spacecraft are slowly and continually rotated about one or more axesIn this condition a ight crew member taking a photo while oating in microgravitycould orient the photo with practically any orientation relative to the horizon (seealso gure 8(b)) Unfortunately since these photos are taken with conventionalhandheld cameras there is no other information available which can be used toresolve the photorsquos rotational ambiguity about the optical axis other then the photoitself This is why the above assumption is used and the footprint computed by thiscalculator is actually a lsquolocator ellipsersquo which estimates the area on the ground whatis likely to be included in a speci c photograph (see gure 6) This locator ellipse ismost accurate for square image formats and subject to additional distortion as thephotograph format becomes more rectangular
If the user wants a more precise calculation of the image footprint the photorsquosrotational ambiguity about the optical axis must be resolved This can be done inthe calculator by adding data for an auxiliary point Detailed instructions regardinghow to prepare and use auxiliary point data in the computations are included in thelsquoHow-To-Usersquo tab of the spreadsheet Basically the user determines which side ofthe photograph is top and then measures the angle between the line from PP to thetop of the photo and from PP to the auxiliary point on the photo ( gure 9)
If the user includes data for an auxiliary point a series of computations arecompleted to resolve the photo rotation ambiguity about the optical axis (ie the
+Z axis in R-Photo) A vector from the Auxiliary Point on the Earth (AE) throughthe photograph perspective centre ( located at SC) is intersected with the photo imageplane (XndashY plane of R-Photo) to compute the coordinates of the Auxiliary Point onthe photo (AP) in R-Photo A two-dimensional angle in the XndashY plane of R-Photofrom the shy X axis to a line from PP to AP is calculated as shown in gure 9 The
shy X axis is used as the origin of the angle since it represents the top of the photoonce it passes through the perspective centre The diVerence between the computedangle and the angle measured by the user on the photo resolves the ambiguity inthe rotation of the photo relative to the principal line ( gures 7 and 9) Thetransformations from R-Spacecraft and R-Photo are then modi ed to include anadditional rotation angle about the +Z axis in R-Photo
535 lsquoFootprint rsquo calculationsThe program next computes the coordinates of eight points about the perimeter
of the image format (ie located at the four photo corners plus a bisector pointalong each edge of the image) These points are identi ed in R-Photo based uponthe photograph format size and then converted to R-Spacecraft Since all computa-tions are done in orthogonal coordinate systems the R-Spacecraft to R-Photo
J A Robinson et al4428
rotation matrix is transposed to produce an R-Photo to R-Spacecraft rotation matrixOnce in R-Spacecraft a unit vector from each of the eight perimeter points throughthe photo perspective centre (the same point as SC) is computed This provides thecoordinates for points about the perimeter of the image format with their directionvectors in a common coordinate system with other key points needed to computethe photo footprint
The next step is to compute the point of intersection between the spherical earthmodel and each of the eight perimeter point vectors The scalar value for eachperimeter point unit vector is computed using two-dimensional planar trigonometryAn angle c is computed using the formula for the cosine of an angle between twoincluded vectors (the perimeter point unit vector and the vector from SC to thecentre of the Earth) Angle y is computed using the Law of Sines Angle e=180degrees shy c shy y The scalar value of the perimeter point vector is computed using eand the Law of Cosines The scalar value is then multiplied by the perimeter pointunit vector to produce the three-dimensional point of intersection of the vector withEarthrsquos surface in R-Spacecraft The process is repeated independently for each ofthe eight perimeter point vectors Aside from its mathematical simplicity the valueof arcsine (computed in step 2) will exceed the normal range for the sin of an anglewhen the perimeter point vectors fail to intersect with the surface of the earth Asimple test based on this principle allows the program to correctly handle obliquephotos which image a portion of the horizon (see results for high oblique photographsin table 5)
The nal step in the process converts the eight earth intersection points from theR-Spacecraft to R-Earth The results are then converted to the standard geographiccoordinate system and displayed on the lsquoInput-Outputrsquo page of the spreadsheet
54 ExamplesAll three formulations were applied to the photographs included in this paper
and the results are compared in table 5 Formulation 1 gives too small a value fordistance across the photograph (D) for all but the most nadir shots and thus servesas an indicator of the best theoretical case but is not a good measure for a speci cphotograph For example the photograph of Lake Eyre taken from 276 km altitude( gure 2(a)) and the photograph of Limmen Bight ( gure 7) were closest to beingnadir views (oVsetslt68 km or tlt15deg table 5) For these photographs D calculatedusing Formulation 1 was similar to the minimum D calculated using Formulations2 and 3 For almost all other more oblique photographs Formulation 1 gave asigni cant underestimate of the distance covered in the photograph For gure 5(A)(the picture of Houston taken with a 40 mm lens) Formulation 1 did not give anunderestimate for D This is because Formulation 1 does not account for curvatureof the Earth in any way With this large eld of view assuming a at Earth in atedthe value of D above the minimum from calculations that included Earth curvature
A major diVerence between Formulations 2 and 3 is the ability to estimate pixelsizes (P) in both directions (along the principal line and perpendicular to the principalline) For the more oblique photographs the vertical estimate of D and pixel sizes ismuch larger than in the horizontal direction (eg the low oblique and high obliquephotographs of Hawaii gure 4 table 5)
For the area of Limmen Bight ( gure 7) table 5 illustrates the improvement inthe estimate of distance and pixel size that can be obtained by re-estimating thelocation of the PC with greater accuracy Centre points in the catalogued data are
Astronaut photography as digital data for remote sensing 4429
plusmn05deg of latitude and longitude When the centre point was re-estimated to plusmn002degwe determined that the photograph was not taken as obliquely as rst thought(change in the estimate of the look angle t from 169 to 128deg table 5) When theauxiliary point was added to the calculations of Formula 3 the calculated look angleshrank further to 110deg indicating that this photograph was taken at very close toa nadir view Of course this improvement in accuracy could have also led to estimatesof greater obliquity and corresponding larger pixel sizes
We also tested the performance of the scale calculator with auxiliary point byestimating the corner and point locations on the photograph using a 11 000 000Operational Navigational Chart For this test we estimated our ability to readcoordinates from the map as plusmn002degand our error in nding the locations of thecorner points as plusmn015deg (this error varies among photographs depending on thedetail that can be matched between photo and map) For Limmen Bight ( gure 7)the mean diVerence between map estimates and calculator estimates for four pointswas 031deg (SD=018 n=8) For a photograph of San Francisco Bay (STS062-151-291) the mean diVerence between map estimates and calculator estimates forfour points was 0064deg (SD=018 n=8) For a photograph of San Francisco Bay(STS062-151-291) the mean diVerence between map estimates and calculator estim-ates for eight points was 0196deg (SD=0146 n=16) Thus in one case the calculatorestimates were better than our estimate of the error in locating corner points on themap It is reasonable to expect that for nadir-viewing photographs the calculatorused with an auxiliary point can estimate locations of the edges of a photograph towithin plusmn03deg
55 Empirical con rmation of spatial resolution estimatesAs stated previously a challenge to estimating system-AWAR for astronaut
photography of Earth is the lack of suitable targets Small features in an image cansometimes be used as a check on the size of objects that can be successfully resolvedgiving an approximate value for GRD Similarly the number of pixels that make upthose features in the digitized image can be used to make an independent calculationof pixel size We have successfully used features such as roads and airport runwaysto make estimates of spatial scale and resolution (eg Robinson et al 2000c) Whilerecognizing that the use of linear detail in an image is a poor approximation to abar target and that linear objects smaller than the resolving power can often bedetected (Charman 1965) few objects other than roads could be found to make anydirect estimates of GRD Thus roads and runways were used in the images ofHouston (where one can readily conduct ground veri cations and where a numberof higher-contrast concrete roadways were available) to obtain empirical estimatesof GRD and pixel size for comparison with table 5
In the all-digital ESC image of Houston ( gure 3(d )) we examined EllingtonField runway 4-22 (centre left of the image) which is 24384 mtimes457 m This runwayis approximately 6ndash7 pixels in width and 304ndash3094 pixels in length so pixelsrepresent an distance on the ground 7ndash8 m Using a lower contrast measure of astreet length between two intersections (2125 m=211 pixels) a pixel width of 101 mis estimated These results compare favourably with the minimum estimate of 81 mpixels using Formulation 1 (table 5) For an estimate of GRD that would be morecomparable to aerial photography of a line target the smallest street without treecover that could be clearly distinguished on the photograph was 792 m wide Thesmallest non-street object (a gap between stages of the Saturn rocket on display in
J A Robinson et al4430
Tab
le5
App
lica
tio
nof
met
hod
sfo
res
tim
ati
ng
spati
alre
solu
tio
no
fast
ron
aut
ph
oto
gra
ph
sin
clu
din
gF
orm
ula
tion
s12
and
3Im
pro
ved
cen
tre
po
int
esti
mat
esan
dauxilia
ryp
oin
tca
lcu
lati
ons
are
sho
wn
for
gure
7V
aria
ble
sas
use
din
the
text
fle
ns
foca
lle
ngt
hH
al
titu
de
PC
ph
oto
gra
ph
cen
tre
poin
tSN
sp
ace
craft
nadir
po
int
D
dis
tance
cover
edo
nth
egr
oun
d
Pd
ista
nce
on
the
grou
nd
repre
sen
ted
by
apix
el
OVse
td
ista
nce
bet
wee
nP
Cand
SNusi
ng
the
grea
tci
rcle
form
ula
t
look
angle
away
from
SN
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
2L
ake
Eyre
ST
S093
-717
-80
250
276
shy29
0137
0shy
28
413
71
607
11
767
413
7609
ndash64
212
013
760
9ndash64
7121
124
ST
S082
-754
-61
250
617
shy28
5137
0shy
28
113
50
135
726
1201
18
013
8ndash14
827
517
9138ndash15
3277
294
Fig
ure
3H
ou
sto
nS
TS
062
-97-
151
4029
829
5shy
95
530
5shy
95
4410
78
8112
20
5348ndash58
990
1205
351
ndash63
8951
10
36
ST
S094
-737
-28
100
294
295
shy
95
528
3shy
95
5162
31
1133
24
415
8ndash20
334
724
3158ndash20
7351
390
ST
S055
-71-
41250
302
295
shy
95
028
2shy
93
766
412
8192
32
5736
ndash84
715
232
273
9ndash85
7154
185
S73
E51
45300
2685
295
shy
95
024
78
0F
igu
re4
Haw
aii
ST
S069
-701
-65
250
370
195
shy
155
520
4shy
153
881
415
7204
28
9876
ndash98
917
928
688
0ndash10
9181
210
ST
S069
-701
-70
250
370
210
shy
157
519
2shy
150
781
415
7738
63
414
9ndash23
336
760
7148ndash55
6394
10
63
ST
S069
-729
-27
100
398
200
shy
156
620
5shy
148
2219
42
1867
65
432
8ndash13
10
158
62
4323+
311
N
A
Astronaut photography as digital data for remote sensing 4431
Tab
le5
(Cont
inued
)
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
7L
imm
enB
igh
tS
TS
093
-704
-50
250
283
shy15
0135
0shy
15
013
58
623
120
859
16
963
0ndash67
3125
16
863
0ndash67
612
613
2P
CR
e-es
tim
ated
shy14
733
1352
67
64
5128
623
ndash65
5123
128
623
ndash65
712
3127
Auxilia
ryP
oin
t611
062
3ndash65
112
312
5F
igu
re10
N
ear
Au
stin
T
XS
TS
095
-716
-46
250
543
30
5shy
975
28
6shy
1007
120
230
375
34
613
5ndash157
281
34
1136ndash16
128
636
0
1 All
pho
togr
aphs
exce
pt
on
eta
ken
wit
hH
ass
elbla
dca
mer
aso
dis
tan
ceacr
oss
the
image
d=
55m
m
for
S73
E51
45
taken
wit
hel
ectr
onic
still
cam
erad=
276
5m
mho
rizo
nta
l2 P
Cand
SN
are
giv
enin
the
form
(lati
tud
elo
ngit
ude)
deg
rees
w
ith
neg
ativ
en
um
ber
sre
pre
senti
ng
south
and
wes
tA
ccura
cyo
fP
Cis
plusmn0
5and
SN
isplusmn
01
as
reco
rded
inth
eA
stro
naut
Ph
oto
gra
phy
Data
base
ex
cep
tw
her
en
ote
dth
atce
ntr
ep
oin
tw
asre
-est
imate
dw
ith
grea
ter
accu
racy
3 C
alc
ula
ted
usi
ng
the
mea
nofth
em
inim
um
an
dm
axim
um
esti
mate
sof
DF
orm
ula
tion
2co
nsi
der
sD
inth
ed
irec
tion
per
pen
dic
ula
rto
the
Pri
nci
pal
Lin
eo
nly
4 C
alc
ula
ted
inho
rizo
nta
lan
dve
rtic
aldir
ecti
on
sb
ut
still
ass
um
ing
geom
etry
of
gure
6(B
)F
irst
nu
mb
eris
the
mea
no
fth
em
inim
um
Dat
the
bott
om
of
the
ph
oto
and
the
maxi
mum
Dat
the
top
of
the
ph
oto
(range
show
nin
the
tab
le)
and
isco
mpara
ble
toF
orm
ula
tio
n2
Sec
ond
num
ber
isth
em
ean
of
Des
tim
ated
alon
gth
eP
rin
cip
alline
and
alo
ng
the
ou
tsid
eed
ge
(range
not
show
n)
5 Alt
itu
de
isap
pro
xim
ate
for
mis
sion
as
exac
tti
me
pho
togr
ap
hw
asta
ken
and
corr
esp
on
din
gn
adir
data
are
no
tava
ilab
le
Lack
of
nad
irdata
pre
ven
tsap
plica
tion
of
Form
ula
tion
s2
an
d3
6 Au
xilliar
ypo
int
add
edw
as
shy14
80
lati
tud
e13
51
2lo
ngit
ude
shy46
5degro
tati
on
angle
in
stru
ctio
ns
for
esti
mati
ng
the
rota
tio
nan
gle
para
met
erare
conta
ined
as
par
tof
the
scal
eca
lcu
lato
rsp
read
shee
tdo
cum
enta
tio
n
J A Robinson et al4432
a park at Johnson Space Center) that could clearly be distinguished on the photo-graph was 853 m wide
For the photograph of Houston taken with a 250 mm lens ( gure 3(C)) anddigitized from second generation lm at 2400 ppi (106 mm pixel Otilde 1 ) Ellington Fieldrunway 4-22 is 3 pixels in width and 1613 pixels in length so pixels represent151ndash152 m on the ground These results compare favourably with the minimumestimate of 152 m using Formulation 2 and 154ndash185 m pixels using Formulation 3(table 5) For an estimate of GRD using an 8times8 inch print (1369 enlargement) and4times magni cation the smallest street that could clearly be distinguished was 822 mwide the same feature could barely be distinguished on the digitized image
We also made an empirical estimate of spatial resolution for lower contrastvegetation boundaries By clearing forest so that a pattern would be visible to landingaircraft a landowner outside Austin Texas (see also aerial photo in Lisheron 2000)created a target that is also useful for evaluating spatial resolution of astronautphotographs The forest was selectively cleared in order to spell the landownerrsquosname lsquoLUECKErsquo with the remaining trees ( gure 10) According to local surveyorswho planned the clearing the plan was to create letters that were 3100 fttimes1700 ft(9449 mtimes5182 m) Photographed at a high altitude relative to most Shuttle missions(543 km) with a 250 mm lens Formula 3 predicts that each pixel would represent anarea 286 mtimes360 m on the ground (table 5) When original lm was digitized at2400 ppi (106 mm pixel Otilde 1 ) letters correspond to 294times188 pixels for a comparablepixel size of 27ndash32 m
6 Summary and conclusionsAstronaut photographs can be an excellent source of data for remote sensing
applications Best-case resolutions are similar to that for Landsat or SPOT withpixels as small as lt10 m It was estimated that of images taken to date 504 hadlens and obliquity characteristics that make them potential candidates for remotesensing information Digitized astronaut photographs can be overlaid with othersatellite data using GIS (Eckardt et al 2000) or used to ll in gaps in time serieswhen other imagery is not available As a source of public-domain information theycan be very useful for scientists who do not have access to satellite imagery eitherbecause of the expense of image acquisition (to the end user) or the computersystems needed for processing satellite images These diVerences in image acquisitioncosts (to the user) are summarized in table 6
Searching of the complete database of NASA astronaut photography includinglow-spatial-resolution browse images is available via the Web (OYce of EarthSciences 2000) This provides nearly global access for identifying images that willcontribute to a speci c scienti c project For the most detailed studies digitalproducts posted to the web will not be of suYcient quality To date we have madeit a practice of digitizing small numbers of images when requested by scientists atno charge Such requests (including a description of the project involved) can bemade through the authors or using the contact information listed on the websiteInvestigators with needs for larger numbers of digital images have also been servedthrough collaborative agreements
In our experience scientists that nd the data most valuable are those in develop-ing countries those studying areas that have not usually been targets for the majorsatellites those having diYculty in nding low-cloud images those interested inconstructing time series or those interested in using a large number of images
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
Astronaut photography as digital data for remote sensing 4407
23 Imagery from scanners versus lm from camerasLike aerial photography systems the GRD of early remote sensing scanners
(electro-optical imaging systems) was calibrated empirically ( gures 1ndash6 in Simonett1983) but such methods are impractical for routine applications Satellites are nowcommon platforms for collecting remote sensing data small-scale aerial photographsare widely available within the USA and Europe and astronaut photographshave become available worldwide Straightforward comparison of resolutions ofphotographs to IFOVs of scanner images is necessary for many applications
How can spatial resolutions of data from diVerent sources be compared usingcommon units Oft-quoted lsquoresolution numbersrsquo actually IFOV of digital scanningsystems should be at least doubled for comparison to the standard method ofexpressing photographi c spatial resolution as GRD simply because of the objectcontrast requirement The following rule of thumb has been used in geographicapplications to compare spatial resolution of scanners and aerial photography (Welch1982 Jensen 1983) a low-contrast target may be converted to an approximate IFOVby dividing the GRD (of the photograph) by 24 Such a rule of thumb might alsobe reasonably applied to astronaut photography making it possible to convertobserved GRD to IFOV and IFOV to an estimated GRD
In this paper we discuss the spatial resolution of astronaut photographs in termsof system properties and compute the area on the ground represented by a singlepixel the equivalent to IFOV To complete our treatment of spatial resolution wealso provide estimates of the sizes of small features identi able in images for severalcases where this can be readily done
3 Factors that determine the footprint (area covered by the photograph )The most fundamental metric that forms the basis for estimating spatial resolution
of astronaut photographs is the size of the footprint or area on the ground capturedin a photograph (see review of satellite photogrammetry by Light 1980) The basicgeometric variables that in uence the area covered by an astronaut photograph are(1) the altitude of the orbit H (2) the focal length of the lens f (3) the actual sizeof the image on the lm d and (4) the orientation of the camera axis relative to theground (the obliquity or look angle t ) The relationships among these parametersare illustrated in gure 1 (also see equation (3) in sect5)
31 AltitudeHuman space ight missions have had a variety of primary objectives that required
diVerent orbital altitudes The higher the altitude the larger the footprint of thephotographs The diVering scale of photographs taken at diVerent altitudes is illus-trated in gure 2 The two photographs of Lake Eyre Australia were taken with thesame camera and lens but on diVerent dates and from diVerent altitudes In (a) thelake is relatively ooded while in (b) it is dry A 23timesdiVerence in altitude leads toa corresponding diVerence in the scales of the resulting photographs
32 L ensesThe longer the lens focal length the more magni cation the greater detail and
the smaller footprint A variety of lenses with diVerent focal lengths are own oneach space mission (table 1) The eVect of lens length on spatial coverage and imagedetail is shown in gure 3 In these views of Houston (a)ndash(c) taken from approxi-mately similar altitudes with the same camera most of the diVerence in scale of the
J A Robinson et al4408
Figure 1 Geometric relationship between look angle (t) spacecraft nadir point (SN) photo-graph centre point (PC) and photograph principal point (PP) The dark shaded arearepresents Earthrsquos surface The distance covered on the ground D corresponds totable 5 Original image size d is listed for various lms in table 1 Variables correspondto equations (2)ndash(9) in sect5
Astronaut photography as digital data for remote sensing 4409
(a) (b)
Figure 2 Example of the eVect of altitude on area covered in a photograph Both photographsof Lake Eyre Australia were taken using a Hasselblad camera with 100 mm lens(a) altitude 283 km STS093-702-062 (b) altitude 617 km STS082-754-61
photographs is due to the diVerent magni cations of the lenses Although taken froma diVerent altitude and using a diVerent camera with a diVerent original image size(see table 2) an electronic still camera image taken with a 300 mm lens is alsoincluded for comparison
For remote sensing longer focal length lenses are generally preferred (250 mm or350 mm for Hasselblad 300 mm or 400 mm lenses for 35 mm format cameras andESCs) Unfortunately longer-focal-length lenses exhibit poorer performance towardthe edge of a frame For example a 250 mm lens (Distagon CF 56) used on theHasselblad camera has spatial resolution of 57 lp mm Otilde 1 at the centre but only51 lp mm Otilde 1 at the edge and 46 lp mm Otilde 1 at the corner (tested with Ektachrome 5017f8 aperture and high contrast) A longer 300 mm lens used on the Nikon camerahas a greater spatial resolution diVerence between centre (82 lp mm Otilde 1 ) and corner(49 lp mm Otilde 1 ) using Kodak 5017 Ektachrome f4 aperture and high contrast FredPearce (unpublished data) There are tradeoVs among lens optics and speed Forexample the lenses for the Linhof system and the 250 mm lens for the Hasselblad(Distagon CF 56 see footnotes to table 1) are limited to apertures smaller thanf56 This becomes an important constraint in selecting shutter speeds (see discussionof shutter speed in sect42)
33 Cameras and actual image sizesAfter passing through the lens the photographic image is projected onto lm
inside the camera The size of this original image is another important propertydetermining spatial resolution and is determined by the camera used Cameraformats include 35 mm and 70 mm (Lowman 1980 Amsbury 1989) and occasionally5times4 inch (127times100 mm) and larger (table 1) Cameras own on each mission arenot metricmdashthey lack vacuum platens or reseau grids image-motion compensationor gyro-stabilized mounts The workhorse for engineering and Earth photographyon NASA missions has been a series of 70 mm Hasselblad cameras (table 1) chosenfor their reliability The modi ed magazine databack imprints a unique number and
J A Robinson et al4410
Tab
le1
Det
ails
on
the
cam
eras
use
dfo
rast
ron
aut
ph
oto
gra
phy
of
Eart
h
Data
incl
ud
ep
ho
togr
aphy
from
all
NA
SA
hu
man
spac
eig
ht
mis
sion
sth
rough
ST
S-9
6(J
un
e19
99
378
461
tota
lfr
am
esin
the
dat
abas
e)1
Imag
esi
zeT
ypic
al
lense
sN
um
ber
of
Cam
era
make
(mm
timesm
m)
use
d(m
m)
ph
oto
gra
ph
sP
erio
dof
use
No
tes
Has
selb
lad
55times
55
40
50100
2502
288
494
1963ndashp
rese
nt
Lin
ho
f10
5times120
90
250
29
373
1984ndashp
rese
nt
No
won
lyo
wn
occ
asio
nal
lyM
ult
ispec
tral
Cam
era
(S19
0A
)57times
57152
32
913
1973ndash19
74S
kyla
b
xed
-mou
nt3
cam
era
(NA
SA
1974
)E
art
hT
erra
inC
amer
a(S
190B
)11
5times11
5457
5478
1973ndash19
74S
kyla
b
xed
-mou
nt3
cam
era
(NA
SA
1974
)R
ole
iex
55times
55
50
100250
4352
1990ndash19
92L
arg
eF
orm
at
Cam
era4
229
times457
305
2143
1984
Mis
sio
nST
Sndash41G
only
(Mer
kel
etal
198
5)
Maure
r55
times55
76
80818
1961ndash19
68
1 Sm
all-f
orm
atca
mer
as
(35
mm
lm
and
ES
C)
are
no
tin
clud
edin
this
table
bec
ause
of
the
larg
enu
mb
erof
len
ses
and
con
gu
rati
on
sth
at
have
bee
nu
sed
and
bec
au
seth
ese
data
are
no
tcu
rren
tly
incl
uded
inth
edata
bas
efo
rall
mis
sion
s2 L
ense
suse
dfo
rH
ass
elbla
dm
anufa
cture
dby
Carl
Zei
ss
Dis
tago
nC
F4
(40
mm
)D
ista
gon
CF
4(5
0m
m)
Dis
tagon
CF
35
(100
mm
)D
ista
go
nC
F5
6(2
50
mm
)S
tand
ard
Has
selb
lad
len
ses
ow
nw
ill
chan
ge
inth
eyea
r20
00
toD
ista
gon
FE
28
(50
mm
)D
ista
go
nF
E2
(110
mm
)T
ele-
Tes
sar
FE
4(2
50m
m)
and
Tel
e-T
essa
rF
E4
(350
mm
)3 T
hes
eca
mer
as
wer
ein
xed
mou
nti
ngs
on
the
outs
ide
of
the
Skyla
bsp
ace
stati
on
A
ltho
ugh
not
hand
hel
d
the
lm
isca
talo
gued
wit
ho
ther
ast
ron
aut
ph
oto
gra
ph
yan
dis
incl
uded
her
efo
rco
mp
lete
nes
s4 A
NA
SA
-des
igned
cam
era
wit
hatt
itud
ere
fere
nce
syst
emfo
rd
eter
min
ing
pre
cise
nadir
poin
ts
Dat
ais
no
tcu
rren
tly
inco
rpora
ted
into
on
lin
ed
atab
ase
bu
tis
cata
loged
by
Mer
kel
etal
(1
985
)
Astronaut photography as digital data for remote sensing 4411
(a) (b)
(c) (d)
Figure 3 Example of the eVect of lens focal length on resolving power Hasselblad imagesof Houston were all taken from an altitude of 294 kmndash302 km with diVerent lenses(a) 40 mm STS062-97-151 (b) 100 mm STS094-737-28 (c) 250 mm STS055-71-41 Forcomparison an Electronic Still Camera image with 300 mm lens at 269 km altitude isalso shown (d ) S73E5145
timestamp on each frame at the time of exposure Cameras are serviced between ights Occasionally there is enough volume and mass allowance so that a Linhof5times4 inch (127times101 mm) format camera can be own Nikon 35 mm cameras are own routinely also because of proven reliability Electronic still cameras (ESC)were tested for Earth photography beginning in 1992 (Lulla and Holland 1993) Inan ESC a CCD (charge-coupled device) is used as a digital replacement for lmrecording the image projected inside the camera An ESC (consisting of a KodakDCS 460c CCD in a Nikon N-90S body) was added as routine equipment forhandheld photographs in 1995 ESCs have also been operated remotely to captureand downlink Earth images through a NASA-sponsored educational programme(EarthKAM) Discussion of CCD spatial array and radiometric sensitivity arebeyond the scope of this paper but are summarized by Robinson et al (2000b) The
J A Robinson et al4412
format of the lm (or CCD) and image size projected onto the lm (or CCD) aresummarized for all the diVerent cameras own in table 2
34 L ook angle or obliquityNo handheld photographs can be considered perfectly nadir they are taken at a
variety of look angles ranging from near vertical ( looking down at approximatelythe nadir position of the spacecraft ) to high oblique (images that include the curvatureof Earth) Imaging at oblique look angles leads to an image where scale degradesaway from nadir A set of views of the same area from diVerent look angles is shownin gure 4 The rst two shots of the island of Hawaii were taken only a few secondsapart and with the same lens The third photograph was taken on a subsequentorbit and with a shorter lens The curvature of the Earth can be seen in the upperleft corner
Obliquity can be described qualitatively ( gure 4) or quantitatively as the lookangle (t gure 1 calculations described in formulation 2 (sect52)) Because obliquityand look angle have such a dramatic in uence on the footprint we summarize thedatabase characteristics relative to these two parameters Figure 5 is a breakdownof the spatial resolution characteristics of low oblique and near vertical photographsin the NASA Astronaut Photography database Number of photographs are grouped(1) by calculated values for look angle (t ) and (2) by altitude After observing theoverlap between near vertical and low oblique classes we are currently restructuringthis variable (lsquotiltrsquo) in the database to provide a measure of t when available Userswill still be able to do searches based on the qualitative measures but these measureswill be more closely tied to actual look angle
341 Obliquity and georeferencing digitised photographsOften the rst step in a remote sensing analysis of a digitized astronaut photo-
graph is to georeference the data and resample it to conform to a known mapprojection Details and recommendations for resampling astronaut photographydata are provided by Robinson et al (2000a 2000c) and a tutorial is also available(McRay et al 2000) Slightly oblique photographs can be geometrically correctedfor remote sensing purposes but extremely oblique photographs are not suited forgeometric correction When obliquity is too great the spatial scale far away fromnadir is much larger than the spatial scale closer to nadir resampling results areunsuitable because pixels near nadir are lost as the image is resampled while manypixels far away from nadir are excessively replicated by resampling
To avoid the generation of excess pixels during georeferencing the pixel sizes ofthe original digitized image should be smaller than the pixels in the nal resampledimage Calculations of original pixel size using methods presented below can beuseful in ensuring meaningful resampling For slightly oblique images formulation 3(sect53) can be used to estimate pixel sizes at various locations in a photograph (nearnadir and away from nadir) and these pixel sizes then used to determine a reasonablepixel scale following resampling
4 Other characteristics that in uence spatial resolutionBeyond basic geometry many other factors internal to the astronaut photography
system impact the observed ground resolved distance in a photograph The lensesand cameras already discussed are imperfect and introduce radiometric degradations(Moik 1980)Vignetting (slight darkening around the edges of an image) occurs in
Astronaut photography as digital data for remote sensing 4413
Tab
le2
Max
imu
mpo
ssib
led
igit
alsp
atia
lre
solu
tio
n(p
ixel
size
inm
)fo
rca
mer
asy
stem
scu
rren
tly
use
dby
ast
ron
auts
top
ho
togr
aph
eart
hb
ase
do
nth
ege
om
etry
of
alt
itud
ele
ns
and
lm
size
C
alc
ula
tion
sas
sum
eth
atco
lour
tran
spar
ency
lm
isdig
itiz
edat
2400
ppi
(pix
els
per
inch
10
6m
mpix
elOtilde
1 )
Min
imu
mgro
un
ddis
tance
repre
sente
dby
1p
ixel
(m)
As
afu
nct
ion
of
alti
tude1
Imag
esi
zeM
inim
um
alt
itud
eM
edia
nalt
itu
de
Maxi
mum
alt
itud
eC
amer
asy
stem
mm
timesm
mp
ixel
s2(2
22k
m)
(326
km
)(6
11
km
)
Lin
ho
f125
mm
90
mm
lens
105
times120
8978times
11
434
261
383
718
250
mm
lens
94
138
259
Has
selb
lad
70
mm
100
mm
lens
55times
55
5198
times519
823
534
564
725
0m
mle
ns
94
138
259
Nik
on
35m
m30
0m
mle
ns
36times
24
3308
times217
478
115
216
400
mm
lens
59
86
162
ESC
35m
m18
0m
mle
ns3
27
6times
18
530
69times
204
311
116
330
630
0m
mle
ns
67
98
184
400
mm
lens
50
74
138
1 Alt
itud
essh
ow
nar
em
inim
um
m
edia
nan
dm
axim
um
thro
ugh
ST
S-8
9(J
anuary
199
8N
=84
mis
sion
s)
2 Act
ual
pix
els
for
ES
C
oth
erw
ise
assu
med
dig
itiz
ati
on
at240
0pp
i(p
ixel
sp
erin
ch)
equ
alto
10
6m
mpix
elOtilde
1 3 T
he
mo
stco
mm
on
con
gura
tion
for
Eart
hph
oto
gra
ph
yas
part
of
the
Eart
hK
AM
pro
ject
J A Robinson et al4414
Figure 4 De nitions of look angle for photographs taken from low orbit around EarthThe example photographs of Hawaii were all taken from approximately the samealtitude (370ndash398 km) and with the same (250 mm) lens on a Hasselblad camera(STS069-701-69 STS069-701-70 and STS069-729-27 [100 mm lens])
astronaut photography due to light path properties of the lenses used A lack of atness of the lm at the moment of exposure (because cameras used in orbit donot incorporate a vacuum platen) also introduces slight degradations A numberof additional sources of image degradation that would aVect GRD of astronautphotographs are listed
41 Films and processing411 Colour reversal lms
Most lms used in NASA handheld cameras in orbit have been E-6 processcolour reversal lms (Ektachromes) that have a nominal speed of 64 to 100 ASAalthough many other lms have been own (table 3) Reversal lms are used becausedamage from radiation exposure while outside the atmospheric protection of Earthis more noticeable in negative lms than in positive lms (Slater 1996) The choiceof ASA has been to balance the coarser grains ( lower lm resolving power) of high-speed lms with the fact that slower lms require longer exposures and thus areaVected more by vehicle motion (approximately 73 km s Otilde 1 relative to Earthrsquos surfacefor the Space Shuttle) Extremely fast lms (gt400 ASA) are also more susceptibleto radiation damage in orbit (particularly during long-duration or high-altitudemissions Slater 1996) and have not traditionally been used for Earth photography The manufacturer stock numbers identifying the lm used are available for eachimage in the Astronaut Photography Database (OYce of Earth Sciences 2000)
412 Colour infrared lmsColour infrared lm (CIR) was used during an Earth-orbiting Apollo mission
(Colwell 1971) in the multispectral S-190A and the high-resolution S-190B camera
Astronaut photography as digital data for remote sensing 4415
Figure 5 (a) Distributions of astronaut photographs by look angle oV-nadir (calculated fromcentre point and nadir point using Formulation 2) Data included for all photographsfor which centre and nadir points are known with high oblique look angles (n=55 155) excluded (Total sample shown is 108 293 of 382 563 total records in thedatabase when compiled For records where nadir data was not available number ofobservations for qualitative look angles are near vertical 14 086 low oblique 115 834undetermined 89 195) (b) Altitude distribution for astronaut photographs of Earthtaken with the Hasselblad camera Data included for Hasselblad camera only focallength determined with high oblique look angles excluded (Total sample shown is159 046 of 206 971 Hasselblad records with known altitude and of 382 563 total recordsin the database when compiled For Hasselblad records with known altitude data notshown includes high oblique photographs using 100 mm and 250 mm lenses n=20 262and all look angles with other lenses n=27 663)
systems on Skylab (NASA 1974 Wilmarth et al 1977) and occasionally on Shuttlemissions and Shuttle-Mir missions (table 3) The CIR lm used is a three-layerAerochrome having one layer that is sensitive to re ected solar infrared energy toapproximately 900 nm (Eastman Kodak Company 1998) protective coatings onmost spacecraft windows also limit IR transmittance between 800 mm and 1000 nmThis layer is extremely sensitive to the temperature of the lm which createsunpredictable degradation of the IR signature under some space ight conditions
J A Robinson et al4416
Tab
le3
Det
ails
on
lm
suse
dfo
rast
ron
aut
pho
togr
aphy
of
Eart
h
Dat
ain
clud
ep
ho
togr
aphy
fro
mall
NA
SA
hum
an
spac
eig
ht
mis
sio
ns
thro
ugh
ST
S-9
6(J
un
e19
99
378
461
tota
lfr
ames
inth
edata
base
)F
ilm
sw
ith
lt50
00fr
am
esex
po
sed
and
lm
suse
dfo
r35
mm
cam
eras
only
are
no
tin
clud
ed
Res
olv
ing
Nu
mb
erof
Film
man
ufa
ctu
rer
AS
Ap
ow
er1
fram
esP
erio
dof
use
Cam
eras
Colo
ur
posi
tive
Kod
akA
eroch
rom
eII
(2447
2448
SO
-131S
O-3
68
)32
40ndash50
513
8196
8ndash19
85
Hass
elbla
dL
inh
of
Kod
akE
kta
chro
me
(5017
502
56
017
6118
SO
-121
64
50
113
932
196
5ndash19
68
Hass
elbla
dL
inh
of
Role
iex
SO
-217
Q
X-8
24
QX
-868
)19
81ndash1993
Mau
rer
Nik
on
Kod
akL
um
iere
(5046
5048
)100
105
743
199
4ndash19
98
Hass
elbla
dL
inho
fK
od
akE
lite
100
S(5
069
)100
41
486
199
6ndashp
rese
nt
Hass
elbla
d2
Fu
jiV
elvia
50C
S135
-36
32
13
882
1992ndash19
97H
ass
elbla
dN
iko
nK
od
akH
i-R
esolu
tio
nC
olo
r(S
O-2
42S
O-3
56
)10
096
74
1973ndash19
75H
ass
elbla
d
S19
0A
S190
B
Infr
are
dand
bla
ck-a
nd-w
hit
e
lms
Kod
akA
eroch
rom
eC
olo
rIn
frar
ed40
32
40
704
196
5ndashp
rese
nt2
Hass
elbla
dL
inh
of
Nik
on
S190
A
(8443
34
432443
)S
190B
Kod
akB
ampW
Infr
are
d(2
424
)32
10
553
197
3ndash19
74
S190
AK
od
akP
anato
mic
-XB
ampW
(SO
-022
)63
10
657
197
3ndash19
74
S190
A
1A
tlo
wco
ntr
ast
(ob
ject
back
grou
nd
rati
o16
1)
aspro
vid
edby
Ko
dak
and
list
edby
Sla
ter
etal
(1
983
256
)R
esolv
ing
pow
erw
as
dis
con
tin
ued
fro
mth
ete
chn
ical
test
sper
form
edb
yK
odak
inapp
roxim
atel
y19
93
an
dit
isn
ot
kn
ow
nif
curr
ent
lm
sh
ave
sim
ilar
reso
lvin
gpo
wer
too
lder
ver
sion
so
fsi
milar
lm
s(K
T
eite
lbaum
K
od
akp
erso
nal
com
mu
nic
atio
n)
Th
egra
nu
lari
tym
easu
reno
wpro
vid
edb
yK
od
ak
can
no
tb
ere
adily
con
ver
ted
toa
spat
ial
reso
luti
on
2 The
Lin
hof
was
use
dag
ain
on
ST
S-1
03in
Dec
emb
er19
99(d
ata
not
cata
logued
asof
the
pre
para
tion
of
this
tab
le)
usi
ng
Ko
dak
Elite
100S
lm
3 B
egin
nin
gin
July
1999
2443
colo
ur-
infr
are
d
lmpro
cess
was
chan
ged
from
EA
-5to
E-6
Astronaut photography as digital data for remote sensing 4417
413 Film duplicationThe original lm is archived permanently after producing about 20 duplicate
printing masters (second generation) Duplicate printing masters are disseminatedto regional archives and used to produce products (thirdndashfourth generation) for themedia the public and for scienti c use When prints slides transparencies anddigital products are produced for public distribution they are often colour adjustedto correspond to a more realistic look Digital products from second generationcopies can be requested by scienti c users (contact the authors for information onobtaining such products for a particular project) and these are recommended forremote sensing applications Care should be taken that the digital product acquiredfor remote sensing analysis has not been colour adjusted for presentation purposesBased on qualitative observations there is little visible increase in GRD in the thirdor fourth generation products There is a signi cant degradation in delity of colourin third and fourth generation duplicates because an increase in contrast occurswith every copy
42 Shutter speedsThe impact of camera motion (both due to the photographer and to the motion
of the spacecraft ) on image quality is determined by shutter speedmdash1250 to 1500second have been used because slower speeds record obvious blurring caused bythe rapid motion of the spacecraft relative to the surface of the Earth Generally1250 was used for ISO 64 lms (and slower) and after the switch to ISO 100 lmsa 1500 setting became standard New Hasselblad cameras that have been ownbeginning with STS-92 in October 2000 vary shutter speed using a focal plane shutterfor exposure bracketing (rather than varying aperture) and 1250 1500 and 11000second are used
Across an altitude range of 120ndash300 nautical miles (222ndash555 km) and orbitalinclinations of 285deg and 570deg the median relative ground velocity of orbitingspacecraft is 73 km s Otilde 1 At a nominal shutter speed of 1500 the expected blur dueto motion relative to the ground would be 146 m When cameras are handheld (asopposed to mounted in a bracket) blur can be reduced by physical compensationfor the motion (tracking) by the photographer many photographs show a level ofdetail indicating that blur at this level did not occur (eg gure 3(d )) Thus motionrelative to the ground is not an absolute barrier to ground resolution for handheldphotographs
43 Spacecraft windowsMost of the photographs in the NASA Astronaut Photography Database were
taken through window ports on the spacecraft used The transmittance of the windowport may be aVected by material inhomogeniety of the glass the number of layersof panes coatings used the quality of the surface polish environmentally inducedchanges in window materials (pressure loads thermal gradients) or deposited con-tamination Such degradation of the window cannot be corrected is diVerent foreach window and changes over time See Eppler et al (1996) and Scott (2000) fordiscussion of the spectral transmittance of the fused quartz window that is part ofthe US Laboratory Module (Destiny) of the International Space Station
44 Digitized images from lmWhen lm is scanned digitally the amount of information retained depends on
the spectral information extracted from the lm at the spatial limits of the scanner
J A Robinson et al4418
To date standard digitizing methodologies for astronaut photographs have not beenestablished and lm is digitized on a case-by-case basis using the equipment available
441 Digitizing and spatial resolutionLight (1993 1996) provided equations for determining the digitizing spatial
resolution needed to preserve spatial resolution of aerial photography based on thestatic resolving power (AWAR) for the system For lms where the manufacturer hasprovided data (Eastman Kodak 1998 K Teitelbaum personal communication) the resolving power of lms used for astronaut photography ranges from 32 lp mm Otilde 1to 100 lp mm Otilde 1 at low contrast (objectbackground ratio 161 table 3) The AWARfor the static case of the Hasselblad camera has been measured at high and lowcontrast (using lenses shown in table 1) with a maximum of approximately55 lp mm Otilde 1 (Fred Pearce unpublished data)
Based on the method of Light (1993 1996) the dimension of one spatial resolutionelement for a photograph with maximum static AWAR of 55 lp mm Otilde 1 would be18 mm lp Otilde 1 The acceptable range of spot size to preserve spatial information wouldthen be
18 mm
2 atilde 2aring scan spot size aring
18 mm
2(1)
and 6 mm aring scan spot size aring 9 mm Similarly for a more typical static AWAR of30 lp mm Otilde 1 (33 mm lpOtilde 1 low contrast Fred Pearce unpublished data) 11 mm aring scanspot sizearing 17 mm These scan spot sizes correspond to digitizing resolutions rangingfrom 4233ndash2822 ppi (pixels inch Otilde 1 ) for AWAR of 55 lp mm Otilde 1 and 2309ndash1494 ppi forAWAR of 30 lp mm Otilde 1 Scan spot sizes calculated for astronaut photography arecomparable to those calculated for the National Aerial Photography Program(9ndash13 mm Light 1996)
Widely available scanners that can digitize colour transparency lm currentlyhave a maximum spatial resolution of approximately 2400 ppi (106 mm pixel Otilde 1) Forexample we routinely use an Agfa Arcus II desktop scanner (see also Baltsavias1996) with 2400 ppi digitizing spatial resolution (2400 ppi optical resolution in onedirection and 1200 ppi interpolated to 2400 ppi in the other direction) and 2400 ppiwas used to calculate IFOV equivalents (tables 2 and 4) For some combinations oflens camera and contrast 2400 ppi will capture nearly all of the spatial informationcontained in the lm However for lm with higher resolving power thanEktachrome-64 for better lenses and for higher contrast targets digitizing at 2400 ppiwill not capture all of the spatial information in the lm
Improvements in digitizing technology will only produce real increases in IFOVto the limit of the AWAR of the photography system The incremental increase inspatial information above 3000 ppi (85 mm pixel Otilde 1 ) is not likely to outweigh thecosts of storing large images (Luman et al 1997) At 2400 ppi a Hasselblad frameis approximately 5200times5200 or 27 million pixels (table 2) while the same imagedigitized at 4000 ppi would contain 75 million pixels
Initial studies using astronaut photographs digitized at 2400 ppi (106 mm pixel Otilde 1 Webb et al 2000 Robinson et al 2000c) indicate that some GRD is lost comparedto photographic products Nevertheless the spatial resolution is still comparablewith other widely used data sources (Webb et al 2000) Robinson et al (2000c)found that digitizing at 21 mm pixel Otilde 1 provided information equivalent to 106 mmpixel Otilde 1 for identifying general urban area boundaries for six cities except for a single
Astronaut photography as digital data for remote sensing 4419
photo that required higher spatial resolution digitizing (it had been taken with ashorter lens and thus had less spatial resolution) Part of the equivalence observedin the urban areas study may be attributable to the fact that the atbed scannerused interpolates from 1200 ppi to 2400 ppi in one direction The appropriate digitiz-ing spatial resolution will in part depend on the scale of features of interest to theresearchers with a maximum upper limit set of a scan spot size of approximately6 mm (4233 ppi)
442 Digitizing and spectral resolutionWhen colour lm is digitized there will be loss of spectral resolution The three
lm emulsion layers (red green blue) have relatively distinct spectral responses butare fused together so that digitizers detect one colour signal and must convert itback into three (red green blue) digital channels Digital scanning is also subject tospectral calibration and reproduction errors Studies using digital astronaut photo-graphs to date have used 8 bits per channel However this is another parameterthat can be controlled when the lm is digitized We do not further address spectralresolution of digitally scanned images in this paper
45 External factors that in uence GRDAlthough not discussed in detail in this paper factors external to the spacecraft
and camera system (as listed by Moik (1980) for remote sensing in general) alsoimpact GRD These include atmospheric interference (due to scattering attenuationhaze) variable surface illumination (diVerences in terrain slope and orientation) andchange of terrain re ectance with viewing angle (bidirectional re ectance) For astro-naut photographs variable illumination is particularly important because orbits arenot sun-synchronous Photographs are illuminated by diVerent sun angles and imagesof a given location will have colour intensities that vary widely In addition theviewing angle has an eVect on the degree of object-to-backgroun d contrast andatmospheric interference
5 Estimating spatial resolution of astronaut photographsIn order to use astronaut photographs for digital remote sensing it is important
to be able to calculate the equivalent to an IFOVmdashthe ground area represented bya single pixel in a digitized orbital photograph The obliquity of most photographsmeans that pixel lsquosizesrsquo vary at diVerent places in an image Given a ground distanceD represented by a photograph in each direction (horizontal vertical ) an approxi-mate average pixel width (P the equivalent of IFOV) for the entire image can becalculated as follows
P=D
dS(2)
where D is the projected distance on the ground covered by the image in the samedirection as the pixel is measured d is the actual width of the image on the original lm (table 1) and S is the digitizing spatial resolution
Here we present three mathematical formulations for estimating the size of thefootprint or area on the ground covered by the image Example results from theapplication of all three formulations are given in table 5 The rst and simplestcalculation (formulation 1) gives an idea of the maximum spatial resolution attainableat a given altitude of orbit with a given lm format and a perfectly vertical (nadir)
J A Robinson et al4420
view downward Formulation 2 takes into account obliquity by calculating lookangle from the diVerence between the location on the ground represented at thecentre of the photograph and the nadir location of the spacecraft at the time thephotograph was taken ( gure 1) Formulation 3 describes an alternate solution tothe oblique look angle problem using coordinate-system transformations Thisformulation has been implemented in a documented spreadsheet and is availablefor download (httpeoljscnasagovsseopFootprintCalculatorxls) and is in theprocess of being implemented on a Web-based user interface to the AstronautPhotography Database (OYce of Earth Sciences 2000)
Although formulations 2 and 3 account for obliquity for the purposes of calcula-tion they treat the position of the spacecraft and position of the camera as one Inactuality astronauts are generally holding the cameras by hand (although camerasbracketed in the window are also used) and the selection of window position of theastronaut in the window and rotation of the camera relative to the movement ofthe spacecraft are not known Thus calculations using only the photo centre point(PC) and spacecraft nadir point (SN) give a locator ellipse and not the locations ofthe corners of the photograph A locator ellipse describes an estimated area on theground that is likely to be included in a speci c photograph regardless of the rotationof the lm plane about the camerarsquos optical axis ( gure 6)
Estimating the corner positions of the photo requires additional user input of asingle auxiliary pointmdasha location on the image that has a known location on theground Addition of this auxiliary point is an option available to users of thespreadsheet An example of the results of adding an auxiliary point is shown in gure 7 with comparisons of the various calculations in table 5
51 Formulation 1 Footprint for a nadir viewThe simplest way to estimate footprint size is to use the geometry of camera lens
and spacecraft altitude to calculate the scaling relationship between the image in the
Figure 6 Sketch representation of the locator ellipse an estimated area on the ground thatis likely to be included in a speci c photograph regardless of the rotation of the lmplane about the camerarsquos optical axis
Astronaut photography as digital data for remote sensing 4421
Figure 7 An example of the application of the Low Oblique Space Photo FootprintCalculator The photograph (STS093-704-50) of Limmen Bight Australia was takenfrom an altitude of 283 km using the Hasselblad camera with a 250 mm lens Mapused is 11 000 000 Operational Navigational Chart The distances shown equate toan average pixel size of 124 m for this photograph
lm and the area covered on the ground For a perfect nadir view the scalerelationship from the geometry of similar triangles is
d
D=
f
H(3)
where d=original image size D=distance of footprint on the ground f =focallength of lens and H=altitude Once D is known pixel size ( length=width) can becalculated from equation (2) These calculations represent the minimum footprintand minimum pixel size possible for a given camera system altitude and digitisingspatial resolution (table 2) Formulation 1 was used for calculating minimum pixelsizes shown in tables 2 and 4
By assuming digitizing at 2400 ppi (106 mm pixel Otilde 1 ) currently a spatial resolutioncommonly attainable from multipurpose colour transparency scanners (see sect441)this formulation was used to convert area covered to an IFOV equivalent for missionsof diVerent altitudes (table 2) Table 4 provides a comparison of IFOV of imagesfrom various satellites including the equivalent for astronaut photography Valuesin this table were derived by using formulation 1 to estimate the area covered becausea perfect nadir view represents the best possible spatial resolution and smallest eldof view that could be obtained
52 Formulation 2 Footprint for oblique views using simpli ed geometry and thegreat circle distance
A more realistic approach to determining the footprint of the photographaccounts for the fact that the camera is not usually pointing perfectly down at the
J A Robinson et al4422
Table 4 Comparison of pixel sizes (instantaneous elds of view) for satellite remote sensorswith pixel sizes for near vertical viewing photography from spacecraft Note that alldata returned from the automated satellites have the same eld of view while indicatedpixel sizes for astronaut photography are best-case for near vertical look angles See gure 5 for distributions of astronaut photographs of various look angles High-altitude aerial photography is also included for reference
Altitude PixelInstrument (km) width (mtimesm) Bands1
Landsat Multipectral Scanner2 (MSS 1974-) 880ndash940 79times56 4ndash5Landsat Thematic Mapper (TM 1982-) ~705 30times30 7Satellite pour lrsquoObservation de la Terre (SPOT 1986-) ~832
SPOT 1-3 Panchromatic 10times10 1SPOT 1-3 Multispectral (XS) 20times20 3
Astronaut Photography (1961-)Skylab3 (1973-1974 ) ~435
Multispectral Camera (6 camera stations) 17ndash19times17ndash19 3ndash8Earth Terrain Camera (hi-res colour) 875times875 3
Space Shuttle5 (1981-) 222ndash611Hasselblad 70 mm (250mm lens colour) vertical view 9ndash26times9ndash26 3Hasselblad 70 mm (100mm lens colour) vertical view 23ndash65times23ndash65 3Nikon 35 mm (400mm lens colour lm or ESC) vertical 5ndash16times5ndash16 3
High Altitude Aerial Photograph (1100 000) ~12 2times2 3
1Three bands listed for aerial photography obtainable by high-resolution colour digitizingFor high-resolution colour lm at 65 lp mmOtilde 1 (K Teitelbaum Eastman Kodak Co personalcommunication) the maximum information contained would be 3302 ppi
2Data from Simonett (1983Table 1-2)3Calculated as recommended by Jensen (19831596) the dividend of estimated ground
resolution at low contrast given in NASA (1974179-180) and 244Six lters plus colour and colour infrared lm (NASA 19747) 5Extracted from table 2
nadir point The look angle (the angle oV nadir that the camera is pointing) can becalculated trigonometrically by assuming a spherical earth and calculating the dis-tance between the coordinates of SN and PC ( gure 1) using the Great Circledistance haversine solution (Sinnott 1984 Snyder 198730ndash32 Chamberlain 1996)The diVerence between the spacecraft centre and nadir point latitudes Dlat=lat 2 shy lat 1 and the diVerence between the spacecraft centre and nadir pointlongitudes Dlon=lon 2 shy lon 1 enter the following equations
a=sin2ADlat
2 B+cos(lat 1) cos(lat 2) sin2ADlon
2 B (4)
c=2 arcsin(min[1 atilde a]) (5)
and oVset=R c with R the radius of a spherical Earth=6370 kmThe look angle (t ) is then given by
t=arctanAoffset
H B (6)
Assuming that the camera was positioned so that the imaginary line between thecentre and nadir points (the principal line) runs vertically through the centre of thephotograph the distance between the geometric centre of the photograph (principalpoint PP) and the top of the photograph is d2 ( gure 8(a)) The scale at any point
Astronaut photography as digital data for remote sensing 4423
(a) (b)
Figure 8 The geometric relationship between look angle lens focal length and the centrepoint and nadir points of a photograph (see also Estes and Simonett 1975) Note thatthe Earthrsquos surface and footprint represented in gure 1 are not shown here only arepresentation of the image area of the photograph Variables shown in the gurelook angle t focal length f PP centre of the image on the lm SN spacecraft nadird original image size (a) The special case where the camera is aligned squarely withthe isometric parallel In this case tilt occurs along a plane parallel with the centre ofthe photograph When the conditions are met calculations using Formulation 3 willgive the ground coordinates of the corner points of the photograph (b) The generalrelationship between these parameters and key photogrammetric points on the photo-graph For this more typical case coordinates of an additional point in the photographmust be input in order to adjust for twist of the camera and to nd the corner pointcoordinates
in the photograph varies as a function of the distance y along the principal linebetween the isocentre and the point according to the relationship
d
D=
f shy ysint
H(7)
(Wong 1980 equation (214) H amp the ground elevation h) Using gure 8(a) at thetop of the photo
y= f tanAt
2B+d
2(8)
and at the bottom of the photo
y= f tanAt
2Bshyd
2(9)
Thus for given PC and SN coordinates and assuming a photo orientation as in gure 8(a) and not gure 8(b) we can estimate a minimum D (using equations (7)and (8)) and a maximum D (using equations (7) and (9)) and then average the twoto determine the pixel size (P ) via equation (2)
J A Robinson et al4424
53 Formulation 3 T he L ow Oblique Space Photo Footprint CalculatorThe Low Oblique Space Photo Footprint Calculator was developed to provide
a more accurate estimation of the geographic coordinates for the footprint of a lowoblique photo of the Earthrsquos surface taken from a human-occupied spacecraft inorbit The calculator performs a series of three-dimensional coordinate transforma-tions to compute the location and orientation of the centre of the photo exposureplane relative to an earth referenced coordinate system The nominal camera focallength is then used to create a vector from the photorsquos perspective point througheach of eight points around the parameter of the image as de ned by the formatsize The geographical coordinates for the photo footprint are then computed byintersecting these photo vectors with a spherical earth model Although more sophist-icated projection algorithms are available no signi cant increase in the accuracy ofthe results would be produced by these algorithms due to inherent uncertainties inthe available input data (ie the spacecraft altitude photo centre location etc)
The calculations were initially implemented within a Microsoft Excel workbookwhich allowed one to embed the mathematical and graphical documentation nextto the actual calculations Thus interested users can inspect the mathematical pro-cessing A set of error traps was also built into the calculations to detect erroneousresults A summary of any errors generated is reported to the user with the resultsof the calculations Interested individuals are invited to download the Excel work-book from httpeoljscnasagovsseopFootprintCalculatorxls The calculations arecurrently being encoded in a high-level programming language and should soon beavailable alongside other background data provided for each photograph at OYceof Earth Sciences (2000)
For the purposes of these calculations a low oblique photograph was de ned as
one with the centre within 10 degrees of latitude and longitude of the spacecraftnadir point For the typical range of spacecraft altitudes to date this restricted thecalculations to photographs in which Earthrsquos horizon does not appear (the generalde nition of a low oblique photograph eg Campbell 199671 and gure 4)
531 Input data and resultsUpon opening the Excel workbook the user is presented with a program intro-
duction providing instructions for using the calculator The second worksheet tablsquoHow-To-Usersquo provides detailed step-by-step instructions for preparing the baselinedata for the program The third tab lsquoInput-Output rsquo contains the user input eldsand displays the results of the calculations The additional worksheets contain theactual calculations and program documentation Although users are welcome toreview these sheets an experienced user need only access the lsquoInput-Outputrsquo
spreadsheetTo begin a calculation the user enters the following information which is available
for each photo in the NASA Astronaut Photography Database (1) SN geographicalcoordinates of spacecraft nadir position at the time of photo (2) H spacecraftaltitude (3) PC the geographical coordinates of the centre of the photo (4) f nominal focal length and (5) d image format size The automatic implementationof the workbook on the web will automatically enter these values and completecalculations
For more accurate results the user may optionally enter the geographic co-ordinates and orientation for an auxiliary point on the photo which resolves thecamerarsquos rotation uncertainty about the optical axis The auxiliary point data must
Astronaut photography as digital data for remote sensing 4425
be computed by the user following the instruction contained in the lsquoHow-To-Usersquotab of the spreadsheet
After entering the input data the geographic coordinates of the photo footprint(ie four photo corner points four points at the bisector of each edge and the centreof the photo) are immediately displayed below the input elds along with any errormessages generated by the user input or by the calculations ( gure 7) Although
results are computed and displayed they should not be used when error messagesare produced by the program The program also computes the tilt angle for each ofthe photo vectors relative to the spacecraft nadir vector To the right of the photofootprint coordinates is displayed the arc distance along the surface of the sphere
between adjacent computed points
532 Calculation assumptions
The mathematical calculations implemented in the Low Oblique Space PhotoFootprint Calculator use the following assumptions
1 The SN location is used as exact even though the true value may vary by upto plusmn01 degree from the location provided with the photo
2 The spacecraft altitude is used as exact Although the determination of the
nadir point at the instant of a known spacecraft vector is relatively precise(plusmn115times10 Otilde 4 degrees) the propagator interpolates between sets of approxi-mately 10ndash40 known vectors per day and the time code recorded on the lmcan drift Thus the true value for SN may vary by up to plusmn01 degree fromthe value provided with the photo
3 The perspective centre of the camera is assumed to be at the given altitudeover the speci ed spacecraft nadir location at the time of photo exposure
4 The PC location is used as exact even though the true value may vary by upto plusmn05deg latitude and plusmn05deg longitude from the location provided with the
photo5 A spherical earth model is used with a nominal radius of 6 372 16154 m (a
common rst-order approximation for a spherical earth used in geodeticcomputations)
6 The nominal lens focal length of the camera lens is used in the computations(calibrated focal length values are not available)
7 The photo projection is based on the classic pin-hole camera model8 No correction for lens distortion or atmospheric re ection is made
9 If no auxiliary point data is provided the lsquoTop of the Imagersquo is orientedperpendicular to the vector from SN towards PC
533 T ransformation from Earth to photo coordinate systemsThe calculations begin by converting the geographic coordinates ( latitude and
longitude) of the SN and PC to a Rectangular Earth-Centred Coordinate System(R-Earth) de ned as shown in gure 9 (with the centre of the Earth at (0 0 0))
Using the vector from the Earthrsquos centre through SN and the spacecraft altitude thespacecraft location (SC) is also computed in R-Earth
For ease of computation a Rectangular Spacecraft-Centred Coordinate System(R-Spacecraft ) is de ned as shown in gure 9 The origin of R-Spacecraft is located
at SC with its +Z-axis aligned with vector from the centre of the Earth throughSN and its +X-axis aligned with the vector from SN to PC ( gure 9) The speci c
J A Robinson et al4426
Figure 9 Representation of the area included in a photograph as a geometric projectiononto the Earthrsquos surface Note that the focal length (the distance from the SpaceShuttle to the photo) is grossly overscale compared to the altitude (the distance fromthe Shuttle to the surface of the Earth
rotations and translation used to convert from the R-Earth to the R-Spacecraft arecomputed and documented in the spreadsheet
With the mathematical positions of the SC SN PC and the centre of the Earthcomputed in R-Spacecraft the program next computes the location of the camerarsquosprincipal point (PP) The principle point is the point of intersection of the opticalaxis of the lens with the image plane (ie the lm) It is nominally positioned at a
Astronaut photography as digital data for remote sensing 4427
distance equal to the focal length from the perspective centre of the camera (whichis assumed to be at SC) along the vector from PC through SC as shown in gure 9
A third coordinate system the Rectangular Photo Coordinate System (R-Photo)is created with its origin at PP its XndashY axial plane normal to the vector from PCthrough SC and its +X axis aligned with the +X axis of R-Spacecraft as shownin gure 9 The XndashY plane of this coordinate system represents the image plane ofthe photograph
534 Auxiliary point calculationsThe calculations above employ a critical assumption that all photos are taken
with the lsquotop of the imagersquo oriented perpendicular to the vector from the SN towardsPC as shown in gure 9 To avoid non-uniform solar heating of the external surfacemost orbiting spacecraft are slowly and continually rotated about one or more axesIn this condition a ight crew member taking a photo while oating in microgravitycould orient the photo with practically any orientation relative to the horizon (seealso gure 8(b)) Unfortunately since these photos are taken with conventionalhandheld cameras there is no other information available which can be used toresolve the photorsquos rotational ambiguity about the optical axis other then the photoitself This is why the above assumption is used and the footprint computed by thiscalculator is actually a lsquolocator ellipsersquo which estimates the area on the ground whatis likely to be included in a speci c photograph (see gure 6) This locator ellipse ismost accurate for square image formats and subject to additional distortion as thephotograph format becomes more rectangular
If the user wants a more precise calculation of the image footprint the photorsquosrotational ambiguity about the optical axis must be resolved This can be done inthe calculator by adding data for an auxiliary point Detailed instructions regardinghow to prepare and use auxiliary point data in the computations are included in thelsquoHow-To-Usersquo tab of the spreadsheet Basically the user determines which side ofthe photograph is top and then measures the angle between the line from PP to thetop of the photo and from PP to the auxiliary point on the photo ( gure 9)
If the user includes data for an auxiliary point a series of computations arecompleted to resolve the photo rotation ambiguity about the optical axis (ie the
+Z axis in R-Photo) A vector from the Auxiliary Point on the Earth (AE) throughthe photograph perspective centre ( located at SC) is intersected with the photo imageplane (XndashY plane of R-Photo) to compute the coordinates of the Auxiliary Point onthe photo (AP) in R-Photo A two-dimensional angle in the XndashY plane of R-Photofrom the shy X axis to a line from PP to AP is calculated as shown in gure 9 The
shy X axis is used as the origin of the angle since it represents the top of the photoonce it passes through the perspective centre The diVerence between the computedangle and the angle measured by the user on the photo resolves the ambiguity inthe rotation of the photo relative to the principal line ( gures 7 and 9) Thetransformations from R-Spacecraft and R-Photo are then modi ed to include anadditional rotation angle about the +Z axis in R-Photo
535 lsquoFootprint rsquo calculationsThe program next computes the coordinates of eight points about the perimeter
of the image format (ie located at the four photo corners plus a bisector pointalong each edge of the image) These points are identi ed in R-Photo based uponthe photograph format size and then converted to R-Spacecraft Since all computa-tions are done in orthogonal coordinate systems the R-Spacecraft to R-Photo
J A Robinson et al4428
rotation matrix is transposed to produce an R-Photo to R-Spacecraft rotation matrixOnce in R-Spacecraft a unit vector from each of the eight perimeter points throughthe photo perspective centre (the same point as SC) is computed This provides thecoordinates for points about the perimeter of the image format with their directionvectors in a common coordinate system with other key points needed to computethe photo footprint
The next step is to compute the point of intersection between the spherical earthmodel and each of the eight perimeter point vectors The scalar value for eachperimeter point unit vector is computed using two-dimensional planar trigonometryAn angle c is computed using the formula for the cosine of an angle between twoincluded vectors (the perimeter point unit vector and the vector from SC to thecentre of the Earth) Angle y is computed using the Law of Sines Angle e=180degrees shy c shy y The scalar value of the perimeter point vector is computed using eand the Law of Cosines The scalar value is then multiplied by the perimeter pointunit vector to produce the three-dimensional point of intersection of the vector withEarthrsquos surface in R-Spacecraft The process is repeated independently for each ofthe eight perimeter point vectors Aside from its mathematical simplicity the valueof arcsine (computed in step 2) will exceed the normal range for the sin of an anglewhen the perimeter point vectors fail to intersect with the surface of the earth Asimple test based on this principle allows the program to correctly handle obliquephotos which image a portion of the horizon (see results for high oblique photographsin table 5)
The nal step in the process converts the eight earth intersection points from theR-Spacecraft to R-Earth The results are then converted to the standard geographiccoordinate system and displayed on the lsquoInput-Outputrsquo page of the spreadsheet
54 ExamplesAll three formulations were applied to the photographs included in this paper
and the results are compared in table 5 Formulation 1 gives too small a value fordistance across the photograph (D) for all but the most nadir shots and thus servesas an indicator of the best theoretical case but is not a good measure for a speci cphotograph For example the photograph of Lake Eyre taken from 276 km altitude( gure 2(a)) and the photograph of Limmen Bight ( gure 7) were closest to beingnadir views (oVsetslt68 km or tlt15deg table 5) For these photographs D calculatedusing Formulation 1 was similar to the minimum D calculated using Formulations2 and 3 For almost all other more oblique photographs Formulation 1 gave asigni cant underestimate of the distance covered in the photograph For gure 5(A)(the picture of Houston taken with a 40 mm lens) Formulation 1 did not give anunderestimate for D This is because Formulation 1 does not account for curvatureof the Earth in any way With this large eld of view assuming a at Earth in atedthe value of D above the minimum from calculations that included Earth curvature
A major diVerence between Formulations 2 and 3 is the ability to estimate pixelsizes (P) in both directions (along the principal line and perpendicular to the principalline) For the more oblique photographs the vertical estimate of D and pixel sizes ismuch larger than in the horizontal direction (eg the low oblique and high obliquephotographs of Hawaii gure 4 table 5)
For the area of Limmen Bight ( gure 7) table 5 illustrates the improvement inthe estimate of distance and pixel size that can be obtained by re-estimating thelocation of the PC with greater accuracy Centre points in the catalogued data are
Astronaut photography as digital data for remote sensing 4429
plusmn05deg of latitude and longitude When the centre point was re-estimated to plusmn002degwe determined that the photograph was not taken as obliquely as rst thought(change in the estimate of the look angle t from 169 to 128deg table 5) When theauxiliary point was added to the calculations of Formula 3 the calculated look angleshrank further to 110deg indicating that this photograph was taken at very close toa nadir view Of course this improvement in accuracy could have also led to estimatesof greater obliquity and corresponding larger pixel sizes
We also tested the performance of the scale calculator with auxiliary point byestimating the corner and point locations on the photograph using a 11 000 000Operational Navigational Chart For this test we estimated our ability to readcoordinates from the map as plusmn002degand our error in nding the locations of thecorner points as plusmn015deg (this error varies among photographs depending on thedetail that can be matched between photo and map) For Limmen Bight ( gure 7)the mean diVerence between map estimates and calculator estimates for four pointswas 031deg (SD=018 n=8) For a photograph of San Francisco Bay (STS062-151-291) the mean diVerence between map estimates and calculator estimates forfour points was 0064deg (SD=018 n=8) For a photograph of San Francisco Bay(STS062-151-291) the mean diVerence between map estimates and calculator estim-ates for eight points was 0196deg (SD=0146 n=16) Thus in one case the calculatorestimates were better than our estimate of the error in locating corner points on themap It is reasonable to expect that for nadir-viewing photographs the calculatorused with an auxiliary point can estimate locations of the edges of a photograph towithin plusmn03deg
55 Empirical con rmation of spatial resolution estimatesAs stated previously a challenge to estimating system-AWAR for astronaut
photography of Earth is the lack of suitable targets Small features in an image cansometimes be used as a check on the size of objects that can be successfully resolvedgiving an approximate value for GRD Similarly the number of pixels that make upthose features in the digitized image can be used to make an independent calculationof pixel size We have successfully used features such as roads and airport runwaysto make estimates of spatial scale and resolution (eg Robinson et al 2000c) Whilerecognizing that the use of linear detail in an image is a poor approximation to abar target and that linear objects smaller than the resolving power can often bedetected (Charman 1965) few objects other than roads could be found to make anydirect estimates of GRD Thus roads and runways were used in the images ofHouston (where one can readily conduct ground veri cations and where a numberof higher-contrast concrete roadways were available) to obtain empirical estimatesof GRD and pixel size for comparison with table 5
In the all-digital ESC image of Houston ( gure 3(d )) we examined EllingtonField runway 4-22 (centre left of the image) which is 24384 mtimes457 m This runwayis approximately 6ndash7 pixels in width and 304ndash3094 pixels in length so pixelsrepresent an distance on the ground 7ndash8 m Using a lower contrast measure of astreet length between two intersections (2125 m=211 pixels) a pixel width of 101 mis estimated These results compare favourably with the minimum estimate of 81 mpixels using Formulation 1 (table 5) For an estimate of GRD that would be morecomparable to aerial photography of a line target the smallest street without treecover that could be clearly distinguished on the photograph was 792 m wide Thesmallest non-street object (a gap between stages of the Saturn rocket on display in
J A Robinson et al4430
Tab
le5
App
lica
tio
nof
met
hod
sfo
res
tim
ati
ng
spati
alre
solu
tio
no
fast
ron
aut
ph
oto
gra
ph
sin
clu
din
gF
orm
ula
tion
s12
and
3Im
pro
ved
cen
tre
po
int
esti
mat
esan
dauxilia
ryp
oin
tca
lcu
lati
ons
are
sho
wn
for
gure
7V
aria
ble
sas
use
din
the
text
fle
ns
foca
lle
ngt
hH
al
titu
de
PC
ph
oto
gra
ph
cen
tre
poin
tSN
sp
ace
craft
nadir
po
int
D
dis
tance
cover
edo
nth
egr
oun
d
Pd
ista
nce
on
the
grou
nd
repre
sen
ted
by
apix
el
OVse
td
ista
nce
bet
wee
nP
Cand
SNusi
ng
the
grea
tci
rcle
form
ula
t
look
angle
away
from
SN
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
2L
ake
Eyre
ST
S093
-717
-80
250
276
shy29
0137
0shy
28
413
71
607
11
767
413
7609
ndash64
212
013
760
9ndash64
7121
124
ST
S082
-754
-61
250
617
shy28
5137
0shy
28
113
50
135
726
1201
18
013
8ndash14
827
517
9138ndash15
3277
294
Fig
ure
3H
ou
sto
nS
TS
062
-97-
151
4029
829
5shy
95
530
5shy
95
4410
78
8112
20
5348ndash58
990
1205
351
ndash63
8951
10
36
ST
S094
-737
-28
100
294
295
shy
95
528
3shy
95
5162
31
1133
24
415
8ndash20
334
724
3158ndash20
7351
390
ST
S055
-71-
41250
302
295
shy
95
028
2shy
93
766
412
8192
32
5736
ndash84
715
232
273
9ndash85
7154
185
S73
E51
45300
2685
295
shy
95
024
78
0F
igu
re4
Haw
aii
ST
S069
-701
-65
250
370
195
shy
155
520
4shy
153
881
415
7204
28
9876
ndash98
917
928
688
0ndash10
9181
210
ST
S069
-701
-70
250
370
210
shy
157
519
2shy
150
781
415
7738
63
414
9ndash23
336
760
7148ndash55
6394
10
63
ST
S069
-729
-27
100
398
200
shy
156
620
5shy
148
2219
42
1867
65
432
8ndash13
10
158
62
4323+
311
N
A
Astronaut photography as digital data for remote sensing 4431
Tab
le5
(Cont
inued
)
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
7L
imm
enB
igh
tS
TS
093
-704
-50
250
283
shy15
0135
0shy
15
013
58
623
120
859
16
963
0ndash67
3125
16
863
0ndash67
612
613
2P
CR
e-es
tim
ated
shy14
733
1352
67
64
5128
623
ndash65
5123
128
623
ndash65
712
3127
Auxilia
ryP
oin
t611
062
3ndash65
112
312
5F
igu
re10
N
ear
Au
stin
T
XS
TS
095
-716
-46
250
543
30
5shy
975
28
6shy
1007
120
230
375
34
613
5ndash157
281
34
1136ndash16
128
636
0
1 All
pho
togr
aphs
exce
pt
on
eta
ken
wit
hH
ass
elbla
dca
mer
aso
dis
tan
ceacr
oss
the
image
d=
55m
m
for
S73
E51
45
taken
wit
hel
ectr
onic
still
cam
erad=
276
5m
mho
rizo
nta
l2 P
Cand
SN
are
giv
enin
the
form
(lati
tud
elo
ngit
ude)
deg
rees
w
ith
neg
ativ
en
um
ber
sre
pre
senti
ng
south
and
wes
tA
ccura
cyo
fP
Cis
plusmn0
5and
SN
isplusmn
01
as
reco
rded
inth
eA
stro
naut
Ph
oto
gra
phy
Data
base
ex
cep
tw
her
en
ote
dth
atce
ntr
ep
oin
tw
asre
-est
imate
dw
ith
grea
ter
accu
racy
3 C
alc
ula
ted
usi
ng
the
mea
nofth
em
inim
um
an
dm
axim
um
esti
mate
sof
DF
orm
ula
tion
2co
nsi
der
sD
inth
ed
irec
tion
per
pen
dic
ula
rto
the
Pri
nci
pal
Lin
eo
nly
4 C
alc
ula
ted
inho
rizo
nta
lan
dve
rtic
aldir
ecti
on
sb
ut
still
ass
um
ing
geom
etry
of
gure
6(B
)F
irst
nu
mb
eris
the
mea
no
fth
em
inim
um
Dat
the
bott
om
of
the
ph
oto
and
the
maxi
mum
Dat
the
top
of
the
ph
oto
(range
show
nin
the
tab
le)
and
isco
mpara
ble
toF
orm
ula
tio
n2
Sec
ond
num
ber
isth
em
ean
of
Des
tim
ated
alon
gth
eP
rin
cip
alline
and
alo
ng
the
ou
tsid
eed
ge
(range
not
show
n)
5 Alt
itu
de
isap
pro
xim
ate
for
mis
sion
as
exac
tti
me
pho
togr
ap
hw
asta
ken
and
corr
esp
on
din
gn
adir
data
are
no
tava
ilab
le
Lack
of
nad
irdata
pre
ven
tsap
plica
tion
of
Form
ula
tion
s2
an
d3
6 Au
xilliar
ypo
int
add
edw
as
shy14
80
lati
tud
e13
51
2lo
ngit
ude
shy46
5degro
tati
on
angle
in
stru
ctio
ns
for
esti
mati
ng
the
rota
tio
nan
gle
para
met
erare
conta
ined
as
par
tof
the
scal
eca
lcu
lato
rsp
read
shee
tdo
cum
enta
tio
n
J A Robinson et al4432
a park at Johnson Space Center) that could clearly be distinguished on the photo-graph was 853 m wide
For the photograph of Houston taken with a 250 mm lens ( gure 3(C)) anddigitized from second generation lm at 2400 ppi (106 mm pixel Otilde 1 ) Ellington Fieldrunway 4-22 is 3 pixels in width and 1613 pixels in length so pixels represent151ndash152 m on the ground These results compare favourably with the minimumestimate of 152 m using Formulation 2 and 154ndash185 m pixels using Formulation 3(table 5) For an estimate of GRD using an 8times8 inch print (1369 enlargement) and4times magni cation the smallest street that could clearly be distinguished was 822 mwide the same feature could barely be distinguished on the digitized image
We also made an empirical estimate of spatial resolution for lower contrastvegetation boundaries By clearing forest so that a pattern would be visible to landingaircraft a landowner outside Austin Texas (see also aerial photo in Lisheron 2000)created a target that is also useful for evaluating spatial resolution of astronautphotographs The forest was selectively cleared in order to spell the landownerrsquosname lsquoLUECKErsquo with the remaining trees ( gure 10) According to local surveyorswho planned the clearing the plan was to create letters that were 3100 fttimes1700 ft(9449 mtimes5182 m) Photographed at a high altitude relative to most Shuttle missions(543 km) with a 250 mm lens Formula 3 predicts that each pixel would represent anarea 286 mtimes360 m on the ground (table 5) When original lm was digitized at2400 ppi (106 mm pixel Otilde 1 ) letters correspond to 294times188 pixels for a comparablepixel size of 27ndash32 m
6 Summary and conclusionsAstronaut photographs can be an excellent source of data for remote sensing
applications Best-case resolutions are similar to that for Landsat or SPOT withpixels as small as lt10 m It was estimated that of images taken to date 504 hadlens and obliquity characteristics that make them potential candidates for remotesensing information Digitized astronaut photographs can be overlaid with othersatellite data using GIS (Eckardt et al 2000) or used to ll in gaps in time serieswhen other imagery is not available As a source of public-domain information theycan be very useful for scientists who do not have access to satellite imagery eitherbecause of the expense of image acquisition (to the end user) or the computersystems needed for processing satellite images These diVerences in image acquisitioncosts (to the user) are summarized in table 6
Searching of the complete database of NASA astronaut photography includinglow-spatial-resolution browse images is available via the Web (OYce of EarthSciences 2000) This provides nearly global access for identifying images that willcontribute to a speci c scienti c project For the most detailed studies digitalproducts posted to the web will not be of suYcient quality To date we have madeit a practice of digitizing small numbers of images when requested by scientists atno charge Such requests (including a description of the project involved) can bemade through the authors or using the contact information listed on the websiteInvestigators with needs for larger numbers of digital images have also been servedthrough collaborative agreements
In our experience scientists that nd the data most valuable are those in develop-ing countries those studying areas that have not usually been targets for the majorsatellites those having diYculty in nding low-cloud images those interested inconstructing time series or those interested in using a large number of images
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
J A Robinson et al4408
Figure 1 Geometric relationship between look angle (t) spacecraft nadir point (SN) photo-graph centre point (PC) and photograph principal point (PP) The dark shaded arearepresents Earthrsquos surface The distance covered on the ground D corresponds totable 5 Original image size d is listed for various lms in table 1 Variables correspondto equations (2)ndash(9) in sect5
Astronaut photography as digital data for remote sensing 4409
(a) (b)
Figure 2 Example of the eVect of altitude on area covered in a photograph Both photographsof Lake Eyre Australia were taken using a Hasselblad camera with 100 mm lens(a) altitude 283 km STS093-702-062 (b) altitude 617 km STS082-754-61
photographs is due to the diVerent magni cations of the lenses Although taken froma diVerent altitude and using a diVerent camera with a diVerent original image size(see table 2) an electronic still camera image taken with a 300 mm lens is alsoincluded for comparison
For remote sensing longer focal length lenses are generally preferred (250 mm or350 mm for Hasselblad 300 mm or 400 mm lenses for 35 mm format cameras andESCs) Unfortunately longer-focal-length lenses exhibit poorer performance towardthe edge of a frame For example a 250 mm lens (Distagon CF 56) used on theHasselblad camera has spatial resolution of 57 lp mm Otilde 1 at the centre but only51 lp mm Otilde 1 at the edge and 46 lp mm Otilde 1 at the corner (tested with Ektachrome 5017f8 aperture and high contrast) A longer 300 mm lens used on the Nikon camerahas a greater spatial resolution diVerence between centre (82 lp mm Otilde 1 ) and corner(49 lp mm Otilde 1 ) using Kodak 5017 Ektachrome f4 aperture and high contrast FredPearce (unpublished data) There are tradeoVs among lens optics and speed Forexample the lenses for the Linhof system and the 250 mm lens for the Hasselblad(Distagon CF 56 see footnotes to table 1) are limited to apertures smaller thanf56 This becomes an important constraint in selecting shutter speeds (see discussionof shutter speed in sect42)
33 Cameras and actual image sizesAfter passing through the lens the photographic image is projected onto lm
inside the camera The size of this original image is another important propertydetermining spatial resolution and is determined by the camera used Cameraformats include 35 mm and 70 mm (Lowman 1980 Amsbury 1989) and occasionally5times4 inch (127times100 mm) and larger (table 1) Cameras own on each mission arenot metricmdashthey lack vacuum platens or reseau grids image-motion compensationor gyro-stabilized mounts The workhorse for engineering and Earth photographyon NASA missions has been a series of 70 mm Hasselblad cameras (table 1) chosenfor their reliability The modi ed magazine databack imprints a unique number and
J A Robinson et al4410
Tab
le1
Det
ails
on
the
cam
eras
use
dfo
rast
ron
aut
ph
oto
gra
phy
of
Eart
h
Data
incl
ud
ep
ho
togr
aphy
from
all
NA
SA
hu
man
spac
eig
ht
mis
sion
sth
rough
ST
S-9
6(J
un
e19
99
378
461
tota
lfr
am
esin
the
dat
abas
e)1
Imag
esi
zeT
ypic
al
lense
sN
um
ber
of
Cam
era
make
(mm
timesm
m)
use
d(m
m)
ph
oto
gra
ph
sP
erio
dof
use
No
tes
Has
selb
lad
55times
55
40
50100
2502
288
494
1963ndashp
rese
nt
Lin
ho
f10
5times120
90
250
29
373
1984ndashp
rese
nt
No
won
lyo
wn
occ
asio
nal
lyM
ult
ispec
tral
Cam
era
(S19
0A
)57times
57152
32
913
1973ndash19
74S
kyla
b
xed
-mou
nt3
cam
era
(NA
SA
1974
)E
art
hT
erra
inC
amer
a(S
190B
)11
5times11
5457
5478
1973ndash19
74S
kyla
b
xed
-mou
nt3
cam
era
(NA
SA
1974
)R
ole
iex
55times
55
50
100250
4352
1990ndash19
92L
arg
eF
orm
at
Cam
era4
229
times457
305
2143
1984
Mis
sio
nST
Sndash41G
only
(Mer
kel
etal
198
5)
Maure
r55
times55
76
80818
1961ndash19
68
1 Sm
all-f
orm
atca
mer
as
(35
mm
lm
and
ES
C)
are
no
tin
clud
edin
this
table
bec
ause
of
the
larg
enu
mb
erof
len
ses
and
con
gu
rati
on
sth
at
have
bee
nu
sed
and
bec
au
seth
ese
data
are
no
tcu
rren
tly
incl
uded
inth
edata
bas
efo
rall
mis
sion
s2 L
ense
suse
dfo
rH
ass
elbla
dm
anufa
cture
dby
Carl
Zei
ss
Dis
tago
nC
F4
(40
mm
)D
ista
gon
CF
4(5
0m
m)
Dis
tagon
CF
35
(100
mm
)D
ista
go
nC
F5
6(2
50
mm
)S
tand
ard
Has
selb
lad
len
ses
ow
nw
ill
chan
ge
inth
eyea
r20
00
toD
ista
gon
FE
28
(50
mm
)D
ista
go
nF
E2
(110
mm
)T
ele-
Tes
sar
FE
4(2
50m
m)
and
Tel
e-T
essa
rF
E4
(350
mm
)3 T
hes
eca
mer
as
wer
ein
xed
mou
nti
ngs
on
the
outs
ide
of
the
Skyla
bsp
ace
stati
on
A
ltho
ugh
not
hand
hel
d
the
lm
isca
talo
gued
wit
ho
ther
ast
ron
aut
ph
oto
gra
ph
yan
dis
incl
uded
her
efo
rco
mp
lete
nes
s4 A
NA
SA
-des
igned
cam
era
wit
hatt
itud
ere
fere
nce
syst
emfo
rd
eter
min
ing
pre
cise
nadir
poin
ts
Dat
ais
no
tcu
rren
tly
inco
rpora
ted
into
on
lin
ed
atab
ase
bu
tis
cata
loged
by
Mer
kel
etal
(1
985
)
Astronaut photography as digital data for remote sensing 4411
(a) (b)
(c) (d)
Figure 3 Example of the eVect of lens focal length on resolving power Hasselblad imagesof Houston were all taken from an altitude of 294 kmndash302 km with diVerent lenses(a) 40 mm STS062-97-151 (b) 100 mm STS094-737-28 (c) 250 mm STS055-71-41 Forcomparison an Electronic Still Camera image with 300 mm lens at 269 km altitude isalso shown (d ) S73E5145
timestamp on each frame at the time of exposure Cameras are serviced between ights Occasionally there is enough volume and mass allowance so that a Linhof5times4 inch (127times101 mm) format camera can be own Nikon 35 mm cameras are own routinely also because of proven reliability Electronic still cameras (ESC)were tested for Earth photography beginning in 1992 (Lulla and Holland 1993) Inan ESC a CCD (charge-coupled device) is used as a digital replacement for lmrecording the image projected inside the camera An ESC (consisting of a KodakDCS 460c CCD in a Nikon N-90S body) was added as routine equipment forhandheld photographs in 1995 ESCs have also been operated remotely to captureand downlink Earth images through a NASA-sponsored educational programme(EarthKAM) Discussion of CCD spatial array and radiometric sensitivity arebeyond the scope of this paper but are summarized by Robinson et al (2000b) The
J A Robinson et al4412
format of the lm (or CCD) and image size projected onto the lm (or CCD) aresummarized for all the diVerent cameras own in table 2
34 L ook angle or obliquityNo handheld photographs can be considered perfectly nadir they are taken at a
variety of look angles ranging from near vertical ( looking down at approximatelythe nadir position of the spacecraft ) to high oblique (images that include the curvatureof Earth) Imaging at oblique look angles leads to an image where scale degradesaway from nadir A set of views of the same area from diVerent look angles is shownin gure 4 The rst two shots of the island of Hawaii were taken only a few secondsapart and with the same lens The third photograph was taken on a subsequentorbit and with a shorter lens The curvature of the Earth can be seen in the upperleft corner
Obliquity can be described qualitatively ( gure 4) or quantitatively as the lookangle (t gure 1 calculations described in formulation 2 (sect52)) Because obliquityand look angle have such a dramatic in uence on the footprint we summarize thedatabase characteristics relative to these two parameters Figure 5 is a breakdownof the spatial resolution characteristics of low oblique and near vertical photographsin the NASA Astronaut Photography database Number of photographs are grouped(1) by calculated values for look angle (t ) and (2) by altitude After observing theoverlap between near vertical and low oblique classes we are currently restructuringthis variable (lsquotiltrsquo) in the database to provide a measure of t when available Userswill still be able to do searches based on the qualitative measures but these measureswill be more closely tied to actual look angle
341 Obliquity and georeferencing digitised photographsOften the rst step in a remote sensing analysis of a digitized astronaut photo-
graph is to georeference the data and resample it to conform to a known mapprojection Details and recommendations for resampling astronaut photographydata are provided by Robinson et al (2000a 2000c) and a tutorial is also available(McRay et al 2000) Slightly oblique photographs can be geometrically correctedfor remote sensing purposes but extremely oblique photographs are not suited forgeometric correction When obliquity is too great the spatial scale far away fromnadir is much larger than the spatial scale closer to nadir resampling results areunsuitable because pixels near nadir are lost as the image is resampled while manypixels far away from nadir are excessively replicated by resampling
To avoid the generation of excess pixels during georeferencing the pixel sizes ofthe original digitized image should be smaller than the pixels in the nal resampledimage Calculations of original pixel size using methods presented below can beuseful in ensuring meaningful resampling For slightly oblique images formulation 3(sect53) can be used to estimate pixel sizes at various locations in a photograph (nearnadir and away from nadir) and these pixel sizes then used to determine a reasonablepixel scale following resampling
4 Other characteristics that in uence spatial resolutionBeyond basic geometry many other factors internal to the astronaut photography
system impact the observed ground resolved distance in a photograph The lensesand cameras already discussed are imperfect and introduce radiometric degradations(Moik 1980)Vignetting (slight darkening around the edges of an image) occurs in
Astronaut photography as digital data for remote sensing 4413
Tab
le2
Max
imu
mpo
ssib
led
igit
alsp
atia
lre
solu
tio
n(p
ixel
size
inm
)fo
rca
mer
asy
stem
scu
rren
tly
use
dby
ast
ron
auts
top
ho
togr
aph
eart
hb
ase
do
nth
ege
om
etry
of
alt
itud
ele
ns
and
lm
size
C
alc
ula
tion
sas
sum
eth
atco
lour
tran
spar
ency
lm
isdig
itiz
edat
2400
ppi
(pix
els
per
inch
10
6m
mpix
elOtilde
1 )
Min
imu
mgro
un
ddis
tance
repre
sente
dby
1p
ixel
(m)
As
afu
nct
ion
of
alti
tude1
Imag
esi
zeM
inim
um
alt
itud
eM
edia
nalt
itu
de
Maxi
mum
alt
itud
eC
amer
asy
stem
mm
timesm
mp
ixel
s2(2
22k
m)
(326
km
)(6
11
km
)
Lin
ho
f125
mm
90
mm
lens
105
times120
8978times
11
434
261
383
718
250
mm
lens
94
138
259
Has
selb
lad
70
mm
100
mm
lens
55times
55
5198
times519
823
534
564
725
0m
mle
ns
94
138
259
Nik
on
35m
m30
0m
mle
ns
36times
24
3308
times217
478
115
216
400
mm
lens
59
86
162
ESC
35m
m18
0m
mle
ns3
27
6times
18
530
69times
204
311
116
330
630
0m
mle
ns
67
98
184
400
mm
lens
50
74
138
1 Alt
itud
essh
ow
nar
em
inim
um
m
edia
nan
dm
axim
um
thro
ugh
ST
S-8
9(J
anuary
199
8N
=84
mis
sion
s)
2 Act
ual
pix
els
for
ES
C
oth
erw
ise
assu
med
dig
itiz
ati
on
at240
0pp
i(p
ixel
sp
erin
ch)
equ
alto
10
6m
mpix
elOtilde
1 3 T
he
mo
stco
mm
on
con
gura
tion
for
Eart
hph
oto
gra
ph
yas
part
of
the
Eart
hK
AM
pro
ject
J A Robinson et al4414
Figure 4 De nitions of look angle for photographs taken from low orbit around EarthThe example photographs of Hawaii were all taken from approximately the samealtitude (370ndash398 km) and with the same (250 mm) lens on a Hasselblad camera(STS069-701-69 STS069-701-70 and STS069-729-27 [100 mm lens])
astronaut photography due to light path properties of the lenses used A lack of atness of the lm at the moment of exposure (because cameras used in orbit donot incorporate a vacuum platen) also introduces slight degradations A numberof additional sources of image degradation that would aVect GRD of astronautphotographs are listed
41 Films and processing411 Colour reversal lms
Most lms used in NASA handheld cameras in orbit have been E-6 processcolour reversal lms (Ektachromes) that have a nominal speed of 64 to 100 ASAalthough many other lms have been own (table 3) Reversal lms are used becausedamage from radiation exposure while outside the atmospheric protection of Earthis more noticeable in negative lms than in positive lms (Slater 1996) The choiceof ASA has been to balance the coarser grains ( lower lm resolving power) of high-speed lms with the fact that slower lms require longer exposures and thus areaVected more by vehicle motion (approximately 73 km s Otilde 1 relative to Earthrsquos surfacefor the Space Shuttle) Extremely fast lms (gt400 ASA) are also more susceptibleto radiation damage in orbit (particularly during long-duration or high-altitudemissions Slater 1996) and have not traditionally been used for Earth photography The manufacturer stock numbers identifying the lm used are available for eachimage in the Astronaut Photography Database (OYce of Earth Sciences 2000)
412 Colour infrared lmsColour infrared lm (CIR) was used during an Earth-orbiting Apollo mission
(Colwell 1971) in the multispectral S-190A and the high-resolution S-190B camera
Astronaut photography as digital data for remote sensing 4415
Figure 5 (a) Distributions of astronaut photographs by look angle oV-nadir (calculated fromcentre point and nadir point using Formulation 2) Data included for all photographsfor which centre and nadir points are known with high oblique look angles (n=55 155) excluded (Total sample shown is 108 293 of 382 563 total records in thedatabase when compiled For records where nadir data was not available number ofobservations for qualitative look angles are near vertical 14 086 low oblique 115 834undetermined 89 195) (b) Altitude distribution for astronaut photographs of Earthtaken with the Hasselblad camera Data included for Hasselblad camera only focallength determined with high oblique look angles excluded (Total sample shown is159 046 of 206 971 Hasselblad records with known altitude and of 382 563 total recordsin the database when compiled For Hasselblad records with known altitude data notshown includes high oblique photographs using 100 mm and 250 mm lenses n=20 262and all look angles with other lenses n=27 663)
systems on Skylab (NASA 1974 Wilmarth et al 1977) and occasionally on Shuttlemissions and Shuttle-Mir missions (table 3) The CIR lm used is a three-layerAerochrome having one layer that is sensitive to re ected solar infrared energy toapproximately 900 nm (Eastman Kodak Company 1998) protective coatings onmost spacecraft windows also limit IR transmittance between 800 mm and 1000 nmThis layer is extremely sensitive to the temperature of the lm which createsunpredictable degradation of the IR signature under some space ight conditions
J A Robinson et al4416
Tab
le3
Det
ails
on
lm
suse
dfo
rast
ron
aut
pho
togr
aphy
of
Eart
h
Dat
ain
clud
ep
ho
togr
aphy
fro
mall
NA
SA
hum
an
spac
eig
ht
mis
sio
ns
thro
ugh
ST
S-9
6(J
un
e19
99
378
461
tota
lfr
ames
inth
edata
base
)F
ilm
sw
ith
lt50
00fr
am
esex
po
sed
and
lm
suse
dfo
r35
mm
cam
eras
only
are
no
tin
clud
ed
Res
olv
ing
Nu
mb
erof
Film
man
ufa
ctu
rer
AS
Ap
ow
er1
fram
esP
erio
dof
use
Cam
eras
Colo
ur
posi
tive
Kod
akA
eroch
rom
eII
(2447
2448
SO
-131S
O-3
68
)32
40ndash50
513
8196
8ndash19
85
Hass
elbla
dL
inh
of
Kod
akE
kta
chro
me
(5017
502
56
017
6118
SO
-121
64
50
113
932
196
5ndash19
68
Hass
elbla
dL
inh
of
Role
iex
SO
-217
Q
X-8
24
QX
-868
)19
81ndash1993
Mau
rer
Nik
on
Kod
akL
um
iere
(5046
5048
)100
105
743
199
4ndash19
98
Hass
elbla
dL
inho
fK
od
akE
lite
100
S(5
069
)100
41
486
199
6ndashp
rese
nt
Hass
elbla
d2
Fu
jiV
elvia
50C
S135
-36
32
13
882
1992ndash19
97H
ass
elbla
dN
iko
nK
od
akH
i-R
esolu
tio
nC
olo
r(S
O-2
42S
O-3
56
)10
096
74
1973ndash19
75H
ass
elbla
d
S19
0A
S190
B
Infr
are
dand
bla
ck-a
nd-w
hit
e
lms
Kod
akA
eroch
rom
eC
olo
rIn
frar
ed40
32
40
704
196
5ndashp
rese
nt2
Hass
elbla
dL
inh
of
Nik
on
S190
A
(8443
34
432443
)S
190B
Kod
akB
ampW
Infr
are
d(2
424
)32
10
553
197
3ndash19
74
S190
AK
od
akP
anato
mic
-XB
ampW
(SO
-022
)63
10
657
197
3ndash19
74
S190
A
1A
tlo
wco
ntr
ast
(ob
ject
back
grou
nd
rati
o16
1)
aspro
vid
edby
Ko
dak
and
list
edby
Sla
ter
etal
(1
983
256
)R
esolv
ing
pow
erw
as
dis
con
tin
ued
fro
mth
ete
chn
ical
test
sper
form
edb
yK
odak
inapp
roxim
atel
y19
93
an
dit
isn
ot
kn
ow
nif
curr
ent
lm
sh
ave
sim
ilar
reso
lvin
gpo
wer
too
lder
ver
sion
so
fsi
milar
lm
s(K
T
eite
lbaum
K
od
akp
erso
nal
com
mu
nic
atio
n)
Th
egra
nu
lari
tym
easu
reno
wpro
vid
edb
yK
od
ak
can
no
tb
ere
adily
con
ver
ted
toa
spat
ial
reso
luti
on
2 The
Lin
hof
was
use
dag
ain
on
ST
S-1
03in
Dec
emb
er19
99(d
ata
not
cata
logued
asof
the
pre
para
tion
of
this
tab
le)
usi
ng
Ko
dak
Elite
100S
lm
3 B
egin
nin
gin
July
1999
2443
colo
ur-
infr
are
d
lmpro
cess
was
chan
ged
from
EA
-5to
E-6
Astronaut photography as digital data for remote sensing 4417
413 Film duplicationThe original lm is archived permanently after producing about 20 duplicate
printing masters (second generation) Duplicate printing masters are disseminatedto regional archives and used to produce products (thirdndashfourth generation) for themedia the public and for scienti c use When prints slides transparencies anddigital products are produced for public distribution they are often colour adjustedto correspond to a more realistic look Digital products from second generationcopies can be requested by scienti c users (contact the authors for information onobtaining such products for a particular project) and these are recommended forremote sensing applications Care should be taken that the digital product acquiredfor remote sensing analysis has not been colour adjusted for presentation purposesBased on qualitative observations there is little visible increase in GRD in the thirdor fourth generation products There is a signi cant degradation in delity of colourin third and fourth generation duplicates because an increase in contrast occurswith every copy
42 Shutter speedsThe impact of camera motion (both due to the photographer and to the motion
of the spacecraft ) on image quality is determined by shutter speedmdash1250 to 1500second have been used because slower speeds record obvious blurring caused bythe rapid motion of the spacecraft relative to the surface of the Earth Generally1250 was used for ISO 64 lms (and slower) and after the switch to ISO 100 lmsa 1500 setting became standard New Hasselblad cameras that have been ownbeginning with STS-92 in October 2000 vary shutter speed using a focal plane shutterfor exposure bracketing (rather than varying aperture) and 1250 1500 and 11000second are used
Across an altitude range of 120ndash300 nautical miles (222ndash555 km) and orbitalinclinations of 285deg and 570deg the median relative ground velocity of orbitingspacecraft is 73 km s Otilde 1 At a nominal shutter speed of 1500 the expected blur dueto motion relative to the ground would be 146 m When cameras are handheld (asopposed to mounted in a bracket) blur can be reduced by physical compensationfor the motion (tracking) by the photographer many photographs show a level ofdetail indicating that blur at this level did not occur (eg gure 3(d )) Thus motionrelative to the ground is not an absolute barrier to ground resolution for handheldphotographs
43 Spacecraft windowsMost of the photographs in the NASA Astronaut Photography Database were
taken through window ports on the spacecraft used The transmittance of the windowport may be aVected by material inhomogeniety of the glass the number of layersof panes coatings used the quality of the surface polish environmentally inducedchanges in window materials (pressure loads thermal gradients) or deposited con-tamination Such degradation of the window cannot be corrected is diVerent foreach window and changes over time See Eppler et al (1996) and Scott (2000) fordiscussion of the spectral transmittance of the fused quartz window that is part ofthe US Laboratory Module (Destiny) of the International Space Station
44 Digitized images from lmWhen lm is scanned digitally the amount of information retained depends on
the spectral information extracted from the lm at the spatial limits of the scanner
J A Robinson et al4418
To date standard digitizing methodologies for astronaut photographs have not beenestablished and lm is digitized on a case-by-case basis using the equipment available
441 Digitizing and spatial resolutionLight (1993 1996) provided equations for determining the digitizing spatial
resolution needed to preserve spatial resolution of aerial photography based on thestatic resolving power (AWAR) for the system For lms where the manufacturer hasprovided data (Eastman Kodak 1998 K Teitelbaum personal communication) the resolving power of lms used for astronaut photography ranges from 32 lp mm Otilde 1to 100 lp mm Otilde 1 at low contrast (objectbackground ratio 161 table 3) The AWARfor the static case of the Hasselblad camera has been measured at high and lowcontrast (using lenses shown in table 1) with a maximum of approximately55 lp mm Otilde 1 (Fred Pearce unpublished data)
Based on the method of Light (1993 1996) the dimension of one spatial resolutionelement for a photograph with maximum static AWAR of 55 lp mm Otilde 1 would be18 mm lp Otilde 1 The acceptable range of spot size to preserve spatial information wouldthen be
18 mm
2 atilde 2aring scan spot size aring
18 mm
2(1)
and 6 mm aring scan spot size aring 9 mm Similarly for a more typical static AWAR of30 lp mm Otilde 1 (33 mm lpOtilde 1 low contrast Fred Pearce unpublished data) 11 mm aring scanspot sizearing 17 mm These scan spot sizes correspond to digitizing resolutions rangingfrom 4233ndash2822 ppi (pixels inch Otilde 1 ) for AWAR of 55 lp mm Otilde 1 and 2309ndash1494 ppi forAWAR of 30 lp mm Otilde 1 Scan spot sizes calculated for astronaut photography arecomparable to those calculated for the National Aerial Photography Program(9ndash13 mm Light 1996)
Widely available scanners that can digitize colour transparency lm currentlyhave a maximum spatial resolution of approximately 2400 ppi (106 mm pixel Otilde 1) Forexample we routinely use an Agfa Arcus II desktop scanner (see also Baltsavias1996) with 2400 ppi digitizing spatial resolution (2400 ppi optical resolution in onedirection and 1200 ppi interpolated to 2400 ppi in the other direction) and 2400 ppiwas used to calculate IFOV equivalents (tables 2 and 4) For some combinations oflens camera and contrast 2400 ppi will capture nearly all of the spatial informationcontained in the lm However for lm with higher resolving power thanEktachrome-64 for better lenses and for higher contrast targets digitizing at 2400 ppiwill not capture all of the spatial information in the lm
Improvements in digitizing technology will only produce real increases in IFOVto the limit of the AWAR of the photography system The incremental increase inspatial information above 3000 ppi (85 mm pixel Otilde 1 ) is not likely to outweigh thecosts of storing large images (Luman et al 1997) At 2400 ppi a Hasselblad frameis approximately 5200times5200 or 27 million pixels (table 2) while the same imagedigitized at 4000 ppi would contain 75 million pixels
Initial studies using astronaut photographs digitized at 2400 ppi (106 mm pixel Otilde 1 Webb et al 2000 Robinson et al 2000c) indicate that some GRD is lost comparedto photographic products Nevertheless the spatial resolution is still comparablewith other widely used data sources (Webb et al 2000) Robinson et al (2000c)found that digitizing at 21 mm pixel Otilde 1 provided information equivalent to 106 mmpixel Otilde 1 for identifying general urban area boundaries for six cities except for a single
Astronaut photography as digital data for remote sensing 4419
photo that required higher spatial resolution digitizing (it had been taken with ashorter lens and thus had less spatial resolution) Part of the equivalence observedin the urban areas study may be attributable to the fact that the atbed scannerused interpolates from 1200 ppi to 2400 ppi in one direction The appropriate digitiz-ing spatial resolution will in part depend on the scale of features of interest to theresearchers with a maximum upper limit set of a scan spot size of approximately6 mm (4233 ppi)
442 Digitizing and spectral resolutionWhen colour lm is digitized there will be loss of spectral resolution The three
lm emulsion layers (red green blue) have relatively distinct spectral responses butare fused together so that digitizers detect one colour signal and must convert itback into three (red green blue) digital channels Digital scanning is also subject tospectral calibration and reproduction errors Studies using digital astronaut photo-graphs to date have used 8 bits per channel However this is another parameterthat can be controlled when the lm is digitized We do not further address spectralresolution of digitally scanned images in this paper
45 External factors that in uence GRDAlthough not discussed in detail in this paper factors external to the spacecraft
and camera system (as listed by Moik (1980) for remote sensing in general) alsoimpact GRD These include atmospheric interference (due to scattering attenuationhaze) variable surface illumination (diVerences in terrain slope and orientation) andchange of terrain re ectance with viewing angle (bidirectional re ectance) For astro-naut photographs variable illumination is particularly important because orbits arenot sun-synchronous Photographs are illuminated by diVerent sun angles and imagesof a given location will have colour intensities that vary widely In addition theviewing angle has an eVect on the degree of object-to-backgroun d contrast andatmospheric interference
5 Estimating spatial resolution of astronaut photographsIn order to use astronaut photographs for digital remote sensing it is important
to be able to calculate the equivalent to an IFOVmdashthe ground area represented bya single pixel in a digitized orbital photograph The obliquity of most photographsmeans that pixel lsquosizesrsquo vary at diVerent places in an image Given a ground distanceD represented by a photograph in each direction (horizontal vertical ) an approxi-mate average pixel width (P the equivalent of IFOV) for the entire image can becalculated as follows
P=D
dS(2)
where D is the projected distance on the ground covered by the image in the samedirection as the pixel is measured d is the actual width of the image on the original lm (table 1) and S is the digitizing spatial resolution
Here we present three mathematical formulations for estimating the size of thefootprint or area on the ground covered by the image Example results from theapplication of all three formulations are given in table 5 The rst and simplestcalculation (formulation 1) gives an idea of the maximum spatial resolution attainableat a given altitude of orbit with a given lm format and a perfectly vertical (nadir)
J A Robinson et al4420
view downward Formulation 2 takes into account obliquity by calculating lookangle from the diVerence between the location on the ground represented at thecentre of the photograph and the nadir location of the spacecraft at the time thephotograph was taken ( gure 1) Formulation 3 describes an alternate solution tothe oblique look angle problem using coordinate-system transformations Thisformulation has been implemented in a documented spreadsheet and is availablefor download (httpeoljscnasagovsseopFootprintCalculatorxls) and is in theprocess of being implemented on a Web-based user interface to the AstronautPhotography Database (OYce of Earth Sciences 2000)
Although formulations 2 and 3 account for obliquity for the purposes of calcula-tion they treat the position of the spacecraft and position of the camera as one Inactuality astronauts are generally holding the cameras by hand (although camerasbracketed in the window are also used) and the selection of window position of theastronaut in the window and rotation of the camera relative to the movement ofthe spacecraft are not known Thus calculations using only the photo centre point(PC) and spacecraft nadir point (SN) give a locator ellipse and not the locations ofthe corners of the photograph A locator ellipse describes an estimated area on theground that is likely to be included in a speci c photograph regardless of the rotationof the lm plane about the camerarsquos optical axis ( gure 6)
Estimating the corner positions of the photo requires additional user input of asingle auxiliary pointmdasha location on the image that has a known location on theground Addition of this auxiliary point is an option available to users of thespreadsheet An example of the results of adding an auxiliary point is shown in gure 7 with comparisons of the various calculations in table 5
51 Formulation 1 Footprint for a nadir viewThe simplest way to estimate footprint size is to use the geometry of camera lens
and spacecraft altitude to calculate the scaling relationship between the image in the
Figure 6 Sketch representation of the locator ellipse an estimated area on the ground thatis likely to be included in a speci c photograph regardless of the rotation of the lmplane about the camerarsquos optical axis
Astronaut photography as digital data for remote sensing 4421
Figure 7 An example of the application of the Low Oblique Space Photo FootprintCalculator The photograph (STS093-704-50) of Limmen Bight Australia was takenfrom an altitude of 283 km using the Hasselblad camera with a 250 mm lens Mapused is 11 000 000 Operational Navigational Chart The distances shown equate toan average pixel size of 124 m for this photograph
lm and the area covered on the ground For a perfect nadir view the scalerelationship from the geometry of similar triangles is
d
D=
f
H(3)
where d=original image size D=distance of footprint on the ground f =focallength of lens and H=altitude Once D is known pixel size ( length=width) can becalculated from equation (2) These calculations represent the minimum footprintand minimum pixel size possible for a given camera system altitude and digitisingspatial resolution (table 2) Formulation 1 was used for calculating minimum pixelsizes shown in tables 2 and 4
By assuming digitizing at 2400 ppi (106 mm pixel Otilde 1 ) currently a spatial resolutioncommonly attainable from multipurpose colour transparency scanners (see sect441)this formulation was used to convert area covered to an IFOV equivalent for missionsof diVerent altitudes (table 2) Table 4 provides a comparison of IFOV of imagesfrom various satellites including the equivalent for astronaut photography Valuesin this table were derived by using formulation 1 to estimate the area covered becausea perfect nadir view represents the best possible spatial resolution and smallest eldof view that could be obtained
52 Formulation 2 Footprint for oblique views using simpli ed geometry and thegreat circle distance
A more realistic approach to determining the footprint of the photographaccounts for the fact that the camera is not usually pointing perfectly down at the
J A Robinson et al4422
Table 4 Comparison of pixel sizes (instantaneous elds of view) for satellite remote sensorswith pixel sizes for near vertical viewing photography from spacecraft Note that alldata returned from the automated satellites have the same eld of view while indicatedpixel sizes for astronaut photography are best-case for near vertical look angles See gure 5 for distributions of astronaut photographs of various look angles High-altitude aerial photography is also included for reference
Altitude PixelInstrument (km) width (mtimesm) Bands1
Landsat Multipectral Scanner2 (MSS 1974-) 880ndash940 79times56 4ndash5Landsat Thematic Mapper (TM 1982-) ~705 30times30 7Satellite pour lrsquoObservation de la Terre (SPOT 1986-) ~832
SPOT 1-3 Panchromatic 10times10 1SPOT 1-3 Multispectral (XS) 20times20 3
Astronaut Photography (1961-)Skylab3 (1973-1974 ) ~435
Multispectral Camera (6 camera stations) 17ndash19times17ndash19 3ndash8Earth Terrain Camera (hi-res colour) 875times875 3
Space Shuttle5 (1981-) 222ndash611Hasselblad 70 mm (250mm lens colour) vertical view 9ndash26times9ndash26 3Hasselblad 70 mm (100mm lens colour) vertical view 23ndash65times23ndash65 3Nikon 35 mm (400mm lens colour lm or ESC) vertical 5ndash16times5ndash16 3
High Altitude Aerial Photograph (1100 000) ~12 2times2 3
1Three bands listed for aerial photography obtainable by high-resolution colour digitizingFor high-resolution colour lm at 65 lp mmOtilde 1 (K Teitelbaum Eastman Kodak Co personalcommunication) the maximum information contained would be 3302 ppi
2Data from Simonett (1983Table 1-2)3Calculated as recommended by Jensen (19831596) the dividend of estimated ground
resolution at low contrast given in NASA (1974179-180) and 244Six lters plus colour and colour infrared lm (NASA 19747) 5Extracted from table 2
nadir point The look angle (the angle oV nadir that the camera is pointing) can becalculated trigonometrically by assuming a spherical earth and calculating the dis-tance between the coordinates of SN and PC ( gure 1) using the Great Circledistance haversine solution (Sinnott 1984 Snyder 198730ndash32 Chamberlain 1996)The diVerence between the spacecraft centre and nadir point latitudes Dlat=lat 2 shy lat 1 and the diVerence between the spacecraft centre and nadir pointlongitudes Dlon=lon 2 shy lon 1 enter the following equations
a=sin2ADlat
2 B+cos(lat 1) cos(lat 2) sin2ADlon
2 B (4)
c=2 arcsin(min[1 atilde a]) (5)
and oVset=R c with R the radius of a spherical Earth=6370 kmThe look angle (t ) is then given by
t=arctanAoffset
H B (6)
Assuming that the camera was positioned so that the imaginary line between thecentre and nadir points (the principal line) runs vertically through the centre of thephotograph the distance between the geometric centre of the photograph (principalpoint PP) and the top of the photograph is d2 ( gure 8(a)) The scale at any point
Astronaut photography as digital data for remote sensing 4423
(a) (b)
Figure 8 The geometric relationship between look angle lens focal length and the centrepoint and nadir points of a photograph (see also Estes and Simonett 1975) Note thatthe Earthrsquos surface and footprint represented in gure 1 are not shown here only arepresentation of the image area of the photograph Variables shown in the gurelook angle t focal length f PP centre of the image on the lm SN spacecraft nadird original image size (a) The special case where the camera is aligned squarely withthe isometric parallel In this case tilt occurs along a plane parallel with the centre ofthe photograph When the conditions are met calculations using Formulation 3 willgive the ground coordinates of the corner points of the photograph (b) The generalrelationship between these parameters and key photogrammetric points on the photo-graph For this more typical case coordinates of an additional point in the photographmust be input in order to adjust for twist of the camera and to nd the corner pointcoordinates
in the photograph varies as a function of the distance y along the principal linebetween the isocentre and the point according to the relationship
d
D=
f shy ysint
H(7)
(Wong 1980 equation (214) H amp the ground elevation h) Using gure 8(a) at thetop of the photo
y= f tanAt
2B+d
2(8)
and at the bottom of the photo
y= f tanAt
2Bshyd
2(9)
Thus for given PC and SN coordinates and assuming a photo orientation as in gure 8(a) and not gure 8(b) we can estimate a minimum D (using equations (7)and (8)) and a maximum D (using equations (7) and (9)) and then average the twoto determine the pixel size (P ) via equation (2)
J A Robinson et al4424
53 Formulation 3 T he L ow Oblique Space Photo Footprint CalculatorThe Low Oblique Space Photo Footprint Calculator was developed to provide
a more accurate estimation of the geographic coordinates for the footprint of a lowoblique photo of the Earthrsquos surface taken from a human-occupied spacecraft inorbit The calculator performs a series of three-dimensional coordinate transforma-tions to compute the location and orientation of the centre of the photo exposureplane relative to an earth referenced coordinate system The nominal camera focallength is then used to create a vector from the photorsquos perspective point througheach of eight points around the parameter of the image as de ned by the formatsize The geographical coordinates for the photo footprint are then computed byintersecting these photo vectors with a spherical earth model Although more sophist-icated projection algorithms are available no signi cant increase in the accuracy ofthe results would be produced by these algorithms due to inherent uncertainties inthe available input data (ie the spacecraft altitude photo centre location etc)
The calculations were initially implemented within a Microsoft Excel workbookwhich allowed one to embed the mathematical and graphical documentation nextto the actual calculations Thus interested users can inspect the mathematical pro-cessing A set of error traps was also built into the calculations to detect erroneousresults A summary of any errors generated is reported to the user with the resultsof the calculations Interested individuals are invited to download the Excel work-book from httpeoljscnasagovsseopFootprintCalculatorxls The calculations arecurrently being encoded in a high-level programming language and should soon beavailable alongside other background data provided for each photograph at OYceof Earth Sciences (2000)
For the purposes of these calculations a low oblique photograph was de ned as
one with the centre within 10 degrees of latitude and longitude of the spacecraftnadir point For the typical range of spacecraft altitudes to date this restricted thecalculations to photographs in which Earthrsquos horizon does not appear (the generalde nition of a low oblique photograph eg Campbell 199671 and gure 4)
531 Input data and resultsUpon opening the Excel workbook the user is presented with a program intro-
duction providing instructions for using the calculator The second worksheet tablsquoHow-To-Usersquo provides detailed step-by-step instructions for preparing the baselinedata for the program The third tab lsquoInput-Output rsquo contains the user input eldsand displays the results of the calculations The additional worksheets contain theactual calculations and program documentation Although users are welcome toreview these sheets an experienced user need only access the lsquoInput-Outputrsquo
spreadsheetTo begin a calculation the user enters the following information which is available
for each photo in the NASA Astronaut Photography Database (1) SN geographicalcoordinates of spacecraft nadir position at the time of photo (2) H spacecraftaltitude (3) PC the geographical coordinates of the centre of the photo (4) f nominal focal length and (5) d image format size The automatic implementationof the workbook on the web will automatically enter these values and completecalculations
For more accurate results the user may optionally enter the geographic co-ordinates and orientation for an auxiliary point on the photo which resolves thecamerarsquos rotation uncertainty about the optical axis The auxiliary point data must
Astronaut photography as digital data for remote sensing 4425
be computed by the user following the instruction contained in the lsquoHow-To-Usersquotab of the spreadsheet
After entering the input data the geographic coordinates of the photo footprint(ie four photo corner points four points at the bisector of each edge and the centreof the photo) are immediately displayed below the input elds along with any errormessages generated by the user input or by the calculations ( gure 7) Although
results are computed and displayed they should not be used when error messagesare produced by the program The program also computes the tilt angle for each ofthe photo vectors relative to the spacecraft nadir vector To the right of the photofootprint coordinates is displayed the arc distance along the surface of the sphere
between adjacent computed points
532 Calculation assumptions
The mathematical calculations implemented in the Low Oblique Space PhotoFootprint Calculator use the following assumptions
1 The SN location is used as exact even though the true value may vary by upto plusmn01 degree from the location provided with the photo
2 The spacecraft altitude is used as exact Although the determination of the
nadir point at the instant of a known spacecraft vector is relatively precise(plusmn115times10 Otilde 4 degrees) the propagator interpolates between sets of approxi-mately 10ndash40 known vectors per day and the time code recorded on the lmcan drift Thus the true value for SN may vary by up to plusmn01 degree fromthe value provided with the photo
3 The perspective centre of the camera is assumed to be at the given altitudeover the speci ed spacecraft nadir location at the time of photo exposure
4 The PC location is used as exact even though the true value may vary by upto plusmn05deg latitude and plusmn05deg longitude from the location provided with the
photo5 A spherical earth model is used with a nominal radius of 6 372 16154 m (a
common rst-order approximation for a spherical earth used in geodeticcomputations)
6 The nominal lens focal length of the camera lens is used in the computations(calibrated focal length values are not available)
7 The photo projection is based on the classic pin-hole camera model8 No correction for lens distortion or atmospheric re ection is made
9 If no auxiliary point data is provided the lsquoTop of the Imagersquo is orientedperpendicular to the vector from SN towards PC
533 T ransformation from Earth to photo coordinate systemsThe calculations begin by converting the geographic coordinates ( latitude and
longitude) of the SN and PC to a Rectangular Earth-Centred Coordinate System(R-Earth) de ned as shown in gure 9 (with the centre of the Earth at (0 0 0))
Using the vector from the Earthrsquos centre through SN and the spacecraft altitude thespacecraft location (SC) is also computed in R-Earth
For ease of computation a Rectangular Spacecraft-Centred Coordinate System(R-Spacecraft ) is de ned as shown in gure 9 The origin of R-Spacecraft is located
at SC with its +Z-axis aligned with vector from the centre of the Earth throughSN and its +X-axis aligned with the vector from SN to PC ( gure 9) The speci c
J A Robinson et al4426
Figure 9 Representation of the area included in a photograph as a geometric projectiononto the Earthrsquos surface Note that the focal length (the distance from the SpaceShuttle to the photo) is grossly overscale compared to the altitude (the distance fromthe Shuttle to the surface of the Earth
rotations and translation used to convert from the R-Earth to the R-Spacecraft arecomputed and documented in the spreadsheet
With the mathematical positions of the SC SN PC and the centre of the Earthcomputed in R-Spacecraft the program next computes the location of the camerarsquosprincipal point (PP) The principle point is the point of intersection of the opticalaxis of the lens with the image plane (ie the lm) It is nominally positioned at a
Astronaut photography as digital data for remote sensing 4427
distance equal to the focal length from the perspective centre of the camera (whichis assumed to be at SC) along the vector from PC through SC as shown in gure 9
A third coordinate system the Rectangular Photo Coordinate System (R-Photo)is created with its origin at PP its XndashY axial plane normal to the vector from PCthrough SC and its +X axis aligned with the +X axis of R-Spacecraft as shownin gure 9 The XndashY plane of this coordinate system represents the image plane ofthe photograph
534 Auxiliary point calculationsThe calculations above employ a critical assumption that all photos are taken
with the lsquotop of the imagersquo oriented perpendicular to the vector from the SN towardsPC as shown in gure 9 To avoid non-uniform solar heating of the external surfacemost orbiting spacecraft are slowly and continually rotated about one or more axesIn this condition a ight crew member taking a photo while oating in microgravitycould orient the photo with practically any orientation relative to the horizon (seealso gure 8(b)) Unfortunately since these photos are taken with conventionalhandheld cameras there is no other information available which can be used toresolve the photorsquos rotational ambiguity about the optical axis other then the photoitself This is why the above assumption is used and the footprint computed by thiscalculator is actually a lsquolocator ellipsersquo which estimates the area on the ground whatis likely to be included in a speci c photograph (see gure 6) This locator ellipse ismost accurate for square image formats and subject to additional distortion as thephotograph format becomes more rectangular
If the user wants a more precise calculation of the image footprint the photorsquosrotational ambiguity about the optical axis must be resolved This can be done inthe calculator by adding data for an auxiliary point Detailed instructions regardinghow to prepare and use auxiliary point data in the computations are included in thelsquoHow-To-Usersquo tab of the spreadsheet Basically the user determines which side ofthe photograph is top and then measures the angle between the line from PP to thetop of the photo and from PP to the auxiliary point on the photo ( gure 9)
If the user includes data for an auxiliary point a series of computations arecompleted to resolve the photo rotation ambiguity about the optical axis (ie the
+Z axis in R-Photo) A vector from the Auxiliary Point on the Earth (AE) throughthe photograph perspective centre ( located at SC) is intersected with the photo imageplane (XndashY plane of R-Photo) to compute the coordinates of the Auxiliary Point onthe photo (AP) in R-Photo A two-dimensional angle in the XndashY plane of R-Photofrom the shy X axis to a line from PP to AP is calculated as shown in gure 9 The
shy X axis is used as the origin of the angle since it represents the top of the photoonce it passes through the perspective centre The diVerence between the computedangle and the angle measured by the user on the photo resolves the ambiguity inthe rotation of the photo relative to the principal line ( gures 7 and 9) Thetransformations from R-Spacecraft and R-Photo are then modi ed to include anadditional rotation angle about the +Z axis in R-Photo
535 lsquoFootprint rsquo calculationsThe program next computes the coordinates of eight points about the perimeter
of the image format (ie located at the four photo corners plus a bisector pointalong each edge of the image) These points are identi ed in R-Photo based uponthe photograph format size and then converted to R-Spacecraft Since all computa-tions are done in orthogonal coordinate systems the R-Spacecraft to R-Photo
J A Robinson et al4428
rotation matrix is transposed to produce an R-Photo to R-Spacecraft rotation matrixOnce in R-Spacecraft a unit vector from each of the eight perimeter points throughthe photo perspective centre (the same point as SC) is computed This provides thecoordinates for points about the perimeter of the image format with their directionvectors in a common coordinate system with other key points needed to computethe photo footprint
The next step is to compute the point of intersection between the spherical earthmodel and each of the eight perimeter point vectors The scalar value for eachperimeter point unit vector is computed using two-dimensional planar trigonometryAn angle c is computed using the formula for the cosine of an angle between twoincluded vectors (the perimeter point unit vector and the vector from SC to thecentre of the Earth) Angle y is computed using the Law of Sines Angle e=180degrees shy c shy y The scalar value of the perimeter point vector is computed using eand the Law of Cosines The scalar value is then multiplied by the perimeter pointunit vector to produce the three-dimensional point of intersection of the vector withEarthrsquos surface in R-Spacecraft The process is repeated independently for each ofthe eight perimeter point vectors Aside from its mathematical simplicity the valueof arcsine (computed in step 2) will exceed the normal range for the sin of an anglewhen the perimeter point vectors fail to intersect with the surface of the earth Asimple test based on this principle allows the program to correctly handle obliquephotos which image a portion of the horizon (see results for high oblique photographsin table 5)
The nal step in the process converts the eight earth intersection points from theR-Spacecraft to R-Earth The results are then converted to the standard geographiccoordinate system and displayed on the lsquoInput-Outputrsquo page of the spreadsheet
54 ExamplesAll three formulations were applied to the photographs included in this paper
and the results are compared in table 5 Formulation 1 gives too small a value fordistance across the photograph (D) for all but the most nadir shots and thus servesas an indicator of the best theoretical case but is not a good measure for a speci cphotograph For example the photograph of Lake Eyre taken from 276 km altitude( gure 2(a)) and the photograph of Limmen Bight ( gure 7) were closest to beingnadir views (oVsetslt68 km or tlt15deg table 5) For these photographs D calculatedusing Formulation 1 was similar to the minimum D calculated using Formulations2 and 3 For almost all other more oblique photographs Formulation 1 gave asigni cant underestimate of the distance covered in the photograph For gure 5(A)(the picture of Houston taken with a 40 mm lens) Formulation 1 did not give anunderestimate for D This is because Formulation 1 does not account for curvatureof the Earth in any way With this large eld of view assuming a at Earth in atedthe value of D above the minimum from calculations that included Earth curvature
A major diVerence between Formulations 2 and 3 is the ability to estimate pixelsizes (P) in both directions (along the principal line and perpendicular to the principalline) For the more oblique photographs the vertical estimate of D and pixel sizes ismuch larger than in the horizontal direction (eg the low oblique and high obliquephotographs of Hawaii gure 4 table 5)
For the area of Limmen Bight ( gure 7) table 5 illustrates the improvement inthe estimate of distance and pixel size that can be obtained by re-estimating thelocation of the PC with greater accuracy Centre points in the catalogued data are
Astronaut photography as digital data for remote sensing 4429
plusmn05deg of latitude and longitude When the centre point was re-estimated to plusmn002degwe determined that the photograph was not taken as obliquely as rst thought(change in the estimate of the look angle t from 169 to 128deg table 5) When theauxiliary point was added to the calculations of Formula 3 the calculated look angleshrank further to 110deg indicating that this photograph was taken at very close toa nadir view Of course this improvement in accuracy could have also led to estimatesof greater obliquity and corresponding larger pixel sizes
We also tested the performance of the scale calculator with auxiliary point byestimating the corner and point locations on the photograph using a 11 000 000Operational Navigational Chart For this test we estimated our ability to readcoordinates from the map as plusmn002degand our error in nding the locations of thecorner points as plusmn015deg (this error varies among photographs depending on thedetail that can be matched between photo and map) For Limmen Bight ( gure 7)the mean diVerence between map estimates and calculator estimates for four pointswas 031deg (SD=018 n=8) For a photograph of San Francisco Bay (STS062-151-291) the mean diVerence between map estimates and calculator estimates forfour points was 0064deg (SD=018 n=8) For a photograph of San Francisco Bay(STS062-151-291) the mean diVerence between map estimates and calculator estim-ates for eight points was 0196deg (SD=0146 n=16) Thus in one case the calculatorestimates were better than our estimate of the error in locating corner points on themap It is reasonable to expect that for nadir-viewing photographs the calculatorused with an auxiliary point can estimate locations of the edges of a photograph towithin plusmn03deg
55 Empirical con rmation of spatial resolution estimatesAs stated previously a challenge to estimating system-AWAR for astronaut
photography of Earth is the lack of suitable targets Small features in an image cansometimes be used as a check on the size of objects that can be successfully resolvedgiving an approximate value for GRD Similarly the number of pixels that make upthose features in the digitized image can be used to make an independent calculationof pixel size We have successfully used features such as roads and airport runwaysto make estimates of spatial scale and resolution (eg Robinson et al 2000c) Whilerecognizing that the use of linear detail in an image is a poor approximation to abar target and that linear objects smaller than the resolving power can often bedetected (Charman 1965) few objects other than roads could be found to make anydirect estimates of GRD Thus roads and runways were used in the images ofHouston (where one can readily conduct ground veri cations and where a numberof higher-contrast concrete roadways were available) to obtain empirical estimatesof GRD and pixel size for comparison with table 5
In the all-digital ESC image of Houston ( gure 3(d )) we examined EllingtonField runway 4-22 (centre left of the image) which is 24384 mtimes457 m This runwayis approximately 6ndash7 pixels in width and 304ndash3094 pixels in length so pixelsrepresent an distance on the ground 7ndash8 m Using a lower contrast measure of astreet length between two intersections (2125 m=211 pixels) a pixel width of 101 mis estimated These results compare favourably with the minimum estimate of 81 mpixels using Formulation 1 (table 5) For an estimate of GRD that would be morecomparable to aerial photography of a line target the smallest street without treecover that could be clearly distinguished on the photograph was 792 m wide Thesmallest non-street object (a gap between stages of the Saturn rocket on display in
J A Robinson et al4430
Tab
le5
App
lica
tio
nof
met
hod
sfo
res
tim
ati
ng
spati
alre
solu
tio
no
fast
ron
aut
ph
oto
gra
ph
sin
clu
din
gF
orm
ula
tion
s12
and
3Im
pro
ved
cen
tre
po
int
esti
mat
esan
dauxilia
ryp
oin
tca
lcu
lati
ons
are
sho
wn
for
gure
7V
aria
ble
sas
use
din
the
text
fle
ns
foca
lle
ngt
hH
al
titu
de
PC
ph
oto
gra
ph
cen
tre
poin
tSN
sp
ace
craft
nadir
po
int
D
dis
tance
cover
edo
nth
egr
oun
d
Pd
ista
nce
on
the
grou
nd
repre
sen
ted
by
apix
el
OVse
td
ista
nce
bet
wee
nP
Cand
SNusi
ng
the
grea
tci
rcle
form
ula
t
look
angle
away
from
SN
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
2L
ake
Eyre
ST
S093
-717
-80
250
276
shy29
0137
0shy
28
413
71
607
11
767
413
7609
ndash64
212
013
760
9ndash64
7121
124
ST
S082
-754
-61
250
617
shy28
5137
0shy
28
113
50
135
726
1201
18
013
8ndash14
827
517
9138ndash15
3277
294
Fig
ure
3H
ou
sto
nS
TS
062
-97-
151
4029
829
5shy
95
530
5shy
95
4410
78
8112
20
5348ndash58
990
1205
351
ndash63
8951
10
36
ST
S094
-737
-28
100
294
295
shy
95
528
3shy
95
5162
31
1133
24
415
8ndash20
334
724
3158ndash20
7351
390
ST
S055
-71-
41250
302
295
shy
95
028
2shy
93
766
412
8192
32
5736
ndash84
715
232
273
9ndash85
7154
185
S73
E51
45300
2685
295
shy
95
024
78
0F
igu
re4
Haw
aii
ST
S069
-701
-65
250
370
195
shy
155
520
4shy
153
881
415
7204
28
9876
ndash98
917
928
688
0ndash10
9181
210
ST
S069
-701
-70
250
370
210
shy
157
519
2shy
150
781
415
7738
63
414
9ndash23
336
760
7148ndash55
6394
10
63
ST
S069
-729
-27
100
398
200
shy
156
620
5shy
148
2219
42
1867
65
432
8ndash13
10
158
62
4323+
311
N
A
Astronaut photography as digital data for remote sensing 4431
Tab
le5
(Cont
inued
)
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
7L
imm
enB
igh
tS
TS
093
-704
-50
250
283
shy15
0135
0shy
15
013
58
623
120
859
16
963
0ndash67
3125
16
863
0ndash67
612
613
2P
CR
e-es
tim
ated
shy14
733
1352
67
64
5128
623
ndash65
5123
128
623
ndash65
712
3127
Auxilia
ryP
oin
t611
062
3ndash65
112
312
5F
igu
re10
N
ear
Au
stin
T
XS
TS
095
-716
-46
250
543
30
5shy
975
28
6shy
1007
120
230
375
34
613
5ndash157
281
34
1136ndash16
128
636
0
1 All
pho
togr
aphs
exce
pt
on
eta
ken
wit
hH
ass
elbla
dca
mer
aso
dis
tan
ceacr
oss
the
image
d=
55m
m
for
S73
E51
45
taken
wit
hel
ectr
onic
still
cam
erad=
276
5m
mho
rizo
nta
l2 P
Cand
SN
are
giv
enin
the
form
(lati
tud
elo
ngit
ude)
deg
rees
w
ith
neg
ativ
en
um
ber
sre
pre
senti
ng
south
and
wes
tA
ccura
cyo
fP
Cis
plusmn0
5and
SN
isplusmn
01
as
reco
rded
inth
eA
stro
naut
Ph
oto
gra
phy
Data
base
ex
cep
tw
her
en
ote
dth
atce
ntr
ep
oin
tw
asre
-est
imate
dw
ith
grea
ter
accu
racy
3 C
alc
ula
ted
usi
ng
the
mea
nofth
em
inim
um
an
dm
axim
um
esti
mate
sof
DF
orm
ula
tion
2co
nsi
der
sD
inth
ed
irec
tion
per
pen
dic
ula
rto
the
Pri
nci
pal
Lin
eo
nly
4 C
alc
ula
ted
inho
rizo
nta
lan
dve
rtic
aldir
ecti
on
sb
ut
still
ass
um
ing
geom
etry
of
gure
6(B
)F
irst
nu
mb
eris
the
mea
no
fth
em
inim
um
Dat
the
bott
om
of
the
ph
oto
and
the
maxi
mum
Dat
the
top
of
the
ph
oto
(range
show
nin
the
tab
le)
and
isco
mpara
ble
toF
orm
ula
tio
n2
Sec
ond
num
ber
isth
em
ean
of
Des
tim
ated
alon
gth
eP
rin
cip
alline
and
alo
ng
the
ou
tsid
eed
ge
(range
not
show
n)
5 Alt
itu
de
isap
pro
xim
ate
for
mis
sion
as
exac
tti
me
pho
togr
ap
hw
asta
ken
and
corr
esp
on
din
gn
adir
data
are
no
tava
ilab
le
Lack
of
nad
irdata
pre
ven
tsap
plica
tion
of
Form
ula
tion
s2
an
d3
6 Au
xilliar
ypo
int
add
edw
as
shy14
80
lati
tud
e13
51
2lo
ngit
ude
shy46
5degro
tati
on
angle
in
stru
ctio
ns
for
esti
mati
ng
the
rota
tio
nan
gle
para
met
erare
conta
ined
as
par
tof
the
scal
eca
lcu
lato
rsp
read
shee
tdo
cum
enta
tio
n
J A Robinson et al4432
a park at Johnson Space Center) that could clearly be distinguished on the photo-graph was 853 m wide
For the photograph of Houston taken with a 250 mm lens ( gure 3(C)) anddigitized from second generation lm at 2400 ppi (106 mm pixel Otilde 1 ) Ellington Fieldrunway 4-22 is 3 pixels in width and 1613 pixels in length so pixels represent151ndash152 m on the ground These results compare favourably with the minimumestimate of 152 m using Formulation 2 and 154ndash185 m pixels using Formulation 3(table 5) For an estimate of GRD using an 8times8 inch print (1369 enlargement) and4times magni cation the smallest street that could clearly be distinguished was 822 mwide the same feature could barely be distinguished on the digitized image
We also made an empirical estimate of spatial resolution for lower contrastvegetation boundaries By clearing forest so that a pattern would be visible to landingaircraft a landowner outside Austin Texas (see also aerial photo in Lisheron 2000)created a target that is also useful for evaluating spatial resolution of astronautphotographs The forest was selectively cleared in order to spell the landownerrsquosname lsquoLUECKErsquo with the remaining trees ( gure 10) According to local surveyorswho planned the clearing the plan was to create letters that were 3100 fttimes1700 ft(9449 mtimes5182 m) Photographed at a high altitude relative to most Shuttle missions(543 km) with a 250 mm lens Formula 3 predicts that each pixel would represent anarea 286 mtimes360 m on the ground (table 5) When original lm was digitized at2400 ppi (106 mm pixel Otilde 1 ) letters correspond to 294times188 pixels for a comparablepixel size of 27ndash32 m
6 Summary and conclusionsAstronaut photographs can be an excellent source of data for remote sensing
applications Best-case resolutions are similar to that for Landsat or SPOT withpixels as small as lt10 m It was estimated that of images taken to date 504 hadlens and obliquity characteristics that make them potential candidates for remotesensing information Digitized astronaut photographs can be overlaid with othersatellite data using GIS (Eckardt et al 2000) or used to ll in gaps in time serieswhen other imagery is not available As a source of public-domain information theycan be very useful for scientists who do not have access to satellite imagery eitherbecause of the expense of image acquisition (to the end user) or the computersystems needed for processing satellite images These diVerences in image acquisitioncosts (to the user) are summarized in table 6
Searching of the complete database of NASA astronaut photography includinglow-spatial-resolution browse images is available via the Web (OYce of EarthSciences 2000) This provides nearly global access for identifying images that willcontribute to a speci c scienti c project For the most detailed studies digitalproducts posted to the web will not be of suYcient quality To date we have madeit a practice of digitizing small numbers of images when requested by scientists atno charge Such requests (including a description of the project involved) can bemade through the authors or using the contact information listed on the websiteInvestigators with needs for larger numbers of digital images have also been servedthrough collaborative agreements
In our experience scientists that nd the data most valuable are those in develop-ing countries those studying areas that have not usually been targets for the majorsatellites those having diYculty in nding low-cloud images those interested inconstructing time series or those interested in using a large number of images
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
Astronaut photography as digital data for remote sensing 4409
(a) (b)
Figure 2 Example of the eVect of altitude on area covered in a photograph Both photographsof Lake Eyre Australia were taken using a Hasselblad camera with 100 mm lens(a) altitude 283 km STS093-702-062 (b) altitude 617 km STS082-754-61
photographs is due to the diVerent magni cations of the lenses Although taken froma diVerent altitude and using a diVerent camera with a diVerent original image size(see table 2) an electronic still camera image taken with a 300 mm lens is alsoincluded for comparison
For remote sensing longer focal length lenses are generally preferred (250 mm or350 mm for Hasselblad 300 mm or 400 mm lenses for 35 mm format cameras andESCs) Unfortunately longer-focal-length lenses exhibit poorer performance towardthe edge of a frame For example a 250 mm lens (Distagon CF 56) used on theHasselblad camera has spatial resolution of 57 lp mm Otilde 1 at the centre but only51 lp mm Otilde 1 at the edge and 46 lp mm Otilde 1 at the corner (tested with Ektachrome 5017f8 aperture and high contrast) A longer 300 mm lens used on the Nikon camerahas a greater spatial resolution diVerence between centre (82 lp mm Otilde 1 ) and corner(49 lp mm Otilde 1 ) using Kodak 5017 Ektachrome f4 aperture and high contrast FredPearce (unpublished data) There are tradeoVs among lens optics and speed Forexample the lenses for the Linhof system and the 250 mm lens for the Hasselblad(Distagon CF 56 see footnotes to table 1) are limited to apertures smaller thanf56 This becomes an important constraint in selecting shutter speeds (see discussionof shutter speed in sect42)
33 Cameras and actual image sizesAfter passing through the lens the photographic image is projected onto lm
inside the camera The size of this original image is another important propertydetermining spatial resolution and is determined by the camera used Cameraformats include 35 mm and 70 mm (Lowman 1980 Amsbury 1989) and occasionally5times4 inch (127times100 mm) and larger (table 1) Cameras own on each mission arenot metricmdashthey lack vacuum platens or reseau grids image-motion compensationor gyro-stabilized mounts The workhorse for engineering and Earth photographyon NASA missions has been a series of 70 mm Hasselblad cameras (table 1) chosenfor their reliability The modi ed magazine databack imprints a unique number and
J A Robinson et al4410
Tab
le1
Det
ails
on
the
cam
eras
use
dfo
rast
ron
aut
ph
oto
gra
phy
of
Eart
h
Data
incl
ud
ep
ho
togr
aphy
from
all
NA
SA
hu
man
spac
eig
ht
mis
sion
sth
rough
ST
S-9
6(J
un
e19
99
378
461
tota
lfr
am
esin
the
dat
abas
e)1
Imag
esi
zeT
ypic
al
lense
sN
um
ber
of
Cam
era
make
(mm
timesm
m)
use
d(m
m)
ph
oto
gra
ph
sP
erio
dof
use
No
tes
Has
selb
lad
55times
55
40
50100
2502
288
494
1963ndashp
rese
nt
Lin
ho
f10
5times120
90
250
29
373
1984ndashp
rese
nt
No
won
lyo
wn
occ
asio
nal
lyM
ult
ispec
tral
Cam
era
(S19
0A
)57times
57152
32
913
1973ndash19
74S
kyla
b
xed
-mou
nt3
cam
era
(NA
SA
1974
)E
art
hT
erra
inC
amer
a(S
190B
)11
5times11
5457
5478
1973ndash19
74S
kyla
b
xed
-mou
nt3
cam
era
(NA
SA
1974
)R
ole
iex
55times
55
50
100250
4352
1990ndash19
92L
arg
eF
orm
at
Cam
era4
229
times457
305
2143
1984
Mis
sio
nST
Sndash41G
only
(Mer
kel
etal
198
5)
Maure
r55
times55
76
80818
1961ndash19
68
1 Sm
all-f
orm
atca
mer
as
(35
mm
lm
and
ES
C)
are
no
tin
clud
edin
this
table
bec
ause
of
the
larg
enu
mb
erof
len
ses
and
con
gu
rati
on
sth
at
have
bee
nu
sed
and
bec
au
seth
ese
data
are
no
tcu
rren
tly
incl
uded
inth
edata
bas
efo
rall
mis
sion
s2 L
ense
suse
dfo
rH
ass
elbla
dm
anufa
cture
dby
Carl
Zei
ss
Dis
tago
nC
F4
(40
mm
)D
ista
gon
CF
4(5
0m
m)
Dis
tagon
CF
35
(100
mm
)D
ista
go
nC
F5
6(2
50
mm
)S
tand
ard
Has
selb
lad
len
ses
ow
nw
ill
chan
ge
inth
eyea
r20
00
toD
ista
gon
FE
28
(50
mm
)D
ista
go
nF
E2
(110
mm
)T
ele-
Tes
sar
FE
4(2
50m
m)
and
Tel
e-T
essa
rF
E4
(350
mm
)3 T
hes
eca
mer
as
wer
ein
xed
mou
nti
ngs
on
the
outs
ide
of
the
Skyla
bsp
ace
stati
on
A
ltho
ugh
not
hand
hel
d
the
lm
isca
talo
gued
wit
ho
ther
ast
ron
aut
ph
oto
gra
ph
yan
dis
incl
uded
her
efo
rco
mp
lete
nes
s4 A
NA
SA
-des
igned
cam
era
wit
hatt
itud
ere
fere
nce
syst
emfo
rd
eter
min
ing
pre
cise
nadir
poin
ts
Dat
ais
no
tcu
rren
tly
inco
rpora
ted
into
on
lin
ed
atab
ase
bu
tis
cata
loged
by
Mer
kel
etal
(1
985
)
Astronaut photography as digital data for remote sensing 4411
(a) (b)
(c) (d)
Figure 3 Example of the eVect of lens focal length on resolving power Hasselblad imagesof Houston were all taken from an altitude of 294 kmndash302 km with diVerent lenses(a) 40 mm STS062-97-151 (b) 100 mm STS094-737-28 (c) 250 mm STS055-71-41 Forcomparison an Electronic Still Camera image with 300 mm lens at 269 km altitude isalso shown (d ) S73E5145
timestamp on each frame at the time of exposure Cameras are serviced between ights Occasionally there is enough volume and mass allowance so that a Linhof5times4 inch (127times101 mm) format camera can be own Nikon 35 mm cameras are own routinely also because of proven reliability Electronic still cameras (ESC)were tested for Earth photography beginning in 1992 (Lulla and Holland 1993) Inan ESC a CCD (charge-coupled device) is used as a digital replacement for lmrecording the image projected inside the camera An ESC (consisting of a KodakDCS 460c CCD in a Nikon N-90S body) was added as routine equipment forhandheld photographs in 1995 ESCs have also been operated remotely to captureand downlink Earth images through a NASA-sponsored educational programme(EarthKAM) Discussion of CCD spatial array and radiometric sensitivity arebeyond the scope of this paper but are summarized by Robinson et al (2000b) The
J A Robinson et al4412
format of the lm (or CCD) and image size projected onto the lm (or CCD) aresummarized for all the diVerent cameras own in table 2
34 L ook angle or obliquityNo handheld photographs can be considered perfectly nadir they are taken at a
variety of look angles ranging from near vertical ( looking down at approximatelythe nadir position of the spacecraft ) to high oblique (images that include the curvatureof Earth) Imaging at oblique look angles leads to an image where scale degradesaway from nadir A set of views of the same area from diVerent look angles is shownin gure 4 The rst two shots of the island of Hawaii were taken only a few secondsapart and with the same lens The third photograph was taken on a subsequentorbit and with a shorter lens The curvature of the Earth can be seen in the upperleft corner
Obliquity can be described qualitatively ( gure 4) or quantitatively as the lookangle (t gure 1 calculations described in formulation 2 (sect52)) Because obliquityand look angle have such a dramatic in uence on the footprint we summarize thedatabase characteristics relative to these two parameters Figure 5 is a breakdownof the spatial resolution characteristics of low oblique and near vertical photographsin the NASA Astronaut Photography database Number of photographs are grouped(1) by calculated values for look angle (t ) and (2) by altitude After observing theoverlap between near vertical and low oblique classes we are currently restructuringthis variable (lsquotiltrsquo) in the database to provide a measure of t when available Userswill still be able to do searches based on the qualitative measures but these measureswill be more closely tied to actual look angle
341 Obliquity and georeferencing digitised photographsOften the rst step in a remote sensing analysis of a digitized astronaut photo-
graph is to georeference the data and resample it to conform to a known mapprojection Details and recommendations for resampling astronaut photographydata are provided by Robinson et al (2000a 2000c) and a tutorial is also available(McRay et al 2000) Slightly oblique photographs can be geometrically correctedfor remote sensing purposes but extremely oblique photographs are not suited forgeometric correction When obliquity is too great the spatial scale far away fromnadir is much larger than the spatial scale closer to nadir resampling results areunsuitable because pixels near nadir are lost as the image is resampled while manypixels far away from nadir are excessively replicated by resampling
To avoid the generation of excess pixels during georeferencing the pixel sizes ofthe original digitized image should be smaller than the pixels in the nal resampledimage Calculations of original pixel size using methods presented below can beuseful in ensuring meaningful resampling For slightly oblique images formulation 3(sect53) can be used to estimate pixel sizes at various locations in a photograph (nearnadir and away from nadir) and these pixel sizes then used to determine a reasonablepixel scale following resampling
4 Other characteristics that in uence spatial resolutionBeyond basic geometry many other factors internal to the astronaut photography
system impact the observed ground resolved distance in a photograph The lensesand cameras already discussed are imperfect and introduce radiometric degradations(Moik 1980)Vignetting (slight darkening around the edges of an image) occurs in
Astronaut photography as digital data for remote sensing 4413
Tab
le2
Max
imu
mpo
ssib
led
igit
alsp
atia
lre
solu
tio
n(p
ixel
size
inm
)fo
rca
mer
asy
stem
scu
rren
tly
use
dby
ast
ron
auts
top
ho
togr
aph
eart
hb
ase
do
nth
ege
om
etry
of
alt
itud
ele
ns
and
lm
size
C
alc
ula
tion
sas
sum
eth
atco
lour
tran
spar
ency
lm
isdig
itiz
edat
2400
ppi
(pix
els
per
inch
10
6m
mpix
elOtilde
1 )
Min
imu
mgro
un
ddis
tance
repre
sente
dby
1p
ixel
(m)
As
afu
nct
ion
of
alti
tude1
Imag
esi
zeM
inim
um
alt
itud
eM
edia
nalt
itu
de
Maxi
mum
alt
itud
eC
amer
asy
stem
mm
timesm
mp
ixel
s2(2
22k
m)
(326
km
)(6
11
km
)
Lin
ho
f125
mm
90
mm
lens
105
times120
8978times
11
434
261
383
718
250
mm
lens
94
138
259
Has
selb
lad
70
mm
100
mm
lens
55times
55
5198
times519
823
534
564
725
0m
mle
ns
94
138
259
Nik
on
35m
m30
0m
mle
ns
36times
24
3308
times217
478
115
216
400
mm
lens
59
86
162
ESC
35m
m18
0m
mle
ns3
27
6times
18
530
69times
204
311
116
330
630
0m
mle
ns
67
98
184
400
mm
lens
50
74
138
1 Alt
itud
essh
ow
nar
em
inim
um
m
edia
nan
dm
axim
um
thro
ugh
ST
S-8
9(J
anuary
199
8N
=84
mis
sion
s)
2 Act
ual
pix
els
for
ES
C
oth
erw
ise
assu
med
dig
itiz
ati
on
at240
0pp
i(p
ixel
sp
erin
ch)
equ
alto
10
6m
mpix
elOtilde
1 3 T
he
mo
stco
mm
on
con
gura
tion
for
Eart
hph
oto
gra
ph
yas
part
of
the
Eart
hK
AM
pro
ject
J A Robinson et al4414
Figure 4 De nitions of look angle for photographs taken from low orbit around EarthThe example photographs of Hawaii were all taken from approximately the samealtitude (370ndash398 km) and with the same (250 mm) lens on a Hasselblad camera(STS069-701-69 STS069-701-70 and STS069-729-27 [100 mm lens])
astronaut photography due to light path properties of the lenses used A lack of atness of the lm at the moment of exposure (because cameras used in orbit donot incorporate a vacuum platen) also introduces slight degradations A numberof additional sources of image degradation that would aVect GRD of astronautphotographs are listed
41 Films and processing411 Colour reversal lms
Most lms used in NASA handheld cameras in orbit have been E-6 processcolour reversal lms (Ektachromes) that have a nominal speed of 64 to 100 ASAalthough many other lms have been own (table 3) Reversal lms are used becausedamage from radiation exposure while outside the atmospheric protection of Earthis more noticeable in negative lms than in positive lms (Slater 1996) The choiceof ASA has been to balance the coarser grains ( lower lm resolving power) of high-speed lms with the fact that slower lms require longer exposures and thus areaVected more by vehicle motion (approximately 73 km s Otilde 1 relative to Earthrsquos surfacefor the Space Shuttle) Extremely fast lms (gt400 ASA) are also more susceptibleto radiation damage in orbit (particularly during long-duration or high-altitudemissions Slater 1996) and have not traditionally been used for Earth photography The manufacturer stock numbers identifying the lm used are available for eachimage in the Astronaut Photography Database (OYce of Earth Sciences 2000)
412 Colour infrared lmsColour infrared lm (CIR) was used during an Earth-orbiting Apollo mission
(Colwell 1971) in the multispectral S-190A and the high-resolution S-190B camera
Astronaut photography as digital data for remote sensing 4415
Figure 5 (a) Distributions of astronaut photographs by look angle oV-nadir (calculated fromcentre point and nadir point using Formulation 2) Data included for all photographsfor which centre and nadir points are known with high oblique look angles (n=55 155) excluded (Total sample shown is 108 293 of 382 563 total records in thedatabase when compiled For records where nadir data was not available number ofobservations for qualitative look angles are near vertical 14 086 low oblique 115 834undetermined 89 195) (b) Altitude distribution for astronaut photographs of Earthtaken with the Hasselblad camera Data included for Hasselblad camera only focallength determined with high oblique look angles excluded (Total sample shown is159 046 of 206 971 Hasselblad records with known altitude and of 382 563 total recordsin the database when compiled For Hasselblad records with known altitude data notshown includes high oblique photographs using 100 mm and 250 mm lenses n=20 262and all look angles with other lenses n=27 663)
systems on Skylab (NASA 1974 Wilmarth et al 1977) and occasionally on Shuttlemissions and Shuttle-Mir missions (table 3) The CIR lm used is a three-layerAerochrome having one layer that is sensitive to re ected solar infrared energy toapproximately 900 nm (Eastman Kodak Company 1998) protective coatings onmost spacecraft windows also limit IR transmittance between 800 mm and 1000 nmThis layer is extremely sensitive to the temperature of the lm which createsunpredictable degradation of the IR signature under some space ight conditions
J A Robinson et al4416
Tab
le3
Det
ails
on
lm
suse
dfo
rast
ron
aut
pho
togr
aphy
of
Eart
h
Dat
ain
clud
ep
ho
togr
aphy
fro
mall
NA
SA
hum
an
spac
eig
ht
mis
sio
ns
thro
ugh
ST
S-9
6(J
un
e19
99
378
461
tota
lfr
ames
inth
edata
base
)F
ilm
sw
ith
lt50
00fr
am
esex
po
sed
and
lm
suse
dfo
r35
mm
cam
eras
only
are
no
tin
clud
ed
Res
olv
ing
Nu
mb
erof
Film
man
ufa
ctu
rer
AS
Ap
ow
er1
fram
esP
erio
dof
use
Cam
eras
Colo
ur
posi
tive
Kod
akA
eroch
rom
eII
(2447
2448
SO
-131S
O-3
68
)32
40ndash50
513
8196
8ndash19
85
Hass
elbla
dL
inh
of
Kod
akE
kta
chro
me
(5017
502
56
017
6118
SO
-121
64
50
113
932
196
5ndash19
68
Hass
elbla
dL
inh
of
Role
iex
SO
-217
Q
X-8
24
QX
-868
)19
81ndash1993
Mau
rer
Nik
on
Kod
akL
um
iere
(5046
5048
)100
105
743
199
4ndash19
98
Hass
elbla
dL
inho
fK
od
akE
lite
100
S(5
069
)100
41
486
199
6ndashp
rese
nt
Hass
elbla
d2
Fu
jiV
elvia
50C
S135
-36
32
13
882
1992ndash19
97H
ass
elbla
dN
iko
nK
od
akH
i-R
esolu
tio
nC
olo
r(S
O-2
42S
O-3
56
)10
096
74
1973ndash19
75H
ass
elbla
d
S19
0A
S190
B
Infr
are
dand
bla
ck-a
nd-w
hit
e
lms
Kod
akA
eroch
rom
eC
olo
rIn
frar
ed40
32
40
704
196
5ndashp
rese
nt2
Hass
elbla
dL
inh
of
Nik
on
S190
A
(8443
34
432443
)S
190B
Kod
akB
ampW
Infr
are
d(2
424
)32
10
553
197
3ndash19
74
S190
AK
od
akP
anato
mic
-XB
ampW
(SO
-022
)63
10
657
197
3ndash19
74
S190
A
1A
tlo
wco
ntr
ast
(ob
ject
back
grou
nd
rati
o16
1)
aspro
vid
edby
Ko
dak
and
list
edby
Sla
ter
etal
(1
983
256
)R
esolv
ing
pow
erw
as
dis
con
tin
ued
fro
mth
ete
chn
ical
test
sper
form
edb
yK
odak
inapp
roxim
atel
y19
93
an
dit
isn
ot
kn
ow
nif
curr
ent
lm
sh
ave
sim
ilar
reso
lvin
gpo
wer
too
lder
ver
sion
so
fsi
milar
lm
s(K
T
eite
lbaum
K
od
akp
erso
nal
com
mu
nic
atio
n)
Th
egra
nu
lari
tym
easu
reno
wpro
vid
edb
yK
od
ak
can
no
tb
ere
adily
con
ver
ted
toa
spat
ial
reso
luti
on
2 The
Lin
hof
was
use
dag
ain
on
ST
S-1
03in
Dec
emb
er19
99(d
ata
not
cata
logued
asof
the
pre
para
tion
of
this
tab
le)
usi
ng
Ko
dak
Elite
100S
lm
3 B
egin
nin
gin
July
1999
2443
colo
ur-
infr
are
d
lmpro
cess
was
chan
ged
from
EA
-5to
E-6
Astronaut photography as digital data for remote sensing 4417
413 Film duplicationThe original lm is archived permanently after producing about 20 duplicate
printing masters (second generation) Duplicate printing masters are disseminatedto regional archives and used to produce products (thirdndashfourth generation) for themedia the public and for scienti c use When prints slides transparencies anddigital products are produced for public distribution they are often colour adjustedto correspond to a more realistic look Digital products from second generationcopies can be requested by scienti c users (contact the authors for information onobtaining such products for a particular project) and these are recommended forremote sensing applications Care should be taken that the digital product acquiredfor remote sensing analysis has not been colour adjusted for presentation purposesBased on qualitative observations there is little visible increase in GRD in the thirdor fourth generation products There is a signi cant degradation in delity of colourin third and fourth generation duplicates because an increase in contrast occurswith every copy
42 Shutter speedsThe impact of camera motion (both due to the photographer and to the motion
of the spacecraft ) on image quality is determined by shutter speedmdash1250 to 1500second have been used because slower speeds record obvious blurring caused bythe rapid motion of the spacecraft relative to the surface of the Earth Generally1250 was used for ISO 64 lms (and slower) and after the switch to ISO 100 lmsa 1500 setting became standard New Hasselblad cameras that have been ownbeginning with STS-92 in October 2000 vary shutter speed using a focal plane shutterfor exposure bracketing (rather than varying aperture) and 1250 1500 and 11000second are used
Across an altitude range of 120ndash300 nautical miles (222ndash555 km) and orbitalinclinations of 285deg and 570deg the median relative ground velocity of orbitingspacecraft is 73 km s Otilde 1 At a nominal shutter speed of 1500 the expected blur dueto motion relative to the ground would be 146 m When cameras are handheld (asopposed to mounted in a bracket) blur can be reduced by physical compensationfor the motion (tracking) by the photographer many photographs show a level ofdetail indicating that blur at this level did not occur (eg gure 3(d )) Thus motionrelative to the ground is not an absolute barrier to ground resolution for handheldphotographs
43 Spacecraft windowsMost of the photographs in the NASA Astronaut Photography Database were
taken through window ports on the spacecraft used The transmittance of the windowport may be aVected by material inhomogeniety of the glass the number of layersof panes coatings used the quality of the surface polish environmentally inducedchanges in window materials (pressure loads thermal gradients) or deposited con-tamination Such degradation of the window cannot be corrected is diVerent foreach window and changes over time See Eppler et al (1996) and Scott (2000) fordiscussion of the spectral transmittance of the fused quartz window that is part ofthe US Laboratory Module (Destiny) of the International Space Station
44 Digitized images from lmWhen lm is scanned digitally the amount of information retained depends on
the spectral information extracted from the lm at the spatial limits of the scanner
J A Robinson et al4418
To date standard digitizing methodologies for astronaut photographs have not beenestablished and lm is digitized on a case-by-case basis using the equipment available
441 Digitizing and spatial resolutionLight (1993 1996) provided equations for determining the digitizing spatial
resolution needed to preserve spatial resolution of aerial photography based on thestatic resolving power (AWAR) for the system For lms where the manufacturer hasprovided data (Eastman Kodak 1998 K Teitelbaum personal communication) the resolving power of lms used for astronaut photography ranges from 32 lp mm Otilde 1to 100 lp mm Otilde 1 at low contrast (objectbackground ratio 161 table 3) The AWARfor the static case of the Hasselblad camera has been measured at high and lowcontrast (using lenses shown in table 1) with a maximum of approximately55 lp mm Otilde 1 (Fred Pearce unpublished data)
Based on the method of Light (1993 1996) the dimension of one spatial resolutionelement for a photograph with maximum static AWAR of 55 lp mm Otilde 1 would be18 mm lp Otilde 1 The acceptable range of spot size to preserve spatial information wouldthen be
18 mm
2 atilde 2aring scan spot size aring
18 mm
2(1)
and 6 mm aring scan spot size aring 9 mm Similarly for a more typical static AWAR of30 lp mm Otilde 1 (33 mm lpOtilde 1 low contrast Fred Pearce unpublished data) 11 mm aring scanspot sizearing 17 mm These scan spot sizes correspond to digitizing resolutions rangingfrom 4233ndash2822 ppi (pixels inch Otilde 1 ) for AWAR of 55 lp mm Otilde 1 and 2309ndash1494 ppi forAWAR of 30 lp mm Otilde 1 Scan spot sizes calculated for astronaut photography arecomparable to those calculated for the National Aerial Photography Program(9ndash13 mm Light 1996)
Widely available scanners that can digitize colour transparency lm currentlyhave a maximum spatial resolution of approximately 2400 ppi (106 mm pixel Otilde 1) Forexample we routinely use an Agfa Arcus II desktop scanner (see also Baltsavias1996) with 2400 ppi digitizing spatial resolution (2400 ppi optical resolution in onedirection and 1200 ppi interpolated to 2400 ppi in the other direction) and 2400 ppiwas used to calculate IFOV equivalents (tables 2 and 4) For some combinations oflens camera and contrast 2400 ppi will capture nearly all of the spatial informationcontained in the lm However for lm with higher resolving power thanEktachrome-64 for better lenses and for higher contrast targets digitizing at 2400 ppiwill not capture all of the spatial information in the lm
Improvements in digitizing technology will only produce real increases in IFOVto the limit of the AWAR of the photography system The incremental increase inspatial information above 3000 ppi (85 mm pixel Otilde 1 ) is not likely to outweigh thecosts of storing large images (Luman et al 1997) At 2400 ppi a Hasselblad frameis approximately 5200times5200 or 27 million pixels (table 2) while the same imagedigitized at 4000 ppi would contain 75 million pixels
Initial studies using astronaut photographs digitized at 2400 ppi (106 mm pixel Otilde 1 Webb et al 2000 Robinson et al 2000c) indicate that some GRD is lost comparedto photographic products Nevertheless the spatial resolution is still comparablewith other widely used data sources (Webb et al 2000) Robinson et al (2000c)found that digitizing at 21 mm pixel Otilde 1 provided information equivalent to 106 mmpixel Otilde 1 for identifying general urban area boundaries for six cities except for a single
Astronaut photography as digital data for remote sensing 4419
photo that required higher spatial resolution digitizing (it had been taken with ashorter lens and thus had less spatial resolution) Part of the equivalence observedin the urban areas study may be attributable to the fact that the atbed scannerused interpolates from 1200 ppi to 2400 ppi in one direction The appropriate digitiz-ing spatial resolution will in part depend on the scale of features of interest to theresearchers with a maximum upper limit set of a scan spot size of approximately6 mm (4233 ppi)
442 Digitizing and spectral resolutionWhen colour lm is digitized there will be loss of spectral resolution The three
lm emulsion layers (red green blue) have relatively distinct spectral responses butare fused together so that digitizers detect one colour signal and must convert itback into three (red green blue) digital channels Digital scanning is also subject tospectral calibration and reproduction errors Studies using digital astronaut photo-graphs to date have used 8 bits per channel However this is another parameterthat can be controlled when the lm is digitized We do not further address spectralresolution of digitally scanned images in this paper
45 External factors that in uence GRDAlthough not discussed in detail in this paper factors external to the spacecraft
and camera system (as listed by Moik (1980) for remote sensing in general) alsoimpact GRD These include atmospheric interference (due to scattering attenuationhaze) variable surface illumination (diVerences in terrain slope and orientation) andchange of terrain re ectance with viewing angle (bidirectional re ectance) For astro-naut photographs variable illumination is particularly important because orbits arenot sun-synchronous Photographs are illuminated by diVerent sun angles and imagesof a given location will have colour intensities that vary widely In addition theviewing angle has an eVect on the degree of object-to-backgroun d contrast andatmospheric interference
5 Estimating spatial resolution of astronaut photographsIn order to use astronaut photographs for digital remote sensing it is important
to be able to calculate the equivalent to an IFOVmdashthe ground area represented bya single pixel in a digitized orbital photograph The obliquity of most photographsmeans that pixel lsquosizesrsquo vary at diVerent places in an image Given a ground distanceD represented by a photograph in each direction (horizontal vertical ) an approxi-mate average pixel width (P the equivalent of IFOV) for the entire image can becalculated as follows
P=D
dS(2)
where D is the projected distance on the ground covered by the image in the samedirection as the pixel is measured d is the actual width of the image on the original lm (table 1) and S is the digitizing spatial resolution
Here we present three mathematical formulations for estimating the size of thefootprint or area on the ground covered by the image Example results from theapplication of all three formulations are given in table 5 The rst and simplestcalculation (formulation 1) gives an idea of the maximum spatial resolution attainableat a given altitude of orbit with a given lm format and a perfectly vertical (nadir)
J A Robinson et al4420
view downward Formulation 2 takes into account obliquity by calculating lookangle from the diVerence between the location on the ground represented at thecentre of the photograph and the nadir location of the spacecraft at the time thephotograph was taken ( gure 1) Formulation 3 describes an alternate solution tothe oblique look angle problem using coordinate-system transformations Thisformulation has been implemented in a documented spreadsheet and is availablefor download (httpeoljscnasagovsseopFootprintCalculatorxls) and is in theprocess of being implemented on a Web-based user interface to the AstronautPhotography Database (OYce of Earth Sciences 2000)
Although formulations 2 and 3 account for obliquity for the purposes of calcula-tion they treat the position of the spacecraft and position of the camera as one Inactuality astronauts are generally holding the cameras by hand (although camerasbracketed in the window are also used) and the selection of window position of theastronaut in the window and rotation of the camera relative to the movement ofthe spacecraft are not known Thus calculations using only the photo centre point(PC) and spacecraft nadir point (SN) give a locator ellipse and not the locations ofthe corners of the photograph A locator ellipse describes an estimated area on theground that is likely to be included in a speci c photograph regardless of the rotationof the lm plane about the camerarsquos optical axis ( gure 6)
Estimating the corner positions of the photo requires additional user input of asingle auxiliary pointmdasha location on the image that has a known location on theground Addition of this auxiliary point is an option available to users of thespreadsheet An example of the results of adding an auxiliary point is shown in gure 7 with comparisons of the various calculations in table 5
51 Formulation 1 Footprint for a nadir viewThe simplest way to estimate footprint size is to use the geometry of camera lens
and spacecraft altitude to calculate the scaling relationship between the image in the
Figure 6 Sketch representation of the locator ellipse an estimated area on the ground thatis likely to be included in a speci c photograph regardless of the rotation of the lmplane about the camerarsquos optical axis
Astronaut photography as digital data for remote sensing 4421
Figure 7 An example of the application of the Low Oblique Space Photo FootprintCalculator The photograph (STS093-704-50) of Limmen Bight Australia was takenfrom an altitude of 283 km using the Hasselblad camera with a 250 mm lens Mapused is 11 000 000 Operational Navigational Chart The distances shown equate toan average pixel size of 124 m for this photograph
lm and the area covered on the ground For a perfect nadir view the scalerelationship from the geometry of similar triangles is
d
D=
f
H(3)
where d=original image size D=distance of footprint on the ground f =focallength of lens and H=altitude Once D is known pixel size ( length=width) can becalculated from equation (2) These calculations represent the minimum footprintand minimum pixel size possible for a given camera system altitude and digitisingspatial resolution (table 2) Formulation 1 was used for calculating minimum pixelsizes shown in tables 2 and 4
By assuming digitizing at 2400 ppi (106 mm pixel Otilde 1 ) currently a spatial resolutioncommonly attainable from multipurpose colour transparency scanners (see sect441)this formulation was used to convert area covered to an IFOV equivalent for missionsof diVerent altitudes (table 2) Table 4 provides a comparison of IFOV of imagesfrom various satellites including the equivalent for astronaut photography Valuesin this table were derived by using formulation 1 to estimate the area covered becausea perfect nadir view represents the best possible spatial resolution and smallest eldof view that could be obtained
52 Formulation 2 Footprint for oblique views using simpli ed geometry and thegreat circle distance
A more realistic approach to determining the footprint of the photographaccounts for the fact that the camera is not usually pointing perfectly down at the
J A Robinson et al4422
Table 4 Comparison of pixel sizes (instantaneous elds of view) for satellite remote sensorswith pixel sizes for near vertical viewing photography from spacecraft Note that alldata returned from the automated satellites have the same eld of view while indicatedpixel sizes for astronaut photography are best-case for near vertical look angles See gure 5 for distributions of astronaut photographs of various look angles High-altitude aerial photography is also included for reference
Altitude PixelInstrument (km) width (mtimesm) Bands1
Landsat Multipectral Scanner2 (MSS 1974-) 880ndash940 79times56 4ndash5Landsat Thematic Mapper (TM 1982-) ~705 30times30 7Satellite pour lrsquoObservation de la Terre (SPOT 1986-) ~832
SPOT 1-3 Panchromatic 10times10 1SPOT 1-3 Multispectral (XS) 20times20 3
Astronaut Photography (1961-)Skylab3 (1973-1974 ) ~435
Multispectral Camera (6 camera stations) 17ndash19times17ndash19 3ndash8Earth Terrain Camera (hi-res colour) 875times875 3
Space Shuttle5 (1981-) 222ndash611Hasselblad 70 mm (250mm lens colour) vertical view 9ndash26times9ndash26 3Hasselblad 70 mm (100mm lens colour) vertical view 23ndash65times23ndash65 3Nikon 35 mm (400mm lens colour lm or ESC) vertical 5ndash16times5ndash16 3
High Altitude Aerial Photograph (1100 000) ~12 2times2 3
1Three bands listed for aerial photography obtainable by high-resolution colour digitizingFor high-resolution colour lm at 65 lp mmOtilde 1 (K Teitelbaum Eastman Kodak Co personalcommunication) the maximum information contained would be 3302 ppi
2Data from Simonett (1983Table 1-2)3Calculated as recommended by Jensen (19831596) the dividend of estimated ground
resolution at low contrast given in NASA (1974179-180) and 244Six lters plus colour and colour infrared lm (NASA 19747) 5Extracted from table 2
nadir point The look angle (the angle oV nadir that the camera is pointing) can becalculated trigonometrically by assuming a spherical earth and calculating the dis-tance between the coordinates of SN and PC ( gure 1) using the Great Circledistance haversine solution (Sinnott 1984 Snyder 198730ndash32 Chamberlain 1996)The diVerence between the spacecraft centre and nadir point latitudes Dlat=lat 2 shy lat 1 and the diVerence between the spacecraft centre and nadir pointlongitudes Dlon=lon 2 shy lon 1 enter the following equations
a=sin2ADlat
2 B+cos(lat 1) cos(lat 2) sin2ADlon
2 B (4)
c=2 arcsin(min[1 atilde a]) (5)
and oVset=R c with R the radius of a spherical Earth=6370 kmThe look angle (t ) is then given by
t=arctanAoffset
H B (6)
Assuming that the camera was positioned so that the imaginary line between thecentre and nadir points (the principal line) runs vertically through the centre of thephotograph the distance between the geometric centre of the photograph (principalpoint PP) and the top of the photograph is d2 ( gure 8(a)) The scale at any point
Astronaut photography as digital data for remote sensing 4423
(a) (b)
Figure 8 The geometric relationship between look angle lens focal length and the centrepoint and nadir points of a photograph (see also Estes and Simonett 1975) Note thatthe Earthrsquos surface and footprint represented in gure 1 are not shown here only arepresentation of the image area of the photograph Variables shown in the gurelook angle t focal length f PP centre of the image on the lm SN spacecraft nadird original image size (a) The special case where the camera is aligned squarely withthe isometric parallel In this case tilt occurs along a plane parallel with the centre ofthe photograph When the conditions are met calculations using Formulation 3 willgive the ground coordinates of the corner points of the photograph (b) The generalrelationship between these parameters and key photogrammetric points on the photo-graph For this more typical case coordinates of an additional point in the photographmust be input in order to adjust for twist of the camera and to nd the corner pointcoordinates
in the photograph varies as a function of the distance y along the principal linebetween the isocentre and the point according to the relationship
d
D=
f shy ysint
H(7)
(Wong 1980 equation (214) H amp the ground elevation h) Using gure 8(a) at thetop of the photo
y= f tanAt
2B+d
2(8)
and at the bottom of the photo
y= f tanAt
2Bshyd
2(9)
Thus for given PC and SN coordinates and assuming a photo orientation as in gure 8(a) and not gure 8(b) we can estimate a minimum D (using equations (7)and (8)) and a maximum D (using equations (7) and (9)) and then average the twoto determine the pixel size (P ) via equation (2)
J A Robinson et al4424
53 Formulation 3 T he L ow Oblique Space Photo Footprint CalculatorThe Low Oblique Space Photo Footprint Calculator was developed to provide
a more accurate estimation of the geographic coordinates for the footprint of a lowoblique photo of the Earthrsquos surface taken from a human-occupied spacecraft inorbit The calculator performs a series of three-dimensional coordinate transforma-tions to compute the location and orientation of the centre of the photo exposureplane relative to an earth referenced coordinate system The nominal camera focallength is then used to create a vector from the photorsquos perspective point througheach of eight points around the parameter of the image as de ned by the formatsize The geographical coordinates for the photo footprint are then computed byintersecting these photo vectors with a spherical earth model Although more sophist-icated projection algorithms are available no signi cant increase in the accuracy ofthe results would be produced by these algorithms due to inherent uncertainties inthe available input data (ie the spacecraft altitude photo centre location etc)
The calculations were initially implemented within a Microsoft Excel workbookwhich allowed one to embed the mathematical and graphical documentation nextto the actual calculations Thus interested users can inspect the mathematical pro-cessing A set of error traps was also built into the calculations to detect erroneousresults A summary of any errors generated is reported to the user with the resultsof the calculations Interested individuals are invited to download the Excel work-book from httpeoljscnasagovsseopFootprintCalculatorxls The calculations arecurrently being encoded in a high-level programming language and should soon beavailable alongside other background data provided for each photograph at OYceof Earth Sciences (2000)
For the purposes of these calculations a low oblique photograph was de ned as
one with the centre within 10 degrees of latitude and longitude of the spacecraftnadir point For the typical range of spacecraft altitudes to date this restricted thecalculations to photographs in which Earthrsquos horizon does not appear (the generalde nition of a low oblique photograph eg Campbell 199671 and gure 4)
531 Input data and resultsUpon opening the Excel workbook the user is presented with a program intro-
duction providing instructions for using the calculator The second worksheet tablsquoHow-To-Usersquo provides detailed step-by-step instructions for preparing the baselinedata for the program The third tab lsquoInput-Output rsquo contains the user input eldsand displays the results of the calculations The additional worksheets contain theactual calculations and program documentation Although users are welcome toreview these sheets an experienced user need only access the lsquoInput-Outputrsquo
spreadsheetTo begin a calculation the user enters the following information which is available
for each photo in the NASA Astronaut Photography Database (1) SN geographicalcoordinates of spacecraft nadir position at the time of photo (2) H spacecraftaltitude (3) PC the geographical coordinates of the centre of the photo (4) f nominal focal length and (5) d image format size The automatic implementationof the workbook on the web will automatically enter these values and completecalculations
For more accurate results the user may optionally enter the geographic co-ordinates and orientation for an auxiliary point on the photo which resolves thecamerarsquos rotation uncertainty about the optical axis The auxiliary point data must
Astronaut photography as digital data for remote sensing 4425
be computed by the user following the instruction contained in the lsquoHow-To-Usersquotab of the spreadsheet
After entering the input data the geographic coordinates of the photo footprint(ie four photo corner points four points at the bisector of each edge and the centreof the photo) are immediately displayed below the input elds along with any errormessages generated by the user input or by the calculations ( gure 7) Although
results are computed and displayed they should not be used when error messagesare produced by the program The program also computes the tilt angle for each ofthe photo vectors relative to the spacecraft nadir vector To the right of the photofootprint coordinates is displayed the arc distance along the surface of the sphere
between adjacent computed points
532 Calculation assumptions
The mathematical calculations implemented in the Low Oblique Space PhotoFootprint Calculator use the following assumptions
1 The SN location is used as exact even though the true value may vary by upto plusmn01 degree from the location provided with the photo
2 The spacecraft altitude is used as exact Although the determination of the
nadir point at the instant of a known spacecraft vector is relatively precise(plusmn115times10 Otilde 4 degrees) the propagator interpolates between sets of approxi-mately 10ndash40 known vectors per day and the time code recorded on the lmcan drift Thus the true value for SN may vary by up to plusmn01 degree fromthe value provided with the photo
3 The perspective centre of the camera is assumed to be at the given altitudeover the speci ed spacecraft nadir location at the time of photo exposure
4 The PC location is used as exact even though the true value may vary by upto plusmn05deg latitude and plusmn05deg longitude from the location provided with the
photo5 A spherical earth model is used with a nominal radius of 6 372 16154 m (a
common rst-order approximation for a spherical earth used in geodeticcomputations)
6 The nominal lens focal length of the camera lens is used in the computations(calibrated focal length values are not available)
7 The photo projection is based on the classic pin-hole camera model8 No correction for lens distortion or atmospheric re ection is made
9 If no auxiliary point data is provided the lsquoTop of the Imagersquo is orientedperpendicular to the vector from SN towards PC
533 T ransformation from Earth to photo coordinate systemsThe calculations begin by converting the geographic coordinates ( latitude and
longitude) of the SN and PC to a Rectangular Earth-Centred Coordinate System(R-Earth) de ned as shown in gure 9 (with the centre of the Earth at (0 0 0))
Using the vector from the Earthrsquos centre through SN and the spacecraft altitude thespacecraft location (SC) is also computed in R-Earth
For ease of computation a Rectangular Spacecraft-Centred Coordinate System(R-Spacecraft ) is de ned as shown in gure 9 The origin of R-Spacecraft is located
at SC with its +Z-axis aligned with vector from the centre of the Earth throughSN and its +X-axis aligned with the vector from SN to PC ( gure 9) The speci c
J A Robinson et al4426
Figure 9 Representation of the area included in a photograph as a geometric projectiononto the Earthrsquos surface Note that the focal length (the distance from the SpaceShuttle to the photo) is grossly overscale compared to the altitude (the distance fromthe Shuttle to the surface of the Earth
rotations and translation used to convert from the R-Earth to the R-Spacecraft arecomputed and documented in the spreadsheet
With the mathematical positions of the SC SN PC and the centre of the Earthcomputed in R-Spacecraft the program next computes the location of the camerarsquosprincipal point (PP) The principle point is the point of intersection of the opticalaxis of the lens with the image plane (ie the lm) It is nominally positioned at a
Astronaut photography as digital data for remote sensing 4427
distance equal to the focal length from the perspective centre of the camera (whichis assumed to be at SC) along the vector from PC through SC as shown in gure 9
A third coordinate system the Rectangular Photo Coordinate System (R-Photo)is created with its origin at PP its XndashY axial plane normal to the vector from PCthrough SC and its +X axis aligned with the +X axis of R-Spacecraft as shownin gure 9 The XndashY plane of this coordinate system represents the image plane ofthe photograph
534 Auxiliary point calculationsThe calculations above employ a critical assumption that all photos are taken
with the lsquotop of the imagersquo oriented perpendicular to the vector from the SN towardsPC as shown in gure 9 To avoid non-uniform solar heating of the external surfacemost orbiting spacecraft are slowly and continually rotated about one or more axesIn this condition a ight crew member taking a photo while oating in microgravitycould orient the photo with practically any orientation relative to the horizon (seealso gure 8(b)) Unfortunately since these photos are taken with conventionalhandheld cameras there is no other information available which can be used toresolve the photorsquos rotational ambiguity about the optical axis other then the photoitself This is why the above assumption is used and the footprint computed by thiscalculator is actually a lsquolocator ellipsersquo which estimates the area on the ground whatis likely to be included in a speci c photograph (see gure 6) This locator ellipse ismost accurate for square image formats and subject to additional distortion as thephotograph format becomes more rectangular
If the user wants a more precise calculation of the image footprint the photorsquosrotational ambiguity about the optical axis must be resolved This can be done inthe calculator by adding data for an auxiliary point Detailed instructions regardinghow to prepare and use auxiliary point data in the computations are included in thelsquoHow-To-Usersquo tab of the spreadsheet Basically the user determines which side ofthe photograph is top and then measures the angle between the line from PP to thetop of the photo and from PP to the auxiliary point on the photo ( gure 9)
If the user includes data for an auxiliary point a series of computations arecompleted to resolve the photo rotation ambiguity about the optical axis (ie the
+Z axis in R-Photo) A vector from the Auxiliary Point on the Earth (AE) throughthe photograph perspective centre ( located at SC) is intersected with the photo imageplane (XndashY plane of R-Photo) to compute the coordinates of the Auxiliary Point onthe photo (AP) in R-Photo A two-dimensional angle in the XndashY plane of R-Photofrom the shy X axis to a line from PP to AP is calculated as shown in gure 9 The
shy X axis is used as the origin of the angle since it represents the top of the photoonce it passes through the perspective centre The diVerence between the computedangle and the angle measured by the user on the photo resolves the ambiguity inthe rotation of the photo relative to the principal line ( gures 7 and 9) Thetransformations from R-Spacecraft and R-Photo are then modi ed to include anadditional rotation angle about the +Z axis in R-Photo
535 lsquoFootprint rsquo calculationsThe program next computes the coordinates of eight points about the perimeter
of the image format (ie located at the four photo corners plus a bisector pointalong each edge of the image) These points are identi ed in R-Photo based uponthe photograph format size and then converted to R-Spacecraft Since all computa-tions are done in orthogonal coordinate systems the R-Spacecraft to R-Photo
J A Robinson et al4428
rotation matrix is transposed to produce an R-Photo to R-Spacecraft rotation matrixOnce in R-Spacecraft a unit vector from each of the eight perimeter points throughthe photo perspective centre (the same point as SC) is computed This provides thecoordinates for points about the perimeter of the image format with their directionvectors in a common coordinate system with other key points needed to computethe photo footprint
The next step is to compute the point of intersection between the spherical earthmodel and each of the eight perimeter point vectors The scalar value for eachperimeter point unit vector is computed using two-dimensional planar trigonometryAn angle c is computed using the formula for the cosine of an angle between twoincluded vectors (the perimeter point unit vector and the vector from SC to thecentre of the Earth) Angle y is computed using the Law of Sines Angle e=180degrees shy c shy y The scalar value of the perimeter point vector is computed using eand the Law of Cosines The scalar value is then multiplied by the perimeter pointunit vector to produce the three-dimensional point of intersection of the vector withEarthrsquos surface in R-Spacecraft The process is repeated independently for each ofthe eight perimeter point vectors Aside from its mathematical simplicity the valueof arcsine (computed in step 2) will exceed the normal range for the sin of an anglewhen the perimeter point vectors fail to intersect with the surface of the earth Asimple test based on this principle allows the program to correctly handle obliquephotos which image a portion of the horizon (see results for high oblique photographsin table 5)
The nal step in the process converts the eight earth intersection points from theR-Spacecraft to R-Earth The results are then converted to the standard geographiccoordinate system and displayed on the lsquoInput-Outputrsquo page of the spreadsheet
54 ExamplesAll three formulations were applied to the photographs included in this paper
and the results are compared in table 5 Formulation 1 gives too small a value fordistance across the photograph (D) for all but the most nadir shots and thus servesas an indicator of the best theoretical case but is not a good measure for a speci cphotograph For example the photograph of Lake Eyre taken from 276 km altitude( gure 2(a)) and the photograph of Limmen Bight ( gure 7) were closest to beingnadir views (oVsetslt68 km or tlt15deg table 5) For these photographs D calculatedusing Formulation 1 was similar to the minimum D calculated using Formulations2 and 3 For almost all other more oblique photographs Formulation 1 gave asigni cant underestimate of the distance covered in the photograph For gure 5(A)(the picture of Houston taken with a 40 mm lens) Formulation 1 did not give anunderestimate for D This is because Formulation 1 does not account for curvatureof the Earth in any way With this large eld of view assuming a at Earth in atedthe value of D above the minimum from calculations that included Earth curvature
A major diVerence between Formulations 2 and 3 is the ability to estimate pixelsizes (P) in both directions (along the principal line and perpendicular to the principalline) For the more oblique photographs the vertical estimate of D and pixel sizes ismuch larger than in the horizontal direction (eg the low oblique and high obliquephotographs of Hawaii gure 4 table 5)
For the area of Limmen Bight ( gure 7) table 5 illustrates the improvement inthe estimate of distance and pixel size that can be obtained by re-estimating thelocation of the PC with greater accuracy Centre points in the catalogued data are
Astronaut photography as digital data for remote sensing 4429
plusmn05deg of latitude and longitude When the centre point was re-estimated to plusmn002degwe determined that the photograph was not taken as obliquely as rst thought(change in the estimate of the look angle t from 169 to 128deg table 5) When theauxiliary point was added to the calculations of Formula 3 the calculated look angleshrank further to 110deg indicating that this photograph was taken at very close toa nadir view Of course this improvement in accuracy could have also led to estimatesof greater obliquity and corresponding larger pixel sizes
We also tested the performance of the scale calculator with auxiliary point byestimating the corner and point locations on the photograph using a 11 000 000Operational Navigational Chart For this test we estimated our ability to readcoordinates from the map as plusmn002degand our error in nding the locations of thecorner points as plusmn015deg (this error varies among photographs depending on thedetail that can be matched between photo and map) For Limmen Bight ( gure 7)the mean diVerence between map estimates and calculator estimates for four pointswas 031deg (SD=018 n=8) For a photograph of San Francisco Bay (STS062-151-291) the mean diVerence between map estimates and calculator estimates forfour points was 0064deg (SD=018 n=8) For a photograph of San Francisco Bay(STS062-151-291) the mean diVerence between map estimates and calculator estim-ates for eight points was 0196deg (SD=0146 n=16) Thus in one case the calculatorestimates were better than our estimate of the error in locating corner points on themap It is reasonable to expect that for nadir-viewing photographs the calculatorused with an auxiliary point can estimate locations of the edges of a photograph towithin plusmn03deg
55 Empirical con rmation of spatial resolution estimatesAs stated previously a challenge to estimating system-AWAR for astronaut
photography of Earth is the lack of suitable targets Small features in an image cansometimes be used as a check on the size of objects that can be successfully resolvedgiving an approximate value for GRD Similarly the number of pixels that make upthose features in the digitized image can be used to make an independent calculationof pixel size We have successfully used features such as roads and airport runwaysto make estimates of spatial scale and resolution (eg Robinson et al 2000c) Whilerecognizing that the use of linear detail in an image is a poor approximation to abar target and that linear objects smaller than the resolving power can often bedetected (Charman 1965) few objects other than roads could be found to make anydirect estimates of GRD Thus roads and runways were used in the images ofHouston (where one can readily conduct ground veri cations and where a numberof higher-contrast concrete roadways were available) to obtain empirical estimatesof GRD and pixel size for comparison with table 5
In the all-digital ESC image of Houston ( gure 3(d )) we examined EllingtonField runway 4-22 (centre left of the image) which is 24384 mtimes457 m This runwayis approximately 6ndash7 pixels in width and 304ndash3094 pixels in length so pixelsrepresent an distance on the ground 7ndash8 m Using a lower contrast measure of astreet length between two intersections (2125 m=211 pixels) a pixel width of 101 mis estimated These results compare favourably with the minimum estimate of 81 mpixels using Formulation 1 (table 5) For an estimate of GRD that would be morecomparable to aerial photography of a line target the smallest street without treecover that could be clearly distinguished on the photograph was 792 m wide Thesmallest non-street object (a gap between stages of the Saturn rocket on display in
J A Robinson et al4430
Tab
le5
App
lica
tio
nof
met
hod
sfo
res
tim
ati
ng
spati
alre
solu
tio
no
fast
ron
aut
ph
oto
gra
ph
sin
clu
din
gF
orm
ula
tion
s12
and
3Im
pro
ved
cen
tre
po
int
esti
mat
esan
dauxilia
ryp
oin
tca
lcu
lati
ons
are
sho
wn
for
gure
7V
aria
ble
sas
use
din
the
text
fle
ns
foca
lle
ngt
hH
al
titu
de
PC
ph
oto
gra
ph
cen
tre
poin
tSN
sp
ace
craft
nadir
po
int
D
dis
tance
cover
edo
nth
egr
oun
d
Pd
ista
nce
on
the
grou
nd
repre
sen
ted
by
apix
el
OVse
td
ista
nce
bet
wee
nP
Cand
SNusi
ng
the
grea
tci
rcle
form
ula
t
look
angle
away
from
SN
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
2L
ake
Eyre
ST
S093
-717
-80
250
276
shy29
0137
0shy
28
413
71
607
11
767
413
7609
ndash64
212
013
760
9ndash64
7121
124
ST
S082
-754
-61
250
617
shy28
5137
0shy
28
113
50
135
726
1201
18
013
8ndash14
827
517
9138ndash15
3277
294
Fig
ure
3H
ou
sto
nS
TS
062
-97-
151
4029
829
5shy
95
530
5shy
95
4410
78
8112
20
5348ndash58
990
1205
351
ndash63
8951
10
36
ST
S094
-737
-28
100
294
295
shy
95
528
3shy
95
5162
31
1133
24
415
8ndash20
334
724
3158ndash20
7351
390
ST
S055
-71-
41250
302
295
shy
95
028
2shy
93
766
412
8192
32
5736
ndash84
715
232
273
9ndash85
7154
185
S73
E51
45300
2685
295
shy
95
024
78
0F
igu
re4
Haw
aii
ST
S069
-701
-65
250
370
195
shy
155
520
4shy
153
881
415
7204
28
9876
ndash98
917
928
688
0ndash10
9181
210
ST
S069
-701
-70
250
370
210
shy
157
519
2shy
150
781
415
7738
63
414
9ndash23
336
760
7148ndash55
6394
10
63
ST
S069
-729
-27
100
398
200
shy
156
620
5shy
148
2219
42
1867
65
432
8ndash13
10
158
62
4323+
311
N
A
Astronaut photography as digital data for remote sensing 4431
Tab
le5
(Cont
inued
)
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
7L
imm
enB
igh
tS
TS
093
-704
-50
250
283
shy15
0135
0shy
15
013
58
623
120
859
16
963
0ndash67
3125
16
863
0ndash67
612
613
2P
CR
e-es
tim
ated
shy14
733
1352
67
64
5128
623
ndash65
5123
128
623
ndash65
712
3127
Auxilia
ryP
oin
t611
062
3ndash65
112
312
5F
igu
re10
N
ear
Au
stin
T
XS
TS
095
-716
-46
250
543
30
5shy
975
28
6shy
1007
120
230
375
34
613
5ndash157
281
34
1136ndash16
128
636
0
1 All
pho
togr
aphs
exce
pt
on
eta
ken
wit
hH
ass
elbla
dca
mer
aso
dis
tan
ceacr
oss
the
image
d=
55m
m
for
S73
E51
45
taken
wit
hel
ectr
onic
still
cam
erad=
276
5m
mho
rizo
nta
l2 P
Cand
SN
are
giv
enin
the
form
(lati
tud
elo
ngit
ude)
deg
rees
w
ith
neg
ativ
en
um
ber
sre
pre
senti
ng
south
and
wes
tA
ccura
cyo
fP
Cis
plusmn0
5and
SN
isplusmn
01
as
reco
rded
inth
eA
stro
naut
Ph
oto
gra
phy
Data
base
ex
cep
tw
her
en
ote
dth
atce
ntr
ep
oin
tw
asre
-est
imate
dw
ith
grea
ter
accu
racy
3 C
alc
ula
ted
usi
ng
the
mea
nofth
em
inim
um
an
dm
axim
um
esti
mate
sof
DF
orm
ula
tion
2co
nsi
der
sD
inth
ed
irec
tion
per
pen
dic
ula
rto
the
Pri
nci
pal
Lin
eo
nly
4 C
alc
ula
ted
inho
rizo
nta
lan
dve
rtic
aldir
ecti
on
sb
ut
still
ass
um
ing
geom
etry
of
gure
6(B
)F
irst
nu
mb
eris
the
mea
no
fth
em
inim
um
Dat
the
bott
om
of
the
ph
oto
and
the
maxi
mum
Dat
the
top
of
the
ph
oto
(range
show
nin
the
tab
le)
and
isco
mpara
ble
toF
orm
ula
tio
n2
Sec
ond
num
ber
isth
em
ean
of
Des
tim
ated
alon
gth
eP
rin
cip
alline
and
alo
ng
the
ou
tsid
eed
ge
(range
not
show
n)
5 Alt
itu
de
isap
pro
xim
ate
for
mis
sion
as
exac
tti
me
pho
togr
ap
hw
asta
ken
and
corr
esp
on
din
gn
adir
data
are
no
tava
ilab
le
Lack
of
nad
irdata
pre
ven
tsap
plica
tion
of
Form
ula
tion
s2
an
d3
6 Au
xilliar
ypo
int
add
edw
as
shy14
80
lati
tud
e13
51
2lo
ngit
ude
shy46
5degro
tati
on
angle
in
stru
ctio
ns
for
esti
mati
ng
the
rota
tio
nan
gle
para
met
erare
conta
ined
as
par
tof
the
scal
eca
lcu
lato
rsp
read
shee
tdo
cum
enta
tio
n
J A Robinson et al4432
a park at Johnson Space Center) that could clearly be distinguished on the photo-graph was 853 m wide
For the photograph of Houston taken with a 250 mm lens ( gure 3(C)) anddigitized from second generation lm at 2400 ppi (106 mm pixel Otilde 1 ) Ellington Fieldrunway 4-22 is 3 pixels in width and 1613 pixels in length so pixels represent151ndash152 m on the ground These results compare favourably with the minimumestimate of 152 m using Formulation 2 and 154ndash185 m pixels using Formulation 3(table 5) For an estimate of GRD using an 8times8 inch print (1369 enlargement) and4times magni cation the smallest street that could clearly be distinguished was 822 mwide the same feature could barely be distinguished on the digitized image
We also made an empirical estimate of spatial resolution for lower contrastvegetation boundaries By clearing forest so that a pattern would be visible to landingaircraft a landowner outside Austin Texas (see also aerial photo in Lisheron 2000)created a target that is also useful for evaluating spatial resolution of astronautphotographs The forest was selectively cleared in order to spell the landownerrsquosname lsquoLUECKErsquo with the remaining trees ( gure 10) According to local surveyorswho planned the clearing the plan was to create letters that were 3100 fttimes1700 ft(9449 mtimes5182 m) Photographed at a high altitude relative to most Shuttle missions(543 km) with a 250 mm lens Formula 3 predicts that each pixel would represent anarea 286 mtimes360 m on the ground (table 5) When original lm was digitized at2400 ppi (106 mm pixel Otilde 1 ) letters correspond to 294times188 pixels for a comparablepixel size of 27ndash32 m
6 Summary and conclusionsAstronaut photographs can be an excellent source of data for remote sensing
applications Best-case resolutions are similar to that for Landsat or SPOT withpixels as small as lt10 m It was estimated that of images taken to date 504 hadlens and obliquity characteristics that make them potential candidates for remotesensing information Digitized astronaut photographs can be overlaid with othersatellite data using GIS (Eckardt et al 2000) or used to ll in gaps in time serieswhen other imagery is not available As a source of public-domain information theycan be very useful for scientists who do not have access to satellite imagery eitherbecause of the expense of image acquisition (to the end user) or the computersystems needed for processing satellite images These diVerences in image acquisitioncosts (to the user) are summarized in table 6
Searching of the complete database of NASA astronaut photography includinglow-spatial-resolution browse images is available via the Web (OYce of EarthSciences 2000) This provides nearly global access for identifying images that willcontribute to a speci c scienti c project For the most detailed studies digitalproducts posted to the web will not be of suYcient quality To date we have madeit a practice of digitizing small numbers of images when requested by scientists atno charge Such requests (including a description of the project involved) can bemade through the authors or using the contact information listed on the websiteInvestigators with needs for larger numbers of digital images have also been servedthrough collaborative agreements
In our experience scientists that nd the data most valuable are those in develop-ing countries those studying areas that have not usually been targets for the majorsatellites those having diYculty in nding low-cloud images those interested inconstructing time series or those interested in using a large number of images
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
J A Robinson et al4410
Tab
le1
Det
ails
on
the
cam
eras
use
dfo
rast
ron
aut
ph
oto
gra
phy
of
Eart
h
Data
incl
ud
ep
ho
togr
aphy
from
all
NA
SA
hu
man
spac
eig
ht
mis
sion
sth
rough
ST
S-9
6(J
un
e19
99
378
461
tota
lfr
am
esin
the
dat
abas
e)1
Imag
esi
zeT
ypic
al
lense
sN
um
ber
of
Cam
era
make
(mm
timesm
m)
use
d(m
m)
ph
oto
gra
ph
sP
erio
dof
use
No
tes
Has
selb
lad
55times
55
40
50100
2502
288
494
1963ndashp
rese
nt
Lin
ho
f10
5times120
90
250
29
373
1984ndashp
rese
nt
No
won
lyo
wn
occ
asio
nal
lyM
ult
ispec
tral
Cam
era
(S19
0A
)57times
57152
32
913
1973ndash19
74S
kyla
b
xed
-mou
nt3
cam
era
(NA
SA
1974
)E
art
hT
erra
inC
amer
a(S
190B
)11
5times11
5457
5478
1973ndash19
74S
kyla
b
xed
-mou
nt3
cam
era
(NA
SA
1974
)R
ole
iex
55times
55
50
100250
4352
1990ndash19
92L
arg
eF
orm
at
Cam
era4
229
times457
305
2143
1984
Mis
sio
nST
Sndash41G
only
(Mer
kel
etal
198
5)
Maure
r55
times55
76
80818
1961ndash19
68
1 Sm
all-f
orm
atca
mer
as
(35
mm
lm
and
ES
C)
are
no
tin
clud
edin
this
table
bec
ause
of
the
larg
enu
mb
erof
len
ses
and
con
gu
rati
on
sth
at
have
bee
nu
sed
and
bec
au
seth
ese
data
are
no
tcu
rren
tly
incl
uded
inth
edata
bas
efo
rall
mis
sion
s2 L
ense
suse
dfo
rH
ass
elbla
dm
anufa
cture
dby
Carl
Zei
ss
Dis
tago
nC
F4
(40
mm
)D
ista
gon
CF
4(5
0m
m)
Dis
tagon
CF
35
(100
mm
)D
ista
go
nC
F5
6(2
50
mm
)S
tand
ard
Has
selb
lad
len
ses
ow
nw
ill
chan
ge
inth
eyea
r20
00
toD
ista
gon
FE
28
(50
mm
)D
ista
go
nF
E2
(110
mm
)T
ele-
Tes
sar
FE
4(2
50m
m)
and
Tel
e-T
essa
rF
E4
(350
mm
)3 T
hes
eca
mer
as
wer
ein
xed
mou
nti
ngs
on
the
outs
ide
of
the
Skyla
bsp
ace
stati
on
A
ltho
ugh
not
hand
hel
d
the
lm
isca
talo
gued
wit
ho
ther
ast
ron
aut
ph
oto
gra
ph
yan
dis
incl
uded
her
efo
rco
mp
lete
nes
s4 A
NA
SA
-des
igned
cam
era
wit
hatt
itud
ere
fere
nce
syst
emfo
rd
eter
min
ing
pre
cise
nadir
poin
ts
Dat
ais
no
tcu
rren
tly
inco
rpora
ted
into
on
lin
ed
atab
ase
bu
tis
cata
loged
by
Mer
kel
etal
(1
985
)
Astronaut photography as digital data for remote sensing 4411
(a) (b)
(c) (d)
Figure 3 Example of the eVect of lens focal length on resolving power Hasselblad imagesof Houston were all taken from an altitude of 294 kmndash302 km with diVerent lenses(a) 40 mm STS062-97-151 (b) 100 mm STS094-737-28 (c) 250 mm STS055-71-41 Forcomparison an Electronic Still Camera image with 300 mm lens at 269 km altitude isalso shown (d ) S73E5145
timestamp on each frame at the time of exposure Cameras are serviced between ights Occasionally there is enough volume and mass allowance so that a Linhof5times4 inch (127times101 mm) format camera can be own Nikon 35 mm cameras are own routinely also because of proven reliability Electronic still cameras (ESC)were tested for Earth photography beginning in 1992 (Lulla and Holland 1993) Inan ESC a CCD (charge-coupled device) is used as a digital replacement for lmrecording the image projected inside the camera An ESC (consisting of a KodakDCS 460c CCD in a Nikon N-90S body) was added as routine equipment forhandheld photographs in 1995 ESCs have also been operated remotely to captureand downlink Earth images through a NASA-sponsored educational programme(EarthKAM) Discussion of CCD spatial array and radiometric sensitivity arebeyond the scope of this paper but are summarized by Robinson et al (2000b) The
J A Robinson et al4412
format of the lm (or CCD) and image size projected onto the lm (or CCD) aresummarized for all the diVerent cameras own in table 2
34 L ook angle or obliquityNo handheld photographs can be considered perfectly nadir they are taken at a
variety of look angles ranging from near vertical ( looking down at approximatelythe nadir position of the spacecraft ) to high oblique (images that include the curvatureof Earth) Imaging at oblique look angles leads to an image where scale degradesaway from nadir A set of views of the same area from diVerent look angles is shownin gure 4 The rst two shots of the island of Hawaii were taken only a few secondsapart and with the same lens The third photograph was taken on a subsequentorbit and with a shorter lens The curvature of the Earth can be seen in the upperleft corner
Obliquity can be described qualitatively ( gure 4) or quantitatively as the lookangle (t gure 1 calculations described in formulation 2 (sect52)) Because obliquityand look angle have such a dramatic in uence on the footprint we summarize thedatabase characteristics relative to these two parameters Figure 5 is a breakdownof the spatial resolution characteristics of low oblique and near vertical photographsin the NASA Astronaut Photography database Number of photographs are grouped(1) by calculated values for look angle (t ) and (2) by altitude After observing theoverlap between near vertical and low oblique classes we are currently restructuringthis variable (lsquotiltrsquo) in the database to provide a measure of t when available Userswill still be able to do searches based on the qualitative measures but these measureswill be more closely tied to actual look angle
341 Obliquity and georeferencing digitised photographsOften the rst step in a remote sensing analysis of a digitized astronaut photo-
graph is to georeference the data and resample it to conform to a known mapprojection Details and recommendations for resampling astronaut photographydata are provided by Robinson et al (2000a 2000c) and a tutorial is also available(McRay et al 2000) Slightly oblique photographs can be geometrically correctedfor remote sensing purposes but extremely oblique photographs are not suited forgeometric correction When obliquity is too great the spatial scale far away fromnadir is much larger than the spatial scale closer to nadir resampling results areunsuitable because pixels near nadir are lost as the image is resampled while manypixels far away from nadir are excessively replicated by resampling
To avoid the generation of excess pixels during georeferencing the pixel sizes ofthe original digitized image should be smaller than the pixels in the nal resampledimage Calculations of original pixel size using methods presented below can beuseful in ensuring meaningful resampling For slightly oblique images formulation 3(sect53) can be used to estimate pixel sizes at various locations in a photograph (nearnadir and away from nadir) and these pixel sizes then used to determine a reasonablepixel scale following resampling
4 Other characteristics that in uence spatial resolutionBeyond basic geometry many other factors internal to the astronaut photography
system impact the observed ground resolved distance in a photograph The lensesand cameras already discussed are imperfect and introduce radiometric degradations(Moik 1980)Vignetting (slight darkening around the edges of an image) occurs in
Astronaut photography as digital data for remote sensing 4413
Tab
le2
Max
imu
mpo
ssib
led
igit
alsp
atia
lre
solu
tio
n(p
ixel
size
inm
)fo
rca
mer
asy
stem
scu
rren
tly
use
dby
ast
ron
auts
top
ho
togr
aph
eart
hb
ase
do
nth
ege
om
etry
of
alt
itud
ele
ns
and
lm
size
C
alc
ula
tion
sas
sum
eth
atco
lour
tran
spar
ency
lm
isdig
itiz
edat
2400
ppi
(pix
els
per
inch
10
6m
mpix
elOtilde
1 )
Min
imu
mgro
un
ddis
tance
repre
sente
dby
1p
ixel
(m)
As
afu
nct
ion
of
alti
tude1
Imag
esi
zeM
inim
um
alt
itud
eM
edia
nalt
itu
de
Maxi
mum
alt
itud
eC
amer
asy
stem
mm
timesm
mp
ixel
s2(2
22k
m)
(326
km
)(6
11
km
)
Lin
ho
f125
mm
90
mm
lens
105
times120
8978times
11
434
261
383
718
250
mm
lens
94
138
259
Has
selb
lad
70
mm
100
mm
lens
55times
55
5198
times519
823
534
564
725
0m
mle
ns
94
138
259
Nik
on
35m
m30
0m
mle
ns
36times
24
3308
times217
478
115
216
400
mm
lens
59
86
162
ESC
35m
m18
0m
mle
ns3
27
6times
18
530
69times
204
311
116
330
630
0m
mle
ns
67
98
184
400
mm
lens
50
74
138
1 Alt
itud
essh
ow
nar
em
inim
um
m
edia
nan
dm
axim
um
thro
ugh
ST
S-8
9(J
anuary
199
8N
=84
mis
sion
s)
2 Act
ual
pix
els
for
ES
C
oth
erw
ise
assu
med
dig
itiz
ati
on
at240
0pp
i(p
ixel
sp
erin
ch)
equ
alto
10
6m
mpix
elOtilde
1 3 T
he
mo
stco
mm
on
con
gura
tion
for
Eart
hph
oto
gra
ph
yas
part
of
the
Eart
hK
AM
pro
ject
J A Robinson et al4414
Figure 4 De nitions of look angle for photographs taken from low orbit around EarthThe example photographs of Hawaii were all taken from approximately the samealtitude (370ndash398 km) and with the same (250 mm) lens on a Hasselblad camera(STS069-701-69 STS069-701-70 and STS069-729-27 [100 mm lens])
astronaut photography due to light path properties of the lenses used A lack of atness of the lm at the moment of exposure (because cameras used in orbit donot incorporate a vacuum platen) also introduces slight degradations A numberof additional sources of image degradation that would aVect GRD of astronautphotographs are listed
41 Films and processing411 Colour reversal lms
Most lms used in NASA handheld cameras in orbit have been E-6 processcolour reversal lms (Ektachromes) that have a nominal speed of 64 to 100 ASAalthough many other lms have been own (table 3) Reversal lms are used becausedamage from radiation exposure while outside the atmospheric protection of Earthis more noticeable in negative lms than in positive lms (Slater 1996) The choiceof ASA has been to balance the coarser grains ( lower lm resolving power) of high-speed lms with the fact that slower lms require longer exposures and thus areaVected more by vehicle motion (approximately 73 km s Otilde 1 relative to Earthrsquos surfacefor the Space Shuttle) Extremely fast lms (gt400 ASA) are also more susceptibleto radiation damage in orbit (particularly during long-duration or high-altitudemissions Slater 1996) and have not traditionally been used for Earth photography The manufacturer stock numbers identifying the lm used are available for eachimage in the Astronaut Photography Database (OYce of Earth Sciences 2000)
412 Colour infrared lmsColour infrared lm (CIR) was used during an Earth-orbiting Apollo mission
(Colwell 1971) in the multispectral S-190A and the high-resolution S-190B camera
Astronaut photography as digital data for remote sensing 4415
Figure 5 (a) Distributions of astronaut photographs by look angle oV-nadir (calculated fromcentre point and nadir point using Formulation 2) Data included for all photographsfor which centre and nadir points are known with high oblique look angles (n=55 155) excluded (Total sample shown is 108 293 of 382 563 total records in thedatabase when compiled For records where nadir data was not available number ofobservations for qualitative look angles are near vertical 14 086 low oblique 115 834undetermined 89 195) (b) Altitude distribution for astronaut photographs of Earthtaken with the Hasselblad camera Data included for Hasselblad camera only focallength determined with high oblique look angles excluded (Total sample shown is159 046 of 206 971 Hasselblad records with known altitude and of 382 563 total recordsin the database when compiled For Hasselblad records with known altitude data notshown includes high oblique photographs using 100 mm and 250 mm lenses n=20 262and all look angles with other lenses n=27 663)
systems on Skylab (NASA 1974 Wilmarth et al 1977) and occasionally on Shuttlemissions and Shuttle-Mir missions (table 3) The CIR lm used is a three-layerAerochrome having one layer that is sensitive to re ected solar infrared energy toapproximately 900 nm (Eastman Kodak Company 1998) protective coatings onmost spacecraft windows also limit IR transmittance between 800 mm and 1000 nmThis layer is extremely sensitive to the temperature of the lm which createsunpredictable degradation of the IR signature under some space ight conditions
J A Robinson et al4416
Tab
le3
Det
ails
on
lm
suse
dfo
rast
ron
aut
pho
togr
aphy
of
Eart
h
Dat
ain
clud
ep
ho
togr
aphy
fro
mall
NA
SA
hum
an
spac
eig
ht
mis
sio
ns
thro
ugh
ST
S-9
6(J
un
e19
99
378
461
tota
lfr
ames
inth
edata
base
)F
ilm
sw
ith
lt50
00fr
am
esex
po
sed
and
lm
suse
dfo
r35
mm
cam
eras
only
are
no
tin
clud
ed
Res
olv
ing
Nu
mb
erof
Film
man
ufa
ctu
rer
AS
Ap
ow
er1
fram
esP
erio
dof
use
Cam
eras
Colo
ur
posi
tive
Kod
akA
eroch
rom
eII
(2447
2448
SO
-131S
O-3
68
)32
40ndash50
513
8196
8ndash19
85
Hass
elbla
dL
inh
of
Kod
akE
kta
chro
me
(5017
502
56
017
6118
SO
-121
64
50
113
932
196
5ndash19
68
Hass
elbla
dL
inh
of
Role
iex
SO
-217
Q
X-8
24
QX
-868
)19
81ndash1993
Mau
rer
Nik
on
Kod
akL
um
iere
(5046
5048
)100
105
743
199
4ndash19
98
Hass
elbla
dL
inho
fK
od
akE
lite
100
S(5
069
)100
41
486
199
6ndashp
rese
nt
Hass
elbla
d2
Fu
jiV
elvia
50C
S135
-36
32
13
882
1992ndash19
97H
ass
elbla
dN
iko
nK
od
akH
i-R
esolu
tio
nC
olo
r(S
O-2
42S
O-3
56
)10
096
74
1973ndash19
75H
ass
elbla
d
S19
0A
S190
B
Infr
are
dand
bla
ck-a
nd-w
hit
e
lms
Kod
akA
eroch
rom
eC
olo
rIn
frar
ed40
32
40
704
196
5ndashp
rese
nt2
Hass
elbla
dL
inh
of
Nik
on
S190
A
(8443
34
432443
)S
190B
Kod
akB
ampW
Infr
are
d(2
424
)32
10
553
197
3ndash19
74
S190
AK
od
akP
anato
mic
-XB
ampW
(SO
-022
)63
10
657
197
3ndash19
74
S190
A
1A
tlo
wco
ntr
ast
(ob
ject
back
grou
nd
rati
o16
1)
aspro
vid
edby
Ko
dak
and
list
edby
Sla
ter
etal
(1
983
256
)R
esolv
ing
pow
erw
as
dis
con
tin
ued
fro
mth
ete
chn
ical
test
sper
form
edb
yK
odak
inapp
roxim
atel
y19
93
an
dit
isn
ot
kn
ow
nif
curr
ent
lm
sh
ave
sim
ilar
reso
lvin
gpo
wer
too
lder
ver
sion
so
fsi
milar
lm
s(K
T
eite
lbaum
K
od
akp
erso
nal
com
mu
nic
atio
n)
Th
egra
nu
lari
tym
easu
reno
wpro
vid
edb
yK
od
ak
can
no
tb
ere
adily
con
ver
ted
toa
spat
ial
reso
luti
on
2 The
Lin
hof
was
use
dag
ain
on
ST
S-1
03in
Dec
emb
er19
99(d
ata
not
cata
logued
asof
the
pre
para
tion
of
this
tab
le)
usi
ng
Ko
dak
Elite
100S
lm
3 B
egin
nin
gin
July
1999
2443
colo
ur-
infr
are
d
lmpro
cess
was
chan
ged
from
EA
-5to
E-6
Astronaut photography as digital data for remote sensing 4417
413 Film duplicationThe original lm is archived permanently after producing about 20 duplicate
printing masters (second generation) Duplicate printing masters are disseminatedto regional archives and used to produce products (thirdndashfourth generation) for themedia the public and for scienti c use When prints slides transparencies anddigital products are produced for public distribution they are often colour adjustedto correspond to a more realistic look Digital products from second generationcopies can be requested by scienti c users (contact the authors for information onobtaining such products for a particular project) and these are recommended forremote sensing applications Care should be taken that the digital product acquiredfor remote sensing analysis has not been colour adjusted for presentation purposesBased on qualitative observations there is little visible increase in GRD in the thirdor fourth generation products There is a signi cant degradation in delity of colourin third and fourth generation duplicates because an increase in contrast occurswith every copy
42 Shutter speedsThe impact of camera motion (both due to the photographer and to the motion
of the spacecraft ) on image quality is determined by shutter speedmdash1250 to 1500second have been used because slower speeds record obvious blurring caused bythe rapid motion of the spacecraft relative to the surface of the Earth Generally1250 was used for ISO 64 lms (and slower) and after the switch to ISO 100 lmsa 1500 setting became standard New Hasselblad cameras that have been ownbeginning with STS-92 in October 2000 vary shutter speed using a focal plane shutterfor exposure bracketing (rather than varying aperture) and 1250 1500 and 11000second are used
Across an altitude range of 120ndash300 nautical miles (222ndash555 km) and orbitalinclinations of 285deg and 570deg the median relative ground velocity of orbitingspacecraft is 73 km s Otilde 1 At a nominal shutter speed of 1500 the expected blur dueto motion relative to the ground would be 146 m When cameras are handheld (asopposed to mounted in a bracket) blur can be reduced by physical compensationfor the motion (tracking) by the photographer many photographs show a level ofdetail indicating that blur at this level did not occur (eg gure 3(d )) Thus motionrelative to the ground is not an absolute barrier to ground resolution for handheldphotographs
43 Spacecraft windowsMost of the photographs in the NASA Astronaut Photography Database were
taken through window ports on the spacecraft used The transmittance of the windowport may be aVected by material inhomogeniety of the glass the number of layersof panes coatings used the quality of the surface polish environmentally inducedchanges in window materials (pressure loads thermal gradients) or deposited con-tamination Such degradation of the window cannot be corrected is diVerent foreach window and changes over time See Eppler et al (1996) and Scott (2000) fordiscussion of the spectral transmittance of the fused quartz window that is part ofthe US Laboratory Module (Destiny) of the International Space Station
44 Digitized images from lmWhen lm is scanned digitally the amount of information retained depends on
the spectral information extracted from the lm at the spatial limits of the scanner
J A Robinson et al4418
To date standard digitizing methodologies for astronaut photographs have not beenestablished and lm is digitized on a case-by-case basis using the equipment available
441 Digitizing and spatial resolutionLight (1993 1996) provided equations for determining the digitizing spatial
resolution needed to preserve spatial resolution of aerial photography based on thestatic resolving power (AWAR) for the system For lms where the manufacturer hasprovided data (Eastman Kodak 1998 K Teitelbaum personal communication) the resolving power of lms used for astronaut photography ranges from 32 lp mm Otilde 1to 100 lp mm Otilde 1 at low contrast (objectbackground ratio 161 table 3) The AWARfor the static case of the Hasselblad camera has been measured at high and lowcontrast (using lenses shown in table 1) with a maximum of approximately55 lp mm Otilde 1 (Fred Pearce unpublished data)
Based on the method of Light (1993 1996) the dimension of one spatial resolutionelement for a photograph with maximum static AWAR of 55 lp mm Otilde 1 would be18 mm lp Otilde 1 The acceptable range of spot size to preserve spatial information wouldthen be
18 mm
2 atilde 2aring scan spot size aring
18 mm
2(1)
and 6 mm aring scan spot size aring 9 mm Similarly for a more typical static AWAR of30 lp mm Otilde 1 (33 mm lpOtilde 1 low contrast Fred Pearce unpublished data) 11 mm aring scanspot sizearing 17 mm These scan spot sizes correspond to digitizing resolutions rangingfrom 4233ndash2822 ppi (pixels inch Otilde 1 ) for AWAR of 55 lp mm Otilde 1 and 2309ndash1494 ppi forAWAR of 30 lp mm Otilde 1 Scan spot sizes calculated for astronaut photography arecomparable to those calculated for the National Aerial Photography Program(9ndash13 mm Light 1996)
Widely available scanners that can digitize colour transparency lm currentlyhave a maximum spatial resolution of approximately 2400 ppi (106 mm pixel Otilde 1) Forexample we routinely use an Agfa Arcus II desktop scanner (see also Baltsavias1996) with 2400 ppi digitizing spatial resolution (2400 ppi optical resolution in onedirection and 1200 ppi interpolated to 2400 ppi in the other direction) and 2400 ppiwas used to calculate IFOV equivalents (tables 2 and 4) For some combinations oflens camera and contrast 2400 ppi will capture nearly all of the spatial informationcontained in the lm However for lm with higher resolving power thanEktachrome-64 for better lenses and for higher contrast targets digitizing at 2400 ppiwill not capture all of the spatial information in the lm
Improvements in digitizing technology will only produce real increases in IFOVto the limit of the AWAR of the photography system The incremental increase inspatial information above 3000 ppi (85 mm pixel Otilde 1 ) is not likely to outweigh thecosts of storing large images (Luman et al 1997) At 2400 ppi a Hasselblad frameis approximately 5200times5200 or 27 million pixels (table 2) while the same imagedigitized at 4000 ppi would contain 75 million pixels
Initial studies using astronaut photographs digitized at 2400 ppi (106 mm pixel Otilde 1 Webb et al 2000 Robinson et al 2000c) indicate that some GRD is lost comparedto photographic products Nevertheless the spatial resolution is still comparablewith other widely used data sources (Webb et al 2000) Robinson et al (2000c)found that digitizing at 21 mm pixel Otilde 1 provided information equivalent to 106 mmpixel Otilde 1 for identifying general urban area boundaries for six cities except for a single
Astronaut photography as digital data for remote sensing 4419
photo that required higher spatial resolution digitizing (it had been taken with ashorter lens and thus had less spatial resolution) Part of the equivalence observedin the urban areas study may be attributable to the fact that the atbed scannerused interpolates from 1200 ppi to 2400 ppi in one direction The appropriate digitiz-ing spatial resolution will in part depend on the scale of features of interest to theresearchers with a maximum upper limit set of a scan spot size of approximately6 mm (4233 ppi)
442 Digitizing and spectral resolutionWhen colour lm is digitized there will be loss of spectral resolution The three
lm emulsion layers (red green blue) have relatively distinct spectral responses butare fused together so that digitizers detect one colour signal and must convert itback into three (red green blue) digital channels Digital scanning is also subject tospectral calibration and reproduction errors Studies using digital astronaut photo-graphs to date have used 8 bits per channel However this is another parameterthat can be controlled when the lm is digitized We do not further address spectralresolution of digitally scanned images in this paper
45 External factors that in uence GRDAlthough not discussed in detail in this paper factors external to the spacecraft
and camera system (as listed by Moik (1980) for remote sensing in general) alsoimpact GRD These include atmospheric interference (due to scattering attenuationhaze) variable surface illumination (diVerences in terrain slope and orientation) andchange of terrain re ectance with viewing angle (bidirectional re ectance) For astro-naut photographs variable illumination is particularly important because orbits arenot sun-synchronous Photographs are illuminated by diVerent sun angles and imagesof a given location will have colour intensities that vary widely In addition theviewing angle has an eVect on the degree of object-to-backgroun d contrast andatmospheric interference
5 Estimating spatial resolution of astronaut photographsIn order to use astronaut photographs for digital remote sensing it is important
to be able to calculate the equivalent to an IFOVmdashthe ground area represented bya single pixel in a digitized orbital photograph The obliquity of most photographsmeans that pixel lsquosizesrsquo vary at diVerent places in an image Given a ground distanceD represented by a photograph in each direction (horizontal vertical ) an approxi-mate average pixel width (P the equivalent of IFOV) for the entire image can becalculated as follows
P=D
dS(2)
where D is the projected distance on the ground covered by the image in the samedirection as the pixel is measured d is the actual width of the image on the original lm (table 1) and S is the digitizing spatial resolution
Here we present three mathematical formulations for estimating the size of thefootprint or area on the ground covered by the image Example results from theapplication of all three formulations are given in table 5 The rst and simplestcalculation (formulation 1) gives an idea of the maximum spatial resolution attainableat a given altitude of orbit with a given lm format and a perfectly vertical (nadir)
J A Robinson et al4420
view downward Formulation 2 takes into account obliquity by calculating lookangle from the diVerence between the location on the ground represented at thecentre of the photograph and the nadir location of the spacecraft at the time thephotograph was taken ( gure 1) Formulation 3 describes an alternate solution tothe oblique look angle problem using coordinate-system transformations Thisformulation has been implemented in a documented spreadsheet and is availablefor download (httpeoljscnasagovsseopFootprintCalculatorxls) and is in theprocess of being implemented on a Web-based user interface to the AstronautPhotography Database (OYce of Earth Sciences 2000)
Although formulations 2 and 3 account for obliquity for the purposes of calcula-tion they treat the position of the spacecraft and position of the camera as one Inactuality astronauts are generally holding the cameras by hand (although camerasbracketed in the window are also used) and the selection of window position of theastronaut in the window and rotation of the camera relative to the movement ofthe spacecraft are not known Thus calculations using only the photo centre point(PC) and spacecraft nadir point (SN) give a locator ellipse and not the locations ofthe corners of the photograph A locator ellipse describes an estimated area on theground that is likely to be included in a speci c photograph regardless of the rotationof the lm plane about the camerarsquos optical axis ( gure 6)
Estimating the corner positions of the photo requires additional user input of asingle auxiliary pointmdasha location on the image that has a known location on theground Addition of this auxiliary point is an option available to users of thespreadsheet An example of the results of adding an auxiliary point is shown in gure 7 with comparisons of the various calculations in table 5
51 Formulation 1 Footprint for a nadir viewThe simplest way to estimate footprint size is to use the geometry of camera lens
and spacecraft altitude to calculate the scaling relationship between the image in the
Figure 6 Sketch representation of the locator ellipse an estimated area on the ground thatis likely to be included in a speci c photograph regardless of the rotation of the lmplane about the camerarsquos optical axis
Astronaut photography as digital data for remote sensing 4421
Figure 7 An example of the application of the Low Oblique Space Photo FootprintCalculator The photograph (STS093-704-50) of Limmen Bight Australia was takenfrom an altitude of 283 km using the Hasselblad camera with a 250 mm lens Mapused is 11 000 000 Operational Navigational Chart The distances shown equate toan average pixel size of 124 m for this photograph
lm and the area covered on the ground For a perfect nadir view the scalerelationship from the geometry of similar triangles is
d
D=
f
H(3)
where d=original image size D=distance of footprint on the ground f =focallength of lens and H=altitude Once D is known pixel size ( length=width) can becalculated from equation (2) These calculations represent the minimum footprintand minimum pixel size possible for a given camera system altitude and digitisingspatial resolution (table 2) Formulation 1 was used for calculating minimum pixelsizes shown in tables 2 and 4
By assuming digitizing at 2400 ppi (106 mm pixel Otilde 1 ) currently a spatial resolutioncommonly attainable from multipurpose colour transparency scanners (see sect441)this formulation was used to convert area covered to an IFOV equivalent for missionsof diVerent altitudes (table 2) Table 4 provides a comparison of IFOV of imagesfrom various satellites including the equivalent for astronaut photography Valuesin this table were derived by using formulation 1 to estimate the area covered becausea perfect nadir view represents the best possible spatial resolution and smallest eldof view that could be obtained
52 Formulation 2 Footprint for oblique views using simpli ed geometry and thegreat circle distance
A more realistic approach to determining the footprint of the photographaccounts for the fact that the camera is not usually pointing perfectly down at the
J A Robinson et al4422
Table 4 Comparison of pixel sizes (instantaneous elds of view) for satellite remote sensorswith pixel sizes for near vertical viewing photography from spacecraft Note that alldata returned from the automated satellites have the same eld of view while indicatedpixel sizes for astronaut photography are best-case for near vertical look angles See gure 5 for distributions of astronaut photographs of various look angles High-altitude aerial photography is also included for reference
Altitude PixelInstrument (km) width (mtimesm) Bands1
Landsat Multipectral Scanner2 (MSS 1974-) 880ndash940 79times56 4ndash5Landsat Thematic Mapper (TM 1982-) ~705 30times30 7Satellite pour lrsquoObservation de la Terre (SPOT 1986-) ~832
SPOT 1-3 Panchromatic 10times10 1SPOT 1-3 Multispectral (XS) 20times20 3
Astronaut Photography (1961-)Skylab3 (1973-1974 ) ~435
Multispectral Camera (6 camera stations) 17ndash19times17ndash19 3ndash8Earth Terrain Camera (hi-res colour) 875times875 3
Space Shuttle5 (1981-) 222ndash611Hasselblad 70 mm (250mm lens colour) vertical view 9ndash26times9ndash26 3Hasselblad 70 mm (100mm lens colour) vertical view 23ndash65times23ndash65 3Nikon 35 mm (400mm lens colour lm or ESC) vertical 5ndash16times5ndash16 3
High Altitude Aerial Photograph (1100 000) ~12 2times2 3
1Three bands listed for aerial photography obtainable by high-resolution colour digitizingFor high-resolution colour lm at 65 lp mmOtilde 1 (K Teitelbaum Eastman Kodak Co personalcommunication) the maximum information contained would be 3302 ppi
2Data from Simonett (1983Table 1-2)3Calculated as recommended by Jensen (19831596) the dividend of estimated ground
resolution at low contrast given in NASA (1974179-180) and 244Six lters plus colour and colour infrared lm (NASA 19747) 5Extracted from table 2
nadir point The look angle (the angle oV nadir that the camera is pointing) can becalculated trigonometrically by assuming a spherical earth and calculating the dis-tance between the coordinates of SN and PC ( gure 1) using the Great Circledistance haversine solution (Sinnott 1984 Snyder 198730ndash32 Chamberlain 1996)The diVerence between the spacecraft centre and nadir point latitudes Dlat=lat 2 shy lat 1 and the diVerence between the spacecraft centre and nadir pointlongitudes Dlon=lon 2 shy lon 1 enter the following equations
a=sin2ADlat
2 B+cos(lat 1) cos(lat 2) sin2ADlon
2 B (4)
c=2 arcsin(min[1 atilde a]) (5)
and oVset=R c with R the radius of a spherical Earth=6370 kmThe look angle (t ) is then given by
t=arctanAoffset
H B (6)
Assuming that the camera was positioned so that the imaginary line between thecentre and nadir points (the principal line) runs vertically through the centre of thephotograph the distance between the geometric centre of the photograph (principalpoint PP) and the top of the photograph is d2 ( gure 8(a)) The scale at any point
Astronaut photography as digital data for remote sensing 4423
(a) (b)
Figure 8 The geometric relationship between look angle lens focal length and the centrepoint and nadir points of a photograph (see also Estes and Simonett 1975) Note thatthe Earthrsquos surface and footprint represented in gure 1 are not shown here only arepresentation of the image area of the photograph Variables shown in the gurelook angle t focal length f PP centre of the image on the lm SN spacecraft nadird original image size (a) The special case where the camera is aligned squarely withthe isometric parallel In this case tilt occurs along a plane parallel with the centre ofthe photograph When the conditions are met calculations using Formulation 3 willgive the ground coordinates of the corner points of the photograph (b) The generalrelationship between these parameters and key photogrammetric points on the photo-graph For this more typical case coordinates of an additional point in the photographmust be input in order to adjust for twist of the camera and to nd the corner pointcoordinates
in the photograph varies as a function of the distance y along the principal linebetween the isocentre and the point according to the relationship
d
D=
f shy ysint
H(7)
(Wong 1980 equation (214) H amp the ground elevation h) Using gure 8(a) at thetop of the photo
y= f tanAt
2B+d
2(8)
and at the bottom of the photo
y= f tanAt
2Bshyd
2(9)
Thus for given PC and SN coordinates and assuming a photo orientation as in gure 8(a) and not gure 8(b) we can estimate a minimum D (using equations (7)and (8)) and a maximum D (using equations (7) and (9)) and then average the twoto determine the pixel size (P ) via equation (2)
J A Robinson et al4424
53 Formulation 3 T he L ow Oblique Space Photo Footprint CalculatorThe Low Oblique Space Photo Footprint Calculator was developed to provide
a more accurate estimation of the geographic coordinates for the footprint of a lowoblique photo of the Earthrsquos surface taken from a human-occupied spacecraft inorbit The calculator performs a series of three-dimensional coordinate transforma-tions to compute the location and orientation of the centre of the photo exposureplane relative to an earth referenced coordinate system The nominal camera focallength is then used to create a vector from the photorsquos perspective point througheach of eight points around the parameter of the image as de ned by the formatsize The geographical coordinates for the photo footprint are then computed byintersecting these photo vectors with a spherical earth model Although more sophist-icated projection algorithms are available no signi cant increase in the accuracy ofthe results would be produced by these algorithms due to inherent uncertainties inthe available input data (ie the spacecraft altitude photo centre location etc)
The calculations were initially implemented within a Microsoft Excel workbookwhich allowed one to embed the mathematical and graphical documentation nextto the actual calculations Thus interested users can inspect the mathematical pro-cessing A set of error traps was also built into the calculations to detect erroneousresults A summary of any errors generated is reported to the user with the resultsof the calculations Interested individuals are invited to download the Excel work-book from httpeoljscnasagovsseopFootprintCalculatorxls The calculations arecurrently being encoded in a high-level programming language and should soon beavailable alongside other background data provided for each photograph at OYceof Earth Sciences (2000)
For the purposes of these calculations a low oblique photograph was de ned as
one with the centre within 10 degrees of latitude and longitude of the spacecraftnadir point For the typical range of spacecraft altitudes to date this restricted thecalculations to photographs in which Earthrsquos horizon does not appear (the generalde nition of a low oblique photograph eg Campbell 199671 and gure 4)
531 Input data and resultsUpon opening the Excel workbook the user is presented with a program intro-
duction providing instructions for using the calculator The second worksheet tablsquoHow-To-Usersquo provides detailed step-by-step instructions for preparing the baselinedata for the program The third tab lsquoInput-Output rsquo contains the user input eldsand displays the results of the calculations The additional worksheets contain theactual calculations and program documentation Although users are welcome toreview these sheets an experienced user need only access the lsquoInput-Outputrsquo
spreadsheetTo begin a calculation the user enters the following information which is available
for each photo in the NASA Astronaut Photography Database (1) SN geographicalcoordinates of spacecraft nadir position at the time of photo (2) H spacecraftaltitude (3) PC the geographical coordinates of the centre of the photo (4) f nominal focal length and (5) d image format size The automatic implementationof the workbook on the web will automatically enter these values and completecalculations
For more accurate results the user may optionally enter the geographic co-ordinates and orientation for an auxiliary point on the photo which resolves thecamerarsquos rotation uncertainty about the optical axis The auxiliary point data must
Astronaut photography as digital data for remote sensing 4425
be computed by the user following the instruction contained in the lsquoHow-To-Usersquotab of the spreadsheet
After entering the input data the geographic coordinates of the photo footprint(ie four photo corner points four points at the bisector of each edge and the centreof the photo) are immediately displayed below the input elds along with any errormessages generated by the user input or by the calculations ( gure 7) Although
results are computed and displayed they should not be used when error messagesare produced by the program The program also computes the tilt angle for each ofthe photo vectors relative to the spacecraft nadir vector To the right of the photofootprint coordinates is displayed the arc distance along the surface of the sphere
between adjacent computed points
532 Calculation assumptions
The mathematical calculations implemented in the Low Oblique Space PhotoFootprint Calculator use the following assumptions
1 The SN location is used as exact even though the true value may vary by upto plusmn01 degree from the location provided with the photo
2 The spacecraft altitude is used as exact Although the determination of the
nadir point at the instant of a known spacecraft vector is relatively precise(plusmn115times10 Otilde 4 degrees) the propagator interpolates between sets of approxi-mately 10ndash40 known vectors per day and the time code recorded on the lmcan drift Thus the true value for SN may vary by up to plusmn01 degree fromthe value provided with the photo
3 The perspective centre of the camera is assumed to be at the given altitudeover the speci ed spacecraft nadir location at the time of photo exposure
4 The PC location is used as exact even though the true value may vary by upto plusmn05deg latitude and plusmn05deg longitude from the location provided with the
photo5 A spherical earth model is used with a nominal radius of 6 372 16154 m (a
common rst-order approximation for a spherical earth used in geodeticcomputations)
6 The nominal lens focal length of the camera lens is used in the computations(calibrated focal length values are not available)
7 The photo projection is based on the classic pin-hole camera model8 No correction for lens distortion or atmospheric re ection is made
9 If no auxiliary point data is provided the lsquoTop of the Imagersquo is orientedperpendicular to the vector from SN towards PC
533 T ransformation from Earth to photo coordinate systemsThe calculations begin by converting the geographic coordinates ( latitude and
longitude) of the SN and PC to a Rectangular Earth-Centred Coordinate System(R-Earth) de ned as shown in gure 9 (with the centre of the Earth at (0 0 0))
Using the vector from the Earthrsquos centre through SN and the spacecraft altitude thespacecraft location (SC) is also computed in R-Earth
For ease of computation a Rectangular Spacecraft-Centred Coordinate System(R-Spacecraft ) is de ned as shown in gure 9 The origin of R-Spacecraft is located
at SC with its +Z-axis aligned with vector from the centre of the Earth throughSN and its +X-axis aligned with the vector from SN to PC ( gure 9) The speci c
J A Robinson et al4426
Figure 9 Representation of the area included in a photograph as a geometric projectiononto the Earthrsquos surface Note that the focal length (the distance from the SpaceShuttle to the photo) is grossly overscale compared to the altitude (the distance fromthe Shuttle to the surface of the Earth
rotations and translation used to convert from the R-Earth to the R-Spacecraft arecomputed and documented in the spreadsheet
With the mathematical positions of the SC SN PC and the centre of the Earthcomputed in R-Spacecraft the program next computes the location of the camerarsquosprincipal point (PP) The principle point is the point of intersection of the opticalaxis of the lens with the image plane (ie the lm) It is nominally positioned at a
Astronaut photography as digital data for remote sensing 4427
distance equal to the focal length from the perspective centre of the camera (whichis assumed to be at SC) along the vector from PC through SC as shown in gure 9
A third coordinate system the Rectangular Photo Coordinate System (R-Photo)is created with its origin at PP its XndashY axial plane normal to the vector from PCthrough SC and its +X axis aligned with the +X axis of R-Spacecraft as shownin gure 9 The XndashY plane of this coordinate system represents the image plane ofthe photograph
534 Auxiliary point calculationsThe calculations above employ a critical assumption that all photos are taken
with the lsquotop of the imagersquo oriented perpendicular to the vector from the SN towardsPC as shown in gure 9 To avoid non-uniform solar heating of the external surfacemost orbiting spacecraft are slowly and continually rotated about one or more axesIn this condition a ight crew member taking a photo while oating in microgravitycould orient the photo with practically any orientation relative to the horizon (seealso gure 8(b)) Unfortunately since these photos are taken with conventionalhandheld cameras there is no other information available which can be used toresolve the photorsquos rotational ambiguity about the optical axis other then the photoitself This is why the above assumption is used and the footprint computed by thiscalculator is actually a lsquolocator ellipsersquo which estimates the area on the ground whatis likely to be included in a speci c photograph (see gure 6) This locator ellipse ismost accurate for square image formats and subject to additional distortion as thephotograph format becomes more rectangular
If the user wants a more precise calculation of the image footprint the photorsquosrotational ambiguity about the optical axis must be resolved This can be done inthe calculator by adding data for an auxiliary point Detailed instructions regardinghow to prepare and use auxiliary point data in the computations are included in thelsquoHow-To-Usersquo tab of the spreadsheet Basically the user determines which side ofthe photograph is top and then measures the angle between the line from PP to thetop of the photo and from PP to the auxiliary point on the photo ( gure 9)
If the user includes data for an auxiliary point a series of computations arecompleted to resolve the photo rotation ambiguity about the optical axis (ie the
+Z axis in R-Photo) A vector from the Auxiliary Point on the Earth (AE) throughthe photograph perspective centre ( located at SC) is intersected with the photo imageplane (XndashY plane of R-Photo) to compute the coordinates of the Auxiliary Point onthe photo (AP) in R-Photo A two-dimensional angle in the XndashY plane of R-Photofrom the shy X axis to a line from PP to AP is calculated as shown in gure 9 The
shy X axis is used as the origin of the angle since it represents the top of the photoonce it passes through the perspective centre The diVerence between the computedangle and the angle measured by the user on the photo resolves the ambiguity inthe rotation of the photo relative to the principal line ( gures 7 and 9) Thetransformations from R-Spacecraft and R-Photo are then modi ed to include anadditional rotation angle about the +Z axis in R-Photo
535 lsquoFootprint rsquo calculationsThe program next computes the coordinates of eight points about the perimeter
of the image format (ie located at the four photo corners plus a bisector pointalong each edge of the image) These points are identi ed in R-Photo based uponthe photograph format size and then converted to R-Spacecraft Since all computa-tions are done in orthogonal coordinate systems the R-Spacecraft to R-Photo
J A Robinson et al4428
rotation matrix is transposed to produce an R-Photo to R-Spacecraft rotation matrixOnce in R-Spacecraft a unit vector from each of the eight perimeter points throughthe photo perspective centre (the same point as SC) is computed This provides thecoordinates for points about the perimeter of the image format with their directionvectors in a common coordinate system with other key points needed to computethe photo footprint
The next step is to compute the point of intersection between the spherical earthmodel and each of the eight perimeter point vectors The scalar value for eachperimeter point unit vector is computed using two-dimensional planar trigonometryAn angle c is computed using the formula for the cosine of an angle between twoincluded vectors (the perimeter point unit vector and the vector from SC to thecentre of the Earth) Angle y is computed using the Law of Sines Angle e=180degrees shy c shy y The scalar value of the perimeter point vector is computed using eand the Law of Cosines The scalar value is then multiplied by the perimeter pointunit vector to produce the three-dimensional point of intersection of the vector withEarthrsquos surface in R-Spacecraft The process is repeated independently for each ofthe eight perimeter point vectors Aside from its mathematical simplicity the valueof arcsine (computed in step 2) will exceed the normal range for the sin of an anglewhen the perimeter point vectors fail to intersect with the surface of the earth Asimple test based on this principle allows the program to correctly handle obliquephotos which image a portion of the horizon (see results for high oblique photographsin table 5)
The nal step in the process converts the eight earth intersection points from theR-Spacecraft to R-Earth The results are then converted to the standard geographiccoordinate system and displayed on the lsquoInput-Outputrsquo page of the spreadsheet
54 ExamplesAll three formulations were applied to the photographs included in this paper
and the results are compared in table 5 Formulation 1 gives too small a value fordistance across the photograph (D) for all but the most nadir shots and thus servesas an indicator of the best theoretical case but is not a good measure for a speci cphotograph For example the photograph of Lake Eyre taken from 276 km altitude( gure 2(a)) and the photograph of Limmen Bight ( gure 7) were closest to beingnadir views (oVsetslt68 km or tlt15deg table 5) For these photographs D calculatedusing Formulation 1 was similar to the minimum D calculated using Formulations2 and 3 For almost all other more oblique photographs Formulation 1 gave asigni cant underestimate of the distance covered in the photograph For gure 5(A)(the picture of Houston taken with a 40 mm lens) Formulation 1 did not give anunderestimate for D This is because Formulation 1 does not account for curvatureof the Earth in any way With this large eld of view assuming a at Earth in atedthe value of D above the minimum from calculations that included Earth curvature
A major diVerence between Formulations 2 and 3 is the ability to estimate pixelsizes (P) in both directions (along the principal line and perpendicular to the principalline) For the more oblique photographs the vertical estimate of D and pixel sizes ismuch larger than in the horizontal direction (eg the low oblique and high obliquephotographs of Hawaii gure 4 table 5)
For the area of Limmen Bight ( gure 7) table 5 illustrates the improvement inthe estimate of distance and pixel size that can be obtained by re-estimating thelocation of the PC with greater accuracy Centre points in the catalogued data are
Astronaut photography as digital data for remote sensing 4429
plusmn05deg of latitude and longitude When the centre point was re-estimated to plusmn002degwe determined that the photograph was not taken as obliquely as rst thought(change in the estimate of the look angle t from 169 to 128deg table 5) When theauxiliary point was added to the calculations of Formula 3 the calculated look angleshrank further to 110deg indicating that this photograph was taken at very close toa nadir view Of course this improvement in accuracy could have also led to estimatesof greater obliquity and corresponding larger pixel sizes
We also tested the performance of the scale calculator with auxiliary point byestimating the corner and point locations on the photograph using a 11 000 000Operational Navigational Chart For this test we estimated our ability to readcoordinates from the map as plusmn002degand our error in nding the locations of thecorner points as plusmn015deg (this error varies among photographs depending on thedetail that can be matched between photo and map) For Limmen Bight ( gure 7)the mean diVerence between map estimates and calculator estimates for four pointswas 031deg (SD=018 n=8) For a photograph of San Francisco Bay (STS062-151-291) the mean diVerence between map estimates and calculator estimates forfour points was 0064deg (SD=018 n=8) For a photograph of San Francisco Bay(STS062-151-291) the mean diVerence between map estimates and calculator estim-ates for eight points was 0196deg (SD=0146 n=16) Thus in one case the calculatorestimates were better than our estimate of the error in locating corner points on themap It is reasonable to expect that for nadir-viewing photographs the calculatorused with an auxiliary point can estimate locations of the edges of a photograph towithin plusmn03deg
55 Empirical con rmation of spatial resolution estimatesAs stated previously a challenge to estimating system-AWAR for astronaut
photography of Earth is the lack of suitable targets Small features in an image cansometimes be used as a check on the size of objects that can be successfully resolvedgiving an approximate value for GRD Similarly the number of pixels that make upthose features in the digitized image can be used to make an independent calculationof pixel size We have successfully used features such as roads and airport runwaysto make estimates of spatial scale and resolution (eg Robinson et al 2000c) Whilerecognizing that the use of linear detail in an image is a poor approximation to abar target and that linear objects smaller than the resolving power can often bedetected (Charman 1965) few objects other than roads could be found to make anydirect estimates of GRD Thus roads and runways were used in the images ofHouston (where one can readily conduct ground veri cations and where a numberof higher-contrast concrete roadways were available) to obtain empirical estimatesof GRD and pixel size for comparison with table 5
In the all-digital ESC image of Houston ( gure 3(d )) we examined EllingtonField runway 4-22 (centre left of the image) which is 24384 mtimes457 m This runwayis approximately 6ndash7 pixels in width and 304ndash3094 pixels in length so pixelsrepresent an distance on the ground 7ndash8 m Using a lower contrast measure of astreet length between two intersections (2125 m=211 pixels) a pixel width of 101 mis estimated These results compare favourably with the minimum estimate of 81 mpixels using Formulation 1 (table 5) For an estimate of GRD that would be morecomparable to aerial photography of a line target the smallest street without treecover that could be clearly distinguished on the photograph was 792 m wide Thesmallest non-street object (a gap between stages of the Saturn rocket on display in
J A Robinson et al4430
Tab
le5
App
lica
tio
nof
met
hod
sfo
res
tim
ati
ng
spati
alre
solu
tio
no
fast
ron
aut
ph
oto
gra
ph
sin
clu
din
gF
orm
ula
tion
s12
and
3Im
pro
ved
cen
tre
po
int
esti
mat
esan
dauxilia
ryp
oin
tca
lcu
lati
ons
are
sho
wn
for
gure
7V
aria
ble
sas
use
din
the
text
fle
ns
foca
lle
ngt
hH
al
titu
de
PC
ph
oto
gra
ph
cen
tre
poin
tSN
sp
ace
craft
nadir
po
int
D
dis
tance
cover
edo
nth
egr
oun
d
Pd
ista
nce
on
the
grou
nd
repre
sen
ted
by
apix
el
OVse
td
ista
nce
bet
wee
nP
Cand
SNusi
ng
the
grea
tci
rcle
form
ula
t
look
angle
away
from
SN
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
2L
ake
Eyre
ST
S093
-717
-80
250
276
shy29
0137
0shy
28
413
71
607
11
767
413
7609
ndash64
212
013
760
9ndash64
7121
124
ST
S082
-754
-61
250
617
shy28
5137
0shy
28
113
50
135
726
1201
18
013
8ndash14
827
517
9138ndash15
3277
294
Fig
ure
3H
ou
sto
nS
TS
062
-97-
151
4029
829
5shy
95
530
5shy
95
4410
78
8112
20
5348ndash58
990
1205
351
ndash63
8951
10
36
ST
S094
-737
-28
100
294
295
shy
95
528
3shy
95
5162
31
1133
24
415
8ndash20
334
724
3158ndash20
7351
390
ST
S055
-71-
41250
302
295
shy
95
028
2shy
93
766
412
8192
32
5736
ndash84
715
232
273
9ndash85
7154
185
S73
E51
45300
2685
295
shy
95
024
78
0F
igu
re4
Haw
aii
ST
S069
-701
-65
250
370
195
shy
155
520
4shy
153
881
415
7204
28
9876
ndash98
917
928
688
0ndash10
9181
210
ST
S069
-701
-70
250
370
210
shy
157
519
2shy
150
781
415
7738
63
414
9ndash23
336
760
7148ndash55
6394
10
63
ST
S069
-729
-27
100
398
200
shy
156
620
5shy
148
2219
42
1867
65
432
8ndash13
10
158
62
4323+
311
N
A
Astronaut photography as digital data for remote sensing 4431
Tab
le5
(Cont
inued
)
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
7L
imm
enB
igh
tS
TS
093
-704
-50
250
283
shy15
0135
0shy
15
013
58
623
120
859
16
963
0ndash67
3125
16
863
0ndash67
612
613
2P
CR
e-es
tim
ated
shy14
733
1352
67
64
5128
623
ndash65
5123
128
623
ndash65
712
3127
Auxilia
ryP
oin
t611
062
3ndash65
112
312
5F
igu
re10
N
ear
Au
stin
T
XS
TS
095
-716
-46
250
543
30
5shy
975
28
6shy
1007
120
230
375
34
613
5ndash157
281
34
1136ndash16
128
636
0
1 All
pho
togr
aphs
exce
pt
on
eta
ken
wit
hH
ass
elbla
dca
mer
aso
dis
tan
ceacr
oss
the
image
d=
55m
m
for
S73
E51
45
taken
wit
hel
ectr
onic
still
cam
erad=
276
5m
mho
rizo
nta
l2 P
Cand
SN
are
giv
enin
the
form
(lati
tud
elo
ngit
ude)
deg
rees
w
ith
neg
ativ
en
um
ber
sre
pre
senti
ng
south
and
wes
tA
ccura
cyo
fP
Cis
plusmn0
5and
SN
isplusmn
01
as
reco
rded
inth
eA
stro
naut
Ph
oto
gra
phy
Data
base
ex
cep
tw
her
en
ote
dth
atce
ntr
ep
oin
tw
asre
-est
imate
dw
ith
grea
ter
accu
racy
3 C
alc
ula
ted
usi
ng
the
mea
nofth
em
inim
um
an
dm
axim
um
esti
mate
sof
DF
orm
ula
tion
2co
nsi
der
sD
inth
ed
irec
tion
per
pen
dic
ula
rto
the
Pri
nci
pal
Lin
eo
nly
4 C
alc
ula
ted
inho
rizo
nta
lan
dve
rtic
aldir
ecti
on
sb
ut
still
ass
um
ing
geom
etry
of
gure
6(B
)F
irst
nu
mb
eris
the
mea
no
fth
em
inim
um
Dat
the
bott
om
of
the
ph
oto
and
the
maxi
mum
Dat
the
top
of
the
ph
oto
(range
show
nin
the
tab
le)
and
isco
mpara
ble
toF
orm
ula
tio
n2
Sec
ond
num
ber
isth
em
ean
of
Des
tim
ated
alon
gth
eP
rin
cip
alline
and
alo
ng
the
ou
tsid
eed
ge
(range
not
show
n)
5 Alt
itu
de
isap
pro
xim
ate
for
mis
sion
as
exac
tti
me
pho
togr
ap
hw
asta
ken
and
corr
esp
on
din
gn
adir
data
are
no
tava
ilab
le
Lack
of
nad
irdata
pre
ven
tsap
plica
tion
of
Form
ula
tion
s2
an
d3
6 Au
xilliar
ypo
int
add
edw
as
shy14
80
lati
tud
e13
51
2lo
ngit
ude
shy46
5degro
tati
on
angle
in
stru
ctio
ns
for
esti
mati
ng
the
rota
tio
nan
gle
para
met
erare
conta
ined
as
par
tof
the
scal
eca
lcu
lato
rsp
read
shee
tdo
cum
enta
tio
n
J A Robinson et al4432
a park at Johnson Space Center) that could clearly be distinguished on the photo-graph was 853 m wide
For the photograph of Houston taken with a 250 mm lens ( gure 3(C)) anddigitized from second generation lm at 2400 ppi (106 mm pixel Otilde 1 ) Ellington Fieldrunway 4-22 is 3 pixels in width and 1613 pixels in length so pixels represent151ndash152 m on the ground These results compare favourably with the minimumestimate of 152 m using Formulation 2 and 154ndash185 m pixels using Formulation 3(table 5) For an estimate of GRD using an 8times8 inch print (1369 enlargement) and4times magni cation the smallest street that could clearly be distinguished was 822 mwide the same feature could barely be distinguished on the digitized image
We also made an empirical estimate of spatial resolution for lower contrastvegetation boundaries By clearing forest so that a pattern would be visible to landingaircraft a landowner outside Austin Texas (see also aerial photo in Lisheron 2000)created a target that is also useful for evaluating spatial resolution of astronautphotographs The forest was selectively cleared in order to spell the landownerrsquosname lsquoLUECKErsquo with the remaining trees ( gure 10) According to local surveyorswho planned the clearing the plan was to create letters that were 3100 fttimes1700 ft(9449 mtimes5182 m) Photographed at a high altitude relative to most Shuttle missions(543 km) with a 250 mm lens Formula 3 predicts that each pixel would represent anarea 286 mtimes360 m on the ground (table 5) When original lm was digitized at2400 ppi (106 mm pixel Otilde 1 ) letters correspond to 294times188 pixels for a comparablepixel size of 27ndash32 m
6 Summary and conclusionsAstronaut photographs can be an excellent source of data for remote sensing
applications Best-case resolutions are similar to that for Landsat or SPOT withpixels as small as lt10 m It was estimated that of images taken to date 504 hadlens and obliquity characteristics that make them potential candidates for remotesensing information Digitized astronaut photographs can be overlaid with othersatellite data using GIS (Eckardt et al 2000) or used to ll in gaps in time serieswhen other imagery is not available As a source of public-domain information theycan be very useful for scientists who do not have access to satellite imagery eitherbecause of the expense of image acquisition (to the end user) or the computersystems needed for processing satellite images These diVerences in image acquisitioncosts (to the user) are summarized in table 6
Searching of the complete database of NASA astronaut photography includinglow-spatial-resolution browse images is available via the Web (OYce of EarthSciences 2000) This provides nearly global access for identifying images that willcontribute to a speci c scienti c project For the most detailed studies digitalproducts posted to the web will not be of suYcient quality To date we have madeit a practice of digitizing small numbers of images when requested by scientists atno charge Such requests (including a description of the project involved) can bemade through the authors or using the contact information listed on the websiteInvestigators with needs for larger numbers of digital images have also been servedthrough collaborative agreements
In our experience scientists that nd the data most valuable are those in develop-ing countries those studying areas that have not usually been targets for the majorsatellites those having diYculty in nding low-cloud images those interested inconstructing time series or those interested in using a large number of images
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
Astronaut photography as digital data for remote sensing 4411
(a) (b)
(c) (d)
Figure 3 Example of the eVect of lens focal length on resolving power Hasselblad imagesof Houston were all taken from an altitude of 294 kmndash302 km with diVerent lenses(a) 40 mm STS062-97-151 (b) 100 mm STS094-737-28 (c) 250 mm STS055-71-41 Forcomparison an Electronic Still Camera image with 300 mm lens at 269 km altitude isalso shown (d ) S73E5145
timestamp on each frame at the time of exposure Cameras are serviced between ights Occasionally there is enough volume and mass allowance so that a Linhof5times4 inch (127times101 mm) format camera can be own Nikon 35 mm cameras are own routinely also because of proven reliability Electronic still cameras (ESC)were tested for Earth photography beginning in 1992 (Lulla and Holland 1993) Inan ESC a CCD (charge-coupled device) is used as a digital replacement for lmrecording the image projected inside the camera An ESC (consisting of a KodakDCS 460c CCD in a Nikon N-90S body) was added as routine equipment forhandheld photographs in 1995 ESCs have also been operated remotely to captureand downlink Earth images through a NASA-sponsored educational programme(EarthKAM) Discussion of CCD spatial array and radiometric sensitivity arebeyond the scope of this paper but are summarized by Robinson et al (2000b) The
J A Robinson et al4412
format of the lm (or CCD) and image size projected onto the lm (or CCD) aresummarized for all the diVerent cameras own in table 2
34 L ook angle or obliquityNo handheld photographs can be considered perfectly nadir they are taken at a
variety of look angles ranging from near vertical ( looking down at approximatelythe nadir position of the spacecraft ) to high oblique (images that include the curvatureof Earth) Imaging at oblique look angles leads to an image where scale degradesaway from nadir A set of views of the same area from diVerent look angles is shownin gure 4 The rst two shots of the island of Hawaii were taken only a few secondsapart and with the same lens The third photograph was taken on a subsequentorbit and with a shorter lens The curvature of the Earth can be seen in the upperleft corner
Obliquity can be described qualitatively ( gure 4) or quantitatively as the lookangle (t gure 1 calculations described in formulation 2 (sect52)) Because obliquityand look angle have such a dramatic in uence on the footprint we summarize thedatabase characteristics relative to these two parameters Figure 5 is a breakdownof the spatial resolution characteristics of low oblique and near vertical photographsin the NASA Astronaut Photography database Number of photographs are grouped(1) by calculated values for look angle (t ) and (2) by altitude After observing theoverlap between near vertical and low oblique classes we are currently restructuringthis variable (lsquotiltrsquo) in the database to provide a measure of t when available Userswill still be able to do searches based on the qualitative measures but these measureswill be more closely tied to actual look angle
341 Obliquity and georeferencing digitised photographsOften the rst step in a remote sensing analysis of a digitized astronaut photo-
graph is to georeference the data and resample it to conform to a known mapprojection Details and recommendations for resampling astronaut photographydata are provided by Robinson et al (2000a 2000c) and a tutorial is also available(McRay et al 2000) Slightly oblique photographs can be geometrically correctedfor remote sensing purposes but extremely oblique photographs are not suited forgeometric correction When obliquity is too great the spatial scale far away fromnadir is much larger than the spatial scale closer to nadir resampling results areunsuitable because pixels near nadir are lost as the image is resampled while manypixels far away from nadir are excessively replicated by resampling
To avoid the generation of excess pixels during georeferencing the pixel sizes ofthe original digitized image should be smaller than the pixels in the nal resampledimage Calculations of original pixel size using methods presented below can beuseful in ensuring meaningful resampling For slightly oblique images formulation 3(sect53) can be used to estimate pixel sizes at various locations in a photograph (nearnadir and away from nadir) and these pixel sizes then used to determine a reasonablepixel scale following resampling
4 Other characteristics that in uence spatial resolutionBeyond basic geometry many other factors internal to the astronaut photography
system impact the observed ground resolved distance in a photograph The lensesand cameras already discussed are imperfect and introduce radiometric degradations(Moik 1980)Vignetting (slight darkening around the edges of an image) occurs in
Astronaut photography as digital data for remote sensing 4413
Tab
le2
Max
imu
mpo
ssib
led
igit
alsp
atia
lre
solu
tio
n(p
ixel
size
inm
)fo
rca
mer
asy
stem
scu
rren
tly
use
dby
ast
ron
auts
top
ho
togr
aph
eart
hb
ase
do
nth
ege
om
etry
of
alt
itud
ele
ns
and
lm
size
C
alc
ula
tion
sas
sum
eth
atco
lour
tran
spar
ency
lm
isdig
itiz
edat
2400
ppi
(pix
els
per
inch
10
6m
mpix
elOtilde
1 )
Min
imu
mgro
un
ddis
tance
repre
sente
dby
1p
ixel
(m)
As
afu
nct
ion
of
alti
tude1
Imag
esi
zeM
inim
um
alt
itud
eM
edia
nalt
itu
de
Maxi
mum
alt
itud
eC
amer
asy
stem
mm
timesm
mp
ixel
s2(2
22k
m)
(326
km
)(6
11
km
)
Lin
ho
f125
mm
90
mm
lens
105
times120
8978times
11
434
261
383
718
250
mm
lens
94
138
259
Has
selb
lad
70
mm
100
mm
lens
55times
55
5198
times519
823
534
564
725
0m
mle
ns
94
138
259
Nik
on
35m
m30
0m
mle
ns
36times
24
3308
times217
478
115
216
400
mm
lens
59
86
162
ESC
35m
m18
0m
mle
ns3
27
6times
18
530
69times
204
311
116
330
630
0m
mle
ns
67
98
184
400
mm
lens
50
74
138
1 Alt
itud
essh
ow
nar
em
inim
um
m
edia
nan
dm
axim
um
thro
ugh
ST
S-8
9(J
anuary
199
8N
=84
mis
sion
s)
2 Act
ual
pix
els
for
ES
C
oth
erw
ise
assu
med
dig
itiz
ati
on
at240
0pp
i(p
ixel
sp
erin
ch)
equ
alto
10
6m
mpix
elOtilde
1 3 T
he
mo
stco
mm
on
con
gura
tion
for
Eart
hph
oto
gra
ph
yas
part
of
the
Eart
hK
AM
pro
ject
J A Robinson et al4414
Figure 4 De nitions of look angle for photographs taken from low orbit around EarthThe example photographs of Hawaii were all taken from approximately the samealtitude (370ndash398 km) and with the same (250 mm) lens on a Hasselblad camera(STS069-701-69 STS069-701-70 and STS069-729-27 [100 mm lens])
astronaut photography due to light path properties of the lenses used A lack of atness of the lm at the moment of exposure (because cameras used in orbit donot incorporate a vacuum platen) also introduces slight degradations A numberof additional sources of image degradation that would aVect GRD of astronautphotographs are listed
41 Films and processing411 Colour reversal lms
Most lms used in NASA handheld cameras in orbit have been E-6 processcolour reversal lms (Ektachromes) that have a nominal speed of 64 to 100 ASAalthough many other lms have been own (table 3) Reversal lms are used becausedamage from radiation exposure while outside the atmospheric protection of Earthis more noticeable in negative lms than in positive lms (Slater 1996) The choiceof ASA has been to balance the coarser grains ( lower lm resolving power) of high-speed lms with the fact that slower lms require longer exposures and thus areaVected more by vehicle motion (approximately 73 km s Otilde 1 relative to Earthrsquos surfacefor the Space Shuttle) Extremely fast lms (gt400 ASA) are also more susceptibleto radiation damage in orbit (particularly during long-duration or high-altitudemissions Slater 1996) and have not traditionally been used for Earth photography The manufacturer stock numbers identifying the lm used are available for eachimage in the Astronaut Photography Database (OYce of Earth Sciences 2000)
412 Colour infrared lmsColour infrared lm (CIR) was used during an Earth-orbiting Apollo mission
(Colwell 1971) in the multispectral S-190A and the high-resolution S-190B camera
Astronaut photography as digital data for remote sensing 4415
Figure 5 (a) Distributions of astronaut photographs by look angle oV-nadir (calculated fromcentre point and nadir point using Formulation 2) Data included for all photographsfor which centre and nadir points are known with high oblique look angles (n=55 155) excluded (Total sample shown is 108 293 of 382 563 total records in thedatabase when compiled For records where nadir data was not available number ofobservations for qualitative look angles are near vertical 14 086 low oblique 115 834undetermined 89 195) (b) Altitude distribution for astronaut photographs of Earthtaken with the Hasselblad camera Data included for Hasselblad camera only focallength determined with high oblique look angles excluded (Total sample shown is159 046 of 206 971 Hasselblad records with known altitude and of 382 563 total recordsin the database when compiled For Hasselblad records with known altitude data notshown includes high oblique photographs using 100 mm and 250 mm lenses n=20 262and all look angles with other lenses n=27 663)
systems on Skylab (NASA 1974 Wilmarth et al 1977) and occasionally on Shuttlemissions and Shuttle-Mir missions (table 3) The CIR lm used is a three-layerAerochrome having one layer that is sensitive to re ected solar infrared energy toapproximately 900 nm (Eastman Kodak Company 1998) protective coatings onmost spacecraft windows also limit IR transmittance between 800 mm and 1000 nmThis layer is extremely sensitive to the temperature of the lm which createsunpredictable degradation of the IR signature under some space ight conditions
J A Robinson et al4416
Tab
le3
Det
ails
on
lm
suse
dfo
rast
ron
aut
pho
togr
aphy
of
Eart
h
Dat
ain
clud
ep
ho
togr
aphy
fro
mall
NA
SA
hum
an
spac
eig
ht
mis
sio
ns
thro
ugh
ST
S-9
6(J
un
e19
99
378
461
tota
lfr
ames
inth
edata
base
)F
ilm
sw
ith
lt50
00fr
am
esex
po
sed
and
lm
suse
dfo
r35
mm
cam
eras
only
are
no
tin
clud
ed
Res
olv
ing
Nu
mb
erof
Film
man
ufa
ctu
rer
AS
Ap
ow
er1
fram
esP
erio
dof
use
Cam
eras
Colo
ur
posi
tive
Kod
akA
eroch
rom
eII
(2447
2448
SO
-131S
O-3
68
)32
40ndash50
513
8196
8ndash19
85
Hass
elbla
dL
inh
of
Kod
akE
kta
chro
me
(5017
502
56
017
6118
SO
-121
64
50
113
932
196
5ndash19
68
Hass
elbla
dL
inh
of
Role
iex
SO
-217
Q
X-8
24
QX
-868
)19
81ndash1993
Mau
rer
Nik
on
Kod
akL
um
iere
(5046
5048
)100
105
743
199
4ndash19
98
Hass
elbla
dL
inho
fK
od
akE
lite
100
S(5
069
)100
41
486
199
6ndashp
rese
nt
Hass
elbla
d2
Fu
jiV
elvia
50C
S135
-36
32
13
882
1992ndash19
97H
ass
elbla
dN
iko
nK
od
akH
i-R
esolu
tio
nC
olo
r(S
O-2
42S
O-3
56
)10
096
74
1973ndash19
75H
ass
elbla
d
S19
0A
S190
B
Infr
are
dand
bla
ck-a
nd-w
hit
e
lms
Kod
akA
eroch
rom
eC
olo
rIn
frar
ed40
32
40
704
196
5ndashp
rese
nt2
Hass
elbla
dL
inh
of
Nik
on
S190
A
(8443
34
432443
)S
190B
Kod
akB
ampW
Infr
are
d(2
424
)32
10
553
197
3ndash19
74
S190
AK
od
akP
anato
mic
-XB
ampW
(SO
-022
)63
10
657
197
3ndash19
74
S190
A
1A
tlo
wco
ntr
ast
(ob
ject
back
grou
nd
rati
o16
1)
aspro
vid
edby
Ko
dak
and
list
edby
Sla
ter
etal
(1
983
256
)R
esolv
ing
pow
erw
as
dis
con
tin
ued
fro
mth
ete
chn
ical
test
sper
form
edb
yK
odak
inapp
roxim
atel
y19
93
an
dit
isn
ot
kn
ow
nif
curr
ent
lm
sh
ave
sim
ilar
reso
lvin
gpo
wer
too
lder
ver
sion
so
fsi
milar
lm
s(K
T
eite
lbaum
K
od
akp
erso
nal
com
mu
nic
atio
n)
Th
egra
nu
lari
tym
easu
reno
wpro
vid
edb
yK
od
ak
can
no
tb
ere
adily
con
ver
ted
toa
spat
ial
reso
luti
on
2 The
Lin
hof
was
use
dag
ain
on
ST
S-1
03in
Dec
emb
er19
99(d
ata
not
cata
logued
asof
the
pre
para
tion
of
this
tab
le)
usi
ng
Ko
dak
Elite
100S
lm
3 B
egin
nin
gin
July
1999
2443
colo
ur-
infr
are
d
lmpro
cess
was
chan
ged
from
EA
-5to
E-6
Astronaut photography as digital data for remote sensing 4417
413 Film duplicationThe original lm is archived permanently after producing about 20 duplicate
printing masters (second generation) Duplicate printing masters are disseminatedto regional archives and used to produce products (thirdndashfourth generation) for themedia the public and for scienti c use When prints slides transparencies anddigital products are produced for public distribution they are often colour adjustedto correspond to a more realistic look Digital products from second generationcopies can be requested by scienti c users (contact the authors for information onobtaining such products for a particular project) and these are recommended forremote sensing applications Care should be taken that the digital product acquiredfor remote sensing analysis has not been colour adjusted for presentation purposesBased on qualitative observations there is little visible increase in GRD in the thirdor fourth generation products There is a signi cant degradation in delity of colourin third and fourth generation duplicates because an increase in contrast occurswith every copy
42 Shutter speedsThe impact of camera motion (both due to the photographer and to the motion
of the spacecraft ) on image quality is determined by shutter speedmdash1250 to 1500second have been used because slower speeds record obvious blurring caused bythe rapid motion of the spacecraft relative to the surface of the Earth Generally1250 was used for ISO 64 lms (and slower) and after the switch to ISO 100 lmsa 1500 setting became standard New Hasselblad cameras that have been ownbeginning with STS-92 in October 2000 vary shutter speed using a focal plane shutterfor exposure bracketing (rather than varying aperture) and 1250 1500 and 11000second are used
Across an altitude range of 120ndash300 nautical miles (222ndash555 km) and orbitalinclinations of 285deg and 570deg the median relative ground velocity of orbitingspacecraft is 73 km s Otilde 1 At a nominal shutter speed of 1500 the expected blur dueto motion relative to the ground would be 146 m When cameras are handheld (asopposed to mounted in a bracket) blur can be reduced by physical compensationfor the motion (tracking) by the photographer many photographs show a level ofdetail indicating that blur at this level did not occur (eg gure 3(d )) Thus motionrelative to the ground is not an absolute barrier to ground resolution for handheldphotographs
43 Spacecraft windowsMost of the photographs in the NASA Astronaut Photography Database were
taken through window ports on the spacecraft used The transmittance of the windowport may be aVected by material inhomogeniety of the glass the number of layersof panes coatings used the quality of the surface polish environmentally inducedchanges in window materials (pressure loads thermal gradients) or deposited con-tamination Such degradation of the window cannot be corrected is diVerent foreach window and changes over time See Eppler et al (1996) and Scott (2000) fordiscussion of the spectral transmittance of the fused quartz window that is part ofthe US Laboratory Module (Destiny) of the International Space Station
44 Digitized images from lmWhen lm is scanned digitally the amount of information retained depends on
the spectral information extracted from the lm at the spatial limits of the scanner
J A Robinson et al4418
To date standard digitizing methodologies for astronaut photographs have not beenestablished and lm is digitized on a case-by-case basis using the equipment available
441 Digitizing and spatial resolutionLight (1993 1996) provided equations for determining the digitizing spatial
resolution needed to preserve spatial resolution of aerial photography based on thestatic resolving power (AWAR) for the system For lms where the manufacturer hasprovided data (Eastman Kodak 1998 K Teitelbaum personal communication) the resolving power of lms used for astronaut photography ranges from 32 lp mm Otilde 1to 100 lp mm Otilde 1 at low contrast (objectbackground ratio 161 table 3) The AWARfor the static case of the Hasselblad camera has been measured at high and lowcontrast (using lenses shown in table 1) with a maximum of approximately55 lp mm Otilde 1 (Fred Pearce unpublished data)
Based on the method of Light (1993 1996) the dimension of one spatial resolutionelement for a photograph with maximum static AWAR of 55 lp mm Otilde 1 would be18 mm lp Otilde 1 The acceptable range of spot size to preserve spatial information wouldthen be
18 mm
2 atilde 2aring scan spot size aring
18 mm
2(1)
and 6 mm aring scan spot size aring 9 mm Similarly for a more typical static AWAR of30 lp mm Otilde 1 (33 mm lpOtilde 1 low contrast Fred Pearce unpublished data) 11 mm aring scanspot sizearing 17 mm These scan spot sizes correspond to digitizing resolutions rangingfrom 4233ndash2822 ppi (pixels inch Otilde 1 ) for AWAR of 55 lp mm Otilde 1 and 2309ndash1494 ppi forAWAR of 30 lp mm Otilde 1 Scan spot sizes calculated for astronaut photography arecomparable to those calculated for the National Aerial Photography Program(9ndash13 mm Light 1996)
Widely available scanners that can digitize colour transparency lm currentlyhave a maximum spatial resolution of approximately 2400 ppi (106 mm pixel Otilde 1) Forexample we routinely use an Agfa Arcus II desktop scanner (see also Baltsavias1996) with 2400 ppi digitizing spatial resolution (2400 ppi optical resolution in onedirection and 1200 ppi interpolated to 2400 ppi in the other direction) and 2400 ppiwas used to calculate IFOV equivalents (tables 2 and 4) For some combinations oflens camera and contrast 2400 ppi will capture nearly all of the spatial informationcontained in the lm However for lm with higher resolving power thanEktachrome-64 for better lenses and for higher contrast targets digitizing at 2400 ppiwill not capture all of the spatial information in the lm
Improvements in digitizing technology will only produce real increases in IFOVto the limit of the AWAR of the photography system The incremental increase inspatial information above 3000 ppi (85 mm pixel Otilde 1 ) is not likely to outweigh thecosts of storing large images (Luman et al 1997) At 2400 ppi a Hasselblad frameis approximately 5200times5200 or 27 million pixels (table 2) while the same imagedigitized at 4000 ppi would contain 75 million pixels
Initial studies using astronaut photographs digitized at 2400 ppi (106 mm pixel Otilde 1 Webb et al 2000 Robinson et al 2000c) indicate that some GRD is lost comparedto photographic products Nevertheless the spatial resolution is still comparablewith other widely used data sources (Webb et al 2000) Robinson et al (2000c)found that digitizing at 21 mm pixel Otilde 1 provided information equivalent to 106 mmpixel Otilde 1 for identifying general urban area boundaries for six cities except for a single
Astronaut photography as digital data for remote sensing 4419
photo that required higher spatial resolution digitizing (it had been taken with ashorter lens and thus had less spatial resolution) Part of the equivalence observedin the urban areas study may be attributable to the fact that the atbed scannerused interpolates from 1200 ppi to 2400 ppi in one direction The appropriate digitiz-ing spatial resolution will in part depend on the scale of features of interest to theresearchers with a maximum upper limit set of a scan spot size of approximately6 mm (4233 ppi)
442 Digitizing and spectral resolutionWhen colour lm is digitized there will be loss of spectral resolution The three
lm emulsion layers (red green blue) have relatively distinct spectral responses butare fused together so that digitizers detect one colour signal and must convert itback into three (red green blue) digital channels Digital scanning is also subject tospectral calibration and reproduction errors Studies using digital astronaut photo-graphs to date have used 8 bits per channel However this is another parameterthat can be controlled when the lm is digitized We do not further address spectralresolution of digitally scanned images in this paper
45 External factors that in uence GRDAlthough not discussed in detail in this paper factors external to the spacecraft
and camera system (as listed by Moik (1980) for remote sensing in general) alsoimpact GRD These include atmospheric interference (due to scattering attenuationhaze) variable surface illumination (diVerences in terrain slope and orientation) andchange of terrain re ectance with viewing angle (bidirectional re ectance) For astro-naut photographs variable illumination is particularly important because orbits arenot sun-synchronous Photographs are illuminated by diVerent sun angles and imagesof a given location will have colour intensities that vary widely In addition theviewing angle has an eVect on the degree of object-to-backgroun d contrast andatmospheric interference
5 Estimating spatial resolution of astronaut photographsIn order to use astronaut photographs for digital remote sensing it is important
to be able to calculate the equivalent to an IFOVmdashthe ground area represented bya single pixel in a digitized orbital photograph The obliquity of most photographsmeans that pixel lsquosizesrsquo vary at diVerent places in an image Given a ground distanceD represented by a photograph in each direction (horizontal vertical ) an approxi-mate average pixel width (P the equivalent of IFOV) for the entire image can becalculated as follows
P=D
dS(2)
where D is the projected distance on the ground covered by the image in the samedirection as the pixel is measured d is the actual width of the image on the original lm (table 1) and S is the digitizing spatial resolution
Here we present three mathematical formulations for estimating the size of thefootprint or area on the ground covered by the image Example results from theapplication of all three formulations are given in table 5 The rst and simplestcalculation (formulation 1) gives an idea of the maximum spatial resolution attainableat a given altitude of orbit with a given lm format and a perfectly vertical (nadir)
J A Robinson et al4420
view downward Formulation 2 takes into account obliquity by calculating lookangle from the diVerence between the location on the ground represented at thecentre of the photograph and the nadir location of the spacecraft at the time thephotograph was taken ( gure 1) Formulation 3 describes an alternate solution tothe oblique look angle problem using coordinate-system transformations Thisformulation has been implemented in a documented spreadsheet and is availablefor download (httpeoljscnasagovsseopFootprintCalculatorxls) and is in theprocess of being implemented on a Web-based user interface to the AstronautPhotography Database (OYce of Earth Sciences 2000)
Although formulations 2 and 3 account for obliquity for the purposes of calcula-tion they treat the position of the spacecraft and position of the camera as one Inactuality astronauts are generally holding the cameras by hand (although camerasbracketed in the window are also used) and the selection of window position of theastronaut in the window and rotation of the camera relative to the movement ofthe spacecraft are not known Thus calculations using only the photo centre point(PC) and spacecraft nadir point (SN) give a locator ellipse and not the locations ofthe corners of the photograph A locator ellipse describes an estimated area on theground that is likely to be included in a speci c photograph regardless of the rotationof the lm plane about the camerarsquos optical axis ( gure 6)
Estimating the corner positions of the photo requires additional user input of asingle auxiliary pointmdasha location on the image that has a known location on theground Addition of this auxiliary point is an option available to users of thespreadsheet An example of the results of adding an auxiliary point is shown in gure 7 with comparisons of the various calculations in table 5
51 Formulation 1 Footprint for a nadir viewThe simplest way to estimate footprint size is to use the geometry of camera lens
and spacecraft altitude to calculate the scaling relationship between the image in the
Figure 6 Sketch representation of the locator ellipse an estimated area on the ground thatis likely to be included in a speci c photograph regardless of the rotation of the lmplane about the camerarsquos optical axis
Astronaut photography as digital data for remote sensing 4421
Figure 7 An example of the application of the Low Oblique Space Photo FootprintCalculator The photograph (STS093-704-50) of Limmen Bight Australia was takenfrom an altitude of 283 km using the Hasselblad camera with a 250 mm lens Mapused is 11 000 000 Operational Navigational Chart The distances shown equate toan average pixel size of 124 m for this photograph
lm and the area covered on the ground For a perfect nadir view the scalerelationship from the geometry of similar triangles is
d
D=
f
H(3)
where d=original image size D=distance of footprint on the ground f =focallength of lens and H=altitude Once D is known pixel size ( length=width) can becalculated from equation (2) These calculations represent the minimum footprintand minimum pixel size possible for a given camera system altitude and digitisingspatial resolution (table 2) Formulation 1 was used for calculating minimum pixelsizes shown in tables 2 and 4
By assuming digitizing at 2400 ppi (106 mm pixel Otilde 1 ) currently a spatial resolutioncommonly attainable from multipurpose colour transparency scanners (see sect441)this formulation was used to convert area covered to an IFOV equivalent for missionsof diVerent altitudes (table 2) Table 4 provides a comparison of IFOV of imagesfrom various satellites including the equivalent for astronaut photography Valuesin this table were derived by using formulation 1 to estimate the area covered becausea perfect nadir view represents the best possible spatial resolution and smallest eldof view that could be obtained
52 Formulation 2 Footprint for oblique views using simpli ed geometry and thegreat circle distance
A more realistic approach to determining the footprint of the photographaccounts for the fact that the camera is not usually pointing perfectly down at the
J A Robinson et al4422
Table 4 Comparison of pixel sizes (instantaneous elds of view) for satellite remote sensorswith pixel sizes for near vertical viewing photography from spacecraft Note that alldata returned from the automated satellites have the same eld of view while indicatedpixel sizes for astronaut photography are best-case for near vertical look angles See gure 5 for distributions of astronaut photographs of various look angles High-altitude aerial photography is also included for reference
Altitude PixelInstrument (km) width (mtimesm) Bands1
Landsat Multipectral Scanner2 (MSS 1974-) 880ndash940 79times56 4ndash5Landsat Thematic Mapper (TM 1982-) ~705 30times30 7Satellite pour lrsquoObservation de la Terre (SPOT 1986-) ~832
SPOT 1-3 Panchromatic 10times10 1SPOT 1-3 Multispectral (XS) 20times20 3
Astronaut Photography (1961-)Skylab3 (1973-1974 ) ~435
Multispectral Camera (6 camera stations) 17ndash19times17ndash19 3ndash8Earth Terrain Camera (hi-res colour) 875times875 3
Space Shuttle5 (1981-) 222ndash611Hasselblad 70 mm (250mm lens colour) vertical view 9ndash26times9ndash26 3Hasselblad 70 mm (100mm lens colour) vertical view 23ndash65times23ndash65 3Nikon 35 mm (400mm lens colour lm or ESC) vertical 5ndash16times5ndash16 3
High Altitude Aerial Photograph (1100 000) ~12 2times2 3
1Three bands listed for aerial photography obtainable by high-resolution colour digitizingFor high-resolution colour lm at 65 lp mmOtilde 1 (K Teitelbaum Eastman Kodak Co personalcommunication) the maximum information contained would be 3302 ppi
2Data from Simonett (1983Table 1-2)3Calculated as recommended by Jensen (19831596) the dividend of estimated ground
resolution at low contrast given in NASA (1974179-180) and 244Six lters plus colour and colour infrared lm (NASA 19747) 5Extracted from table 2
nadir point The look angle (the angle oV nadir that the camera is pointing) can becalculated trigonometrically by assuming a spherical earth and calculating the dis-tance between the coordinates of SN and PC ( gure 1) using the Great Circledistance haversine solution (Sinnott 1984 Snyder 198730ndash32 Chamberlain 1996)The diVerence between the spacecraft centre and nadir point latitudes Dlat=lat 2 shy lat 1 and the diVerence between the spacecraft centre and nadir pointlongitudes Dlon=lon 2 shy lon 1 enter the following equations
a=sin2ADlat
2 B+cos(lat 1) cos(lat 2) sin2ADlon
2 B (4)
c=2 arcsin(min[1 atilde a]) (5)
and oVset=R c with R the radius of a spherical Earth=6370 kmThe look angle (t ) is then given by
t=arctanAoffset
H B (6)
Assuming that the camera was positioned so that the imaginary line between thecentre and nadir points (the principal line) runs vertically through the centre of thephotograph the distance between the geometric centre of the photograph (principalpoint PP) and the top of the photograph is d2 ( gure 8(a)) The scale at any point
Astronaut photography as digital data for remote sensing 4423
(a) (b)
Figure 8 The geometric relationship between look angle lens focal length and the centrepoint and nadir points of a photograph (see also Estes and Simonett 1975) Note thatthe Earthrsquos surface and footprint represented in gure 1 are not shown here only arepresentation of the image area of the photograph Variables shown in the gurelook angle t focal length f PP centre of the image on the lm SN spacecraft nadird original image size (a) The special case where the camera is aligned squarely withthe isometric parallel In this case tilt occurs along a plane parallel with the centre ofthe photograph When the conditions are met calculations using Formulation 3 willgive the ground coordinates of the corner points of the photograph (b) The generalrelationship between these parameters and key photogrammetric points on the photo-graph For this more typical case coordinates of an additional point in the photographmust be input in order to adjust for twist of the camera and to nd the corner pointcoordinates
in the photograph varies as a function of the distance y along the principal linebetween the isocentre and the point according to the relationship
d
D=
f shy ysint
H(7)
(Wong 1980 equation (214) H amp the ground elevation h) Using gure 8(a) at thetop of the photo
y= f tanAt
2B+d
2(8)
and at the bottom of the photo
y= f tanAt
2Bshyd
2(9)
Thus for given PC and SN coordinates and assuming a photo orientation as in gure 8(a) and not gure 8(b) we can estimate a minimum D (using equations (7)and (8)) and a maximum D (using equations (7) and (9)) and then average the twoto determine the pixel size (P ) via equation (2)
J A Robinson et al4424
53 Formulation 3 T he L ow Oblique Space Photo Footprint CalculatorThe Low Oblique Space Photo Footprint Calculator was developed to provide
a more accurate estimation of the geographic coordinates for the footprint of a lowoblique photo of the Earthrsquos surface taken from a human-occupied spacecraft inorbit The calculator performs a series of three-dimensional coordinate transforma-tions to compute the location and orientation of the centre of the photo exposureplane relative to an earth referenced coordinate system The nominal camera focallength is then used to create a vector from the photorsquos perspective point througheach of eight points around the parameter of the image as de ned by the formatsize The geographical coordinates for the photo footprint are then computed byintersecting these photo vectors with a spherical earth model Although more sophist-icated projection algorithms are available no signi cant increase in the accuracy ofthe results would be produced by these algorithms due to inherent uncertainties inthe available input data (ie the spacecraft altitude photo centre location etc)
The calculations were initially implemented within a Microsoft Excel workbookwhich allowed one to embed the mathematical and graphical documentation nextto the actual calculations Thus interested users can inspect the mathematical pro-cessing A set of error traps was also built into the calculations to detect erroneousresults A summary of any errors generated is reported to the user with the resultsof the calculations Interested individuals are invited to download the Excel work-book from httpeoljscnasagovsseopFootprintCalculatorxls The calculations arecurrently being encoded in a high-level programming language and should soon beavailable alongside other background data provided for each photograph at OYceof Earth Sciences (2000)
For the purposes of these calculations a low oblique photograph was de ned as
one with the centre within 10 degrees of latitude and longitude of the spacecraftnadir point For the typical range of spacecraft altitudes to date this restricted thecalculations to photographs in which Earthrsquos horizon does not appear (the generalde nition of a low oblique photograph eg Campbell 199671 and gure 4)
531 Input data and resultsUpon opening the Excel workbook the user is presented with a program intro-
duction providing instructions for using the calculator The second worksheet tablsquoHow-To-Usersquo provides detailed step-by-step instructions for preparing the baselinedata for the program The third tab lsquoInput-Output rsquo contains the user input eldsand displays the results of the calculations The additional worksheets contain theactual calculations and program documentation Although users are welcome toreview these sheets an experienced user need only access the lsquoInput-Outputrsquo
spreadsheetTo begin a calculation the user enters the following information which is available
for each photo in the NASA Astronaut Photography Database (1) SN geographicalcoordinates of spacecraft nadir position at the time of photo (2) H spacecraftaltitude (3) PC the geographical coordinates of the centre of the photo (4) f nominal focal length and (5) d image format size The automatic implementationof the workbook on the web will automatically enter these values and completecalculations
For more accurate results the user may optionally enter the geographic co-ordinates and orientation for an auxiliary point on the photo which resolves thecamerarsquos rotation uncertainty about the optical axis The auxiliary point data must
Astronaut photography as digital data for remote sensing 4425
be computed by the user following the instruction contained in the lsquoHow-To-Usersquotab of the spreadsheet
After entering the input data the geographic coordinates of the photo footprint(ie four photo corner points four points at the bisector of each edge and the centreof the photo) are immediately displayed below the input elds along with any errormessages generated by the user input or by the calculations ( gure 7) Although
results are computed and displayed they should not be used when error messagesare produced by the program The program also computes the tilt angle for each ofthe photo vectors relative to the spacecraft nadir vector To the right of the photofootprint coordinates is displayed the arc distance along the surface of the sphere
between adjacent computed points
532 Calculation assumptions
The mathematical calculations implemented in the Low Oblique Space PhotoFootprint Calculator use the following assumptions
1 The SN location is used as exact even though the true value may vary by upto plusmn01 degree from the location provided with the photo
2 The spacecraft altitude is used as exact Although the determination of the
nadir point at the instant of a known spacecraft vector is relatively precise(plusmn115times10 Otilde 4 degrees) the propagator interpolates between sets of approxi-mately 10ndash40 known vectors per day and the time code recorded on the lmcan drift Thus the true value for SN may vary by up to plusmn01 degree fromthe value provided with the photo
3 The perspective centre of the camera is assumed to be at the given altitudeover the speci ed spacecraft nadir location at the time of photo exposure
4 The PC location is used as exact even though the true value may vary by upto plusmn05deg latitude and plusmn05deg longitude from the location provided with the
photo5 A spherical earth model is used with a nominal radius of 6 372 16154 m (a
common rst-order approximation for a spherical earth used in geodeticcomputations)
6 The nominal lens focal length of the camera lens is used in the computations(calibrated focal length values are not available)
7 The photo projection is based on the classic pin-hole camera model8 No correction for lens distortion or atmospheric re ection is made
9 If no auxiliary point data is provided the lsquoTop of the Imagersquo is orientedperpendicular to the vector from SN towards PC
533 T ransformation from Earth to photo coordinate systemsThe calculations begin by converting the geographic coordinates ( latitude and
longitude) of the SN and PC to a Rectangular Earth-Centred Coordinate System(R-Earth) de ned as shown in gure 9 (with the centre of the Earth at (0 0 0))
Using the vector from the Earthrsquos centre through SN and the spacecraft altitude thespacecraft location (SC) is also computed in R-Earth
For ease of computation a Rectangular Spacecraft-Centred Coordinate System(R-Spacecraft ) is de ned as shown in gure 9 The origin of R-Spacecraft is located
at SC with its +Z-axis aligned with vector from the centre of the Earth throughSN and its +X-axis aligned with the vector from SN to PC ( gure 9) The speci c
J A Robinson et al4426
Figure 9 Representation of the area included in a photograph as a geometric projectiononto the Earthrsquos surface Note that the focal length (the distance from the SpaceShuttle to the photo) is grossly overscale compared to the altitude (the distance fromthe Shuttle to the surface of the Earth
rotations and translation used to convert from the R-Earth to the R-Spacecraft arecomputed and documented in the spreadsheet
With the mathematical positions of the SC SN PC and the centre of the Earthcomputed in R-Spacecraft the program next computes the location of the camerarsquosprincipal point (PP) The principle point is the point of intersection of the opticalaxis of the lens with the image plane (ie the lm) It is nominally positioned at a
Astronaut photography as digital data for remote sensing 4427
distance equal to the focal length from the perspective centre of the camera (whichis assumed to be at SC) along the vector from PC through SC as shown in gure 9
A third coordinate system the Rectangular Photo Coordinate System (R-Photo)is created with its origin at PP its XndashY axial plane normal to the vector from PCthrough SC and its +X axis aligned with the +X axis of R-Spacecraft as shownin gure 9 The XndashY plane of this coordinate system represents the image plane ofthe photograph
534 Auxiliary point calculationsThe calculations above employ a critical assumption that all photos are taken
with the lsquotop of the imagersquo oriented perpendicular to the vector from the SN towardsPC as shown in gure 9 To avoid non-uniform solar heating of the external surfacemost orbiting spacecraft are slowly and continually rotated about one or more axesIn this condition a ight crew member taking a photo while oating in microgravitycould orient the photo with practically any orientation relative to the horizon (seealso gure 8(b)) Unfortunately since these photos are taken with conventionalhandheld cameras there is no other information available which can be used toresolve the photorsquos rotational ambiguity about the optical axis other then the photoitself This is why the above assumption is used and the footprint computed by thiscalculator is actually a lsquolocator ellipsersquo which estimates the area on the ground whatis likely to be included in a speci c photograph (see gure 6) This locator ellipse ismost accurate for square image formats and subject to additional distortion as thephotograph format becomes more rectangular
If the user wants a more precise calculation of the image footprint the photorsquosrotational ambiguity about the optical axis must be resolved This can be done inthe calculator by adding data for an auxiliary point Detailed instructions regardinghow to prepare and use auxiliary point data in the computations are included in thelsquoHow-To-Usersquo tab of the spreadsheet Basically the user determines which side ofthe photograph is top and then measures the angle between the line from PP to thetop of the photo and from PP to the auxiliary point on the photo ( gure 9)
If the user includes data for an auxiliary point a series of computations arecompleted to resolve the photo rotation ambiguity about the optical axis (ie the
+Z axis in R-Photo) A vector from the Auxiliary Point on the Earth (AE) throughthe photograph perspective centre ( located at SC) is intersected with the photo imageplane (XndashY plane of R-Photo) to compute the coordinates of the Auxiliary Point onthe photo (AP) in R-Photo A two-dimensional angle in the XndashY plane of R-Photofrom the shy X axis to a line from PP to AP is calculated as shown in gure 9 The
shy X axis is used as the origin of the angle since it represents the top of the photoonce it passes through the perspective centre The diVerence between the computedangle and the angle measured by the user on the photo resolves the ambiguity inthe rotation of the photo relative to the principal line ( gures 7 and 9) Thetransformations from R-Spacecraft and R-Photo are then modi ed to include anadditional rotation angle about the +Z axis in R-Photo
535 lsquoFootprint rsquo calculationsThe program next computes the coordinates of eight points about the perimeter
of the image format (ie located at the four photo corners plus a bisector pointalong each edge of the image) These points are identi ed in R-Photo based uponthe photograph format size and then converted to R-Spacecraft Since all computa-tions are done in orthogonal coordinate systems the R-Spacecraft to R-Photo
J A Robinson et al4428
rotation matrix is transposed to produce an R-Photo to R-Spacecraft rotation matrixOnce in R-Spacecraft a unit vector from each of the eight perimeter points throughthe photo perspective centre (the same point as SC) is computed This provides thecoordinates for points about the perimeter of the image format with their directionvectors in a common coordinate system with other key points needed to computethe photo footprint
The next step is to compute the point of intersection between the spherical earthmodel and each of the eight perimeter point vectors The scalar value for eachperimeter point unit vector is computed using two-dimensional planar trigonometryAn angle c is computed using the formula for the cosine of an angle between twoincluded vectors (the perimeter point unit vector and the vector from SC to thecentre of the Earth) Angle y is computed using the Law of Sines Angle e=180degrees shy c shy y The scalar value of the perimeter point vector is computed using eand the Law of Cosines The scalar value is then multiplied by the perimeter pointunit vector to produce the three-dimensional point of intersection of the vector withEarthrsquos surface in R-Spacecraft The process is repeated independently for each ofthe eight perimeter point vectors Aside from its mathematical simplicity the valueof arcsine (computed in step 2) will exceed the normal range for the sin of an anglewhen the perimeter point vectors fail to intersect with the surface of the earth Asimple test based on this principle allows the program to correctly handle obliquephotos which image a portion of the horizon (see results for high oblique photographsin table 5)
The nal step in the process converts the eight earth intersection points from theR-Spacecraft to R-Earth The results are then converted to the standard geographiccoordinate system and displayed on the lsquoInput-Outputrsquo page of the spreadsheet
54 ExamplesAll three formulations were applied to the photographs included in this paper
and the results are compared in table 5 Formulation 1 gives too small a value fordistance across the photograph (D) for all but the most nadir shots and thus servesas an indicator of the best theoretical case but is not a good measure for a speci cphotograph For example the photograph of Lake Eyre taken from 276 km altitude( gure 2(a)) and the photograph of Limmen Bight ( gure 7) were closest to beingnadir views (oVsetslt68 km or tlt15deg table 5) For these photographs D calculatedusing Formulation 1 was similar to the minimum D calculated using Formulations2 and 3 For almost all other more oblique photographs Formulation 1 gave asigni cant underestimate of the distance covered in the photograph For gure 5(A)(the picture of Houston taken with a 40 mm lens) Formulation 1 did not give anunderestimate for D This is because Formulation 1 does not account for curvatureof the Earth in any way With this large eld of view assuming a at Earth in atedthe value of D above the minimum from calculations that included Earth curvature
A major diVerence between Formulations 2 and 3 is the ability to estimate pixelsizes (P) in both directions (along the principal line and perpendicular to the principalline) For the more oblique photographs the vertical estimate of D and pixel sizes ismuch larger than in the horizontal direction (eg the low oblique and high obliquephotographs of Hawaii gure 4 table 5)
For the area of Limmen Bight ( gure 7) table 5 illustrates the improvement inthe estimate of distance and pixel size that can be obtained by re-estimating thelocation of the PC with greater accuracy Centre points in the catalogued data are
Astronaut photography as digital data for remote sensing 4429
plusmn05deg of latitude and longitude When the centre point was re-estimated to plusmn002degwe determined that the photograph was not taken as obliquely as rst thought(change in the estimate of the look angle t from 169 to 128deg table 5) When theauxiliary point was added to the calculations of Formula 3 the calculated look angleshrank further to 110deg indicating that this photograph was taken at very close toa nadir view Of course this improvement in accuracy could have also led to estimatesof greater obliquity and corresponding larger pixel sizes
We also tested the performance of the scale calculator with auxiliary point byestimating the corner and point locations on the photograph using a 11 000 000Operational Navigational Chart For this test we estimated our ability to readcoordinates from the map as plusmn002degand our error in nding the locations of thecorner points as plusmn015deg (this error varies among photographs depending on thedetail that can be matched between photo and map) For Limmen Bight ( gure 7)the mean diVerence between map estimates and calculator estimates for four pointswas 031deg (SD=018 n=8) For a photograph of San Francisco Bay (STS062-151-291) the mean diVerence between map estimates and calculator estimates forfour points was 0064deg (SD=018 n=8) For a photograph of San Francisco Bay(STS062-151-291) the mean diVerence between map estimates and calculator estim-ates for eight points was 0196deg (SD=0146 n=16) Thus in one case the calculatorestimates were better than our estimate of the error in locating corner points on themap It is reasonable to expect that for nadir-viewing photographs the calculatorused with an auxiliary point can estimate locations of the edges of a photograph towithin plusmn03deg
55 Empirical con rmation of spatial resolution estimatesAs stated previously a challenge to estimating system-AWAR for astronaut
photography of Earth is the lack of suitable targets Small features in an image cansometimes be used as a check on the size of objects that can be successfully resolvedgiving an approximate value for GRD Similarly the number of pixels that make upthose features in the digitized image can be used to make an independent calculationof pixel size We have successfully used features such as roads and airport runwaysto make estimates of spatial scale and resolution (eg Robinson et al 2000c) Whilerecognizing that the use of linear detail in an image is a poor approximation to abar target and that linear objects smaller than the resolving power can often bedetected (Charman 1965) few objects other than roads could be found to make anydirect estimates of GRD Thus roads and runways were used in the images ofHouston (where one can readily conduct ground veri cations and where a numberof higher-contrast concrete roadways were available) to obtain empirical estimatesof GRD and pixel size for comparison with table 5
In the all-digital ESC image of Houston ( gure 3(d )) we examined EllingtonField runway 4-22 (centre left of the image) which is 24384 mtimes457 m This runwayis approximately 6ndash7 pixels in width and 304ndash3094 pixels in length so pixelsrepresent an distance on the ground 7ndash8 m Using a lower contrast measure of astreet length between two intersections (2125 m=211 pixels) a pixel width of 101 mis estimated These results compare favourably with the minimum estimate of 81 mpixels using Formulation 1 (table 5) For an estimate of GRD that would be morecomparable to aerial photography of a line target the smallest street without treecover that could be clearly distinguished on the photograph was 792 m wide Thesmallest non-street object (a gap between stages of the Saturn rocket on display in
J A Robinson et al4430
Tab
le5
App
lica
tio
nof
met
hod
sfo
res
tim
ati
ng
spati
alre
solu
tio
no
fast
ron
aut
ph
oto
gra
ph
sin
clu
din
gF
orm
ula
tion
s12
and
3Im
pro
ved
cen
tre
po
int
esti
mat
esan
dauxilia
ryp
oin
tca
lcu
lati
ons
are
sho
wn
for
gure
7V
aria
ble
sas
use
din
the
text
fle
ns
foca
lle
ngt
hH
al
titu
de
PC
ph
oto
gra
ph
cen
tre
poin
tSN
sp
ace
craft
nadir
po
int
D
dis
tance
cover
edo
nth
egr
oun
d
Pd
ista
nce
on
the
grou
nd
repre
sen
ted
by
apix
el
OVse
td
ista
nce
bet
wee
nP
Cand
SNusi
ng
the
grea
tci
rcle
form
ula
t
look
angle
away
from
SN
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
2L
ake
Eyre
ST
S093
-717
-80
250
276
shy29
0137
0shy
28
413
71
607
11
767
413
7609
ndash64
212
013
760
9ndash64
7121
124
ST
S082
-754
-61
250
617
shy28
5137
0shy
28
113
50
135
726
1201
18
013
8ndash14
827
517
9138ndash15
3277
294
Fig
ure
3H
ou
sto
nS
TS
062
-97-
151
4029
829
5shy
95
530
5shy
95
4410
78
8112
20
5348ndash58
990
1205
351
ndash63
8951
10
36
ST
S094
-737
-28
100
294
295
shy
95
528
3shy
95
5162
31
1133
24
415
8ndash20
334
724
3158ndash20
7351
390
ST
S055
-71-
41250
302
295
shy
95
028
2shy
93
766
412
8192
32
5736
ndash84
715
232
273
9ndash85
7154
185
S73
E51
45300
2685
295
shy
95
024
78
0F
igu
re4
Haw
aii
ST
S069
-701
-65
250
370
195
shy
155
520
4shy
153
881
415
7204
28
9876
ndash98
917
928
688
0ndash10
9181
210
ST
S069
-701
-70
250
370
210
shy
157
519
2shy
150
781
415
7738
63
414
9ndash23
336
760
7148ndash55
6394
10
63
ST
S069
-729
-27
100
398
200
shy
156
620
5shy
148
2219
42
1867
65
432
8ndash13
10
158
62
4323+
311
N
A
Astronaut photography as digital data for remote sensing 4431
Tab
le5
(Cont
inued
)
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
7L
imm
enB
igh
tS
TS
093
-704
-50
250
283
shy15
0135
0shy
15
013
58
623
120
859
16
963
0ndash67
3125
16
863
0ndash67
612
613
2P
CR
e-es
tim
ated
shy14
733
1352
67
64
5128
623
ndash65
5123
128
623
ndash65
712
3127
Auxilia
ryP
oin
t611
062
3ndash65
112
312
5F
igu
re10
N
ear
Au
stin
T
XS
TS
095
-716
-46
250
543
30
5shy
975
28
6shy
1007
120
230
375
34
613
5ndash157
281
34
1136ndash16
128
636
0
1 All
pho
togr
aphs
exce
pt
on
eta
ken
wit
hH
ass
elbla
dca
mer
aso
dis
tan
ceacr
oss
the
image
d=
55m
m
for
S73
E51
45
taken
wit
hel
ectr
onic
still
cam
erad=
276
5m
mho
rizo
nta
l2 P
Cand
SN
are
giv
enin
the
form
(lati
tud
elo
ngit
ude)
deg
rees
w
ith
neg
ativ
en
um
ber
sre
pre
senti
ng
south
and
wes
tA
ccura
cyo
fP
Cis
plusmn0
5and
SN
isplusmn
01
as
reco
rded
inth
eA
stro
naut
Ph
oto
gra
phy
Data
base
ex
cep
tw
her
en
ote
dth
atce
ntr
ep
oin
tw
asre
-est
imate
dw
ith
grea
ter
accu
racy
3 C
alc
ula
ted
usi
ng
the
mea
nofth
em
inim
um
an
dm
axim
um
esti
mate
sof
DF
orm
ula
tion
2co
nsi
der
sD
inth
ed
irec
tion
per
pen
dic
ula
rto
the
Pri
nci
pal
Lin
eo
nly
4 C
alc
ula
ted
inho
rizo
nta
lan
dve
rtic
aldir
ecti
on
sb
ut
still
ass
um
ing
geom
etry
of
gure
6(B
)F
irst
nu
mb
eris
the
mea
no
fth
em
inim
um
Dat
the
bott
om
of
the
ph
oto
and
the
maxi
mum
Dat
the
top
of
the
ph
oto
(range
show
nin
the
tab
le)
and
isco
mpara
ble
toF
orm
ula
tio
n2
Sec
ond
num
ber
isth
em
ean
of
Des
tim
ated
alon
gth
eP
rin
cip
alline
and
alo
ng
the
ou
tsid
eed
ge
(range
not
show
n)
5 Alt
itu
de
isap
pro
xim
ate
for
mis
sion
as
exac
tti
me
pho
togr
ap
hw
asta
ken
and
corr
esp
on
din
gn
adir
data
are
no
tava
ilab
le
Lack
of
nad
irdata
pre
ven
tsap
plica
tion
of
Form
ula
tion
s2
an
d3
6 Au
xilliar
ypo
int
add
edw
as
shy14
80
lati
tud
e13
51
2lo
ngit
ude
shy46
5degro
tati
on
angle
in
stru
ctio
ns
for
esti
mati
ng
the
rota
tio
nan
gle
para
met
erare
conta
ined
as
par
tof
the
scal
eca
lcu
lato
rsp
read
shee
tdo
cum
enta
tio
n
J A Robinson et al4432
a park at Johnson Space Center) that could clearly be distinguished on the photo-graph was 853 m wide
For the photograph of Houston taken with a 250 mm lens ( gure 3(C)) anddigitized from second generation lm at 2400 ppi (106 mm pixel Otilde 1 ) Ellington Fieldrunway 4-22 is 3 pixels in width and 1613 pixels in length so pixels represent151ndash152 m on the ground These results compare favourably with the minimumestimate of 152 m using Formulation 2 and 154ndash185 m pixels using Formulation 3(table 5) For an estimate of GRD using an 8times8 inch print (1369 enlargement) and4times magni cation the smallest street that could clearly be distinguished was 822 mwide the same feature could barely be distinguished on the digitized image
We also made an empirical estimate of spatial resolution for lower contrastvegetation boundaries By clearing forest so that a pattern would be visible to landingaircraft a landowner outside Austin Texas (see also aerial photo in Lisheron 2000)created a target that is also useful for evaluating spatial resolution of astronautphotographs The forest was selectively cleared in order to spell the landownerrsquosname lsquoLUECKErsquo with the remaining trees ( gure 10) According to local surveyorswho planned the clearing the plan was to create letters that were 3100 fttimes1700 ft(9449 mtimes5182 m) Photographed at a high altitude relative to most Shuttle missions(543 km) with a 250 mm lens Formula 3 predicts that each pixel would represent anarea 286 mtimes360 m on the ground (table 5) When original lm was digitized at2400 ppi (106 mm pixel Otilde 1 ) letters correspond to 294times188 pixels for a comparablepixel size of 27ndash32 m
6 Summary and conclusionsAstronaut photographs can be an excellent source of data for remote sensing
applications Best-case resolutions are similar to that for Landsat or SPOT withpixels as small as lt10 m It was estimated that of images taken to date 504 hadlens and obliquity characteristics that make them potential candidates for remotesensing information Digitized astronaut photographs can be overlaid with othersatellite data using GIS (Eckardt et al 2000) or used to ll in gaps in time serieswhen other imagery is not available As a source of public-domain information theycan be very useful for scientists who do not have access to satellite imagery eitherbecause of the expense of image acquisition (to the end user) or the computersystems needed for processing satellite images These diVerences in image acquisitioncosts (to the user) are summarized in table 6
Searching of the complete database of NASA astronaut photography includinglow-spatial-resolution browse images is available via the Web (OYce of EarthSciences 2000) This provides nearly global access for identifying images that willcontribute to a speci c scienti c project For the most detailed studies digitalproducts posted to the web will not be of suYcient quality To date we have madeit a practice of digitizing small numbers of images when requested by scientists atno charge Such requests (including a description of the project involved) can bemade through the authors or using the contact information listed on the websiteInvestigators with needs for larger numbers of digital images have also been servedthrough collaborative agreements
In our experience scientists that nd the data most valuable are those in develop-ing countries those studying areas that have not usually been targets for the majorsatellites those having diYculty in nding low-cloud images those interested inconstructing time series or those interested in using a large number of images
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
J A Robinson et al4412
format of the lm (or CCD) and image size projected onto the lm (or CCD) aresummarized for all the diVerent cameras own in table 2
34 L ook angle or obliquityNo handheld photographs can be considered perfectly nadir they are taken at a
variety of look angles ranging from near vertical ( looking down at approximatelythe nadir position of the spacecraft ) to high oblique (images that include the curvatureof Earth) Imaging at oblique look angles leads to an image where scale degradesaway from nadir A set of views of the same area from diVerent look angles is shownin gure 4 The rst two shots of the island of Hawaii were taken only a few secondsapart and with the same lens The third photograph was taken on a subsequentorbit and with a shorter lens The curvature of the Earth can be seen in the upperleft corner
Obliquity can be described qualitatively ( gure 4) or quantitatively as the lookangle (t gure 1 calculations described in formulation 2 (sect52)) Because obliquityand look angle have such a dramatic in uence on the footprint we summarize thedatabase characteristics relative to these two parameters Figure 5 is a breakdownof the spatial resolution characteristics of low oblique and near vertical photographsin the NASA Astronaut Photography database Number of photographs are grouped(1) by calculated values for look angle (t ) and (2) by altitude After observing theoverlap between near vertical and low oblique classes we are currently restructuringthis variable (lsquotiltrsquo) in the database to provide a measure of t when available Userswill still be able to do searches based on the qualitative measures but these measureswill be more closely tied to actual look angle
341 Obliquity and georeferencing digitised photographsOften the rst step in a remote sensing analysis of a digitized astronaut photo-
graph is to georeference the data and resample it to conform to a known mapprojection Details and recommendations for resampling astronaut photographydata are provided by Robinson et al (2000a 2000c) and a tutorial is also available(McRay et al 2000) Slightly oblique photographs can be geometrically correctedfor remote sensing purposes but extremely oblique photographs are not suited forgeometric correction When obliquity is too great the spatial scale far away fromnadir is much larger than the spatial scale closer to nadir resampling results areunsuitable because pixels near nadir are lost as the image is resampled while manypixels far away from nadir are excessively replicated by resampling
To avoid the generation of excess pixels during georeferencing the pixel sizes ofthe original digitized image should be smaller than the pixels in the nal resampledimage Calculations of original pixel size using methods presented below can beuseful in ensuring meaningful resampling For slightly oblique images formulation 3(sect53) can be used to estimate pixel sizes at various locations in a photograph (nearnadir and away from nadir) and these pixel sizes then used to determine a reasonablepixel scale following resampling
4 Other characteristics that in uence spatial resolutionBeyond basic geometry many other factors internal to the astronaut photography
system impact the observed ground resolved distance in a photograph The lensesand cameras already discussed are imperfect and introduce radiometric degradations(Moik 1980)Vignetting (slight darkening around the edges of an image) occurs in
Astronaut photography as digital data for remote sensing 4413
Tab
le2
Max
imu
mpo
ssib
led
igit
alsp
atia
lre
solu
tio
n(p
ixel
size
inm
)fo
rca
mer
asy
stem
scu
rren
tly
use
dby
ast
ron
auts
top
ho
togr
aph
eart
hb
ase
do
nth
ege
om
etry
of
alt
itud
ele
ns
and
lm
size
C
alc
ula
tion
sas
sum
eth
atco
lour
tran
spar
ency
lm
isdig
itiz
edat
2400
ppi
(pix
els
per
inch
10
6m
mpix
elOtilde
1 )
Min
imu
mgro
un
ddis
tance
repre
sente
dby
1p
ixel
(m)
As
afu
nct
ion
of
alti
tude1
Imag
esi
zeM
inim
um
alt
itud
eM
edia
nalt
itu
de
Maxi
mum
alt
itud
eC
amer
asy
stem
mm
timesm
mp
ixel
s2(2
22k
m)
(326
km
)(6
11
km
)
Lin
ho
f125
mm
90
mm
lens
105
times120
8978times
11
434
261
383
718
250
mm
lens
94
138
259
Has
selb
lad
70
mm
100
mm
lens
55times
55
5198
times519
823
534
564
725
0m
mle
ns
94
138
259
Nik
on
35m
m30
0m
mle
ns
36times
24
3308
times217
478
115
216
400
mm
lens
59
86
162
ESC
35m
m18
0m
mle
ns3
27
6times
18
530
69times
204
311
116
330
630
0m
mle
ns
67
98
184
400
mm
lens
50
74
138
1 Alt
itud
essh
ow
nar
em
inim
um
m
edia
nan
dm
axim
um
thro
ugh
ST
S-8
9(J
anuary
199
8N
=84
mis
sion
s)
2 Act
ual
pix
els
for
ES
C
oth
erw
ise
assu
med
dig
itiz
ati
on
at240
0pp
i(p
ixel
sp
erin
ch)
equ
alto
10
6m
mpix
elOtilde
1 3 T
he
mo
stco
mm
on
con
gura
tion
for
Eart
hph
oto
gra
ph
yas
part
of
the
Eart
hK
AM
pro
ject
J A Robinson et al4414
Figure 4 De nitions of look angle for photographs taken from low orbit around EarthThe example photographs of Hawaii were all taken from approximately the samealtitude (370ndash398 km) and with the same (250 mm) lens on a Hasselblad camera(STS069-701-69 STS069-701-70 and STS069-729-27 [100 mm lens])
astronaut photography due to light path properties of the lenses used A lack of atness of the lm at the moment of exposure (because cameras used in orbit donot incorporate a vacuum platen) also introduces slight degradations A numberof additional sources of image degradation that would aVect GRD of astronautphotographs are listed
41 Films and processing411 Colour reversal lms
Most lms used in NASA handheld cameras in orbit have been E-6 processcolour reversal lms (Ektachromes) that have a nominal speed of 64 to 100 ASAalthough many other lms have been own (table 3) Reversal lms are used becausedamage from radiation exposure while outside the atmospheric protection of Earthis more noticeable in negative lms than in positive lms (Slater 1996) The choiceof ASA has been to balance the coarser grains ( lower lm resolving power) of high-speed lms with the fact that slower lms require longer exposures and thus areaVected more by vehicle motion (approximately 73 km s Otilde 1 relative to Earthrsquos surfacefor the Space Shuttle) Extremely fast lms (gt400 ASA) are also more susceptibleto radiation damage in orbit (particularly during long-duration or high-altitudemissions Slater 1996) and have not traditionally been used for Earth photography The manufacturer stock numbers identifying the lm used are available for eachimage in the Astronaut Photography Database (OYce of Earth Sciences 2000)
412 Colour infrared lmsColour infrared lm (CIR) was used during an Earth-orbiting Apollo mission
(Colwell 1971) in the multispectral S-190A and the high-resolution S-190B camera
Astronaut photography as digital data for remote sensing 4415
Figure 5 (a) Distributions of astronaut photographs by look angle oV-nadir (calculated fromcentre point and nadir point using Formulation 2) Data included for all photographsfor which centre and nadir points are known with high oblique look angles (n=55 155) excluded (Total sample shown is 108 293 of 382 563 total records in thedatabase when compiled For records where nadir data was not available number ofobservations for qualitative look angles are near vertical 14 086 low oblique 115 834undetermined 89 195) (b) Altitude distribution for astronaut photographs of Earthtaken with the Hasselblad camera Data included for Hasselblad camera only focallength determined with high oblique look angles excluded (Total sample shown is159 046 of 206 971 Hasselblad records with known altitude and of 382 563 total recordsin the database when compiled For Hasselblad records with known altitude data notshown includes high oblique photographs using 100 mm and 250 mm lenses n=20 262and all look angles with other lenses n=27 663)
systems on Skylab (NASA 1974 Wilmarth et al 1977) and occasionally on Shuttlemissions and Shuttle-Mir missions (table 3) The CIR lm used is a three-layerAerochrome having one layer that is sensitive to re ected solar infrared energy toapproximately 900 nm (Eastman Kodak Company 1998) protective coatings onmost spacecraft windows also limit IR transmittance between 800 mm and 1000 nmThis layer is extremely sensitive to the temperature of the lm which createsunpredictable degradation of the IR signature under some space ight conditions
J A Robinson et al4416
Tab
le3
Det
ails
on
lm
suse
dfo
rast
ron
aut
pho
togr
aphy
of
Eart
h
Dat
ain
clud
ep
ho
togr
aphy
fro
mall
NA
SA
hum
an
spac
eig
ht
mis
sio
ns
thro
ugh
ST
S-9
6(J
un
e19
99
378
461
tota
lfr
ames
inth
edata
base
)F
ilm
sw
ith
lt50
00fr
am
esex
po
sed
and
lm
suse
dfo
r35
mm
cam
eras
only
are
no
tin
clud
ed
Res
olv
ing
Nu
mb
erof
Film
man
ufa
ctu
rer
AS
Ap
ow
er1
fram
esP
erio
dof
use
Cam
eras
Colo
ur
posi
tive
Kod
akA
eroch
rom
eII
(2447
2448
SO
-131S
O-3
68
)32
40ndash50
513
8196
8ndash19
85
Hass
elbla
dL
inh
of
Kod
akE
kta
chro
me
(5017
502
56
017
6118
SO
-121
64
50
113
932
196
5ndash19
68
Hass
elbla
dL
inh
of
Role
iex
SO
-217
Q
X-8
24
QX
-868
)19
81ndash1993
Mau
rer
Nik
on
Kod
akL
um
iere
(5046
5048
)100
105
743
199
4ndash19
98
Hass
elbla
dL
inho
fK
od
akE
lite
100
S(5
069
)100
41
486
199
6ndashp
rese
nt
Hass
elbla
d2
Fu
jiV
elvia
50C
S135
-36
32
13
882
1992ndash19
97H
ass
elbla
dN
iko
nK
od
akH
i-R
esolu
tio
nC
olo
r(S
O-2
42S
O-3
56
)10
096
74
1973ndash19
75H
ass
elbla
d
S19
0A
S190
B
Infr
are
dand
bla
ck-a
nd-w
hit
e
lms
Kod
akA
eroch
rom
eC
olo
rIn
frar
ed40
32
40
704
196
5ndashp
rese
nt2
Hass
elbla
dL
inh
of
Nik
on
S190
A
(8443
34
432443
)S
190B
Kod
akB
ampW
Infr
are
d(2
424
)32
10
553
197
3ndash19
74
S190
AK
od
akP
anato
mic
-XB
ampW
(SO
-022
)63
10
657
197
3ndash19
74
S190
A
1A
tlo
wco
ntr
ast
(ob
ject
back
grou
nd
rati
o16
1)
aspro
vid
edby
Ko
dak
and
list
edby
Sla
ter
etal
(1
983
256
)R
esolv
ing
pow
erw
as
dis
con
tin
ued
fro
mth
ete
chn
ical
test
sper
form
edb
yK
odak
inapp
roxim
atel
y19
93
an
dit
isn
ot
kn
ow
nif
curr
ent
lm
sh
ave
sim
ilar
reso
lvin
gpo
wer
too
lder
ver
sion
so
fsi
milar
lm
s(K
T
eite
lbaum
K
od
akp
erso
nal
com
mu
nic
atio
n)
Th
egra
nu
lari
tym
easu
reno
wpro
vid
edb
yK
od
ak
can
no
tb
ere
adily
con
ver
ted
toa
spat
ial
reso
luti
on
2 The
Lin
hof
was
use
dag
ain
on
ST
S-1
03in
Dec
emb
er19
99(d
ata
not
cata
logued
asof
the
pre
para
tion
of
this
tab
le)
usi
ng
Ko
dak
Elite
100S
lm
3 B
egin
nin
gin
July
1999
2443
colo
ur-
infr
are
d
lmpro
cess
was
chan
ged
from
EA
-5to
E-6
Astronaut photography as digital data for remote sensing 4417
413 Film duplicationThe original lm is archived permanently after producing about 20 duplicate
printing masters (second generation) Duplicate printing masters are disseminatedto regional archives and used to produce products (thirdndashfourth generation) for themedia the public and for scienti c use When prints slides transparencies anddigital products are produced for public distribution they are often colour adjustedto correspond to a more realistic look Digital products from second generationcopies can be requested by scienti c users (contact the authors for information onobtaining such products for a particular project) and these are recommended forremote sensing applications Care should be taken that the digital product acquiredfor remote sensing analysis has not been colour adjusted for presentation purposesBased on qualitative observations there is little visible increase in GRD in the thirdor fourth generation products There is a signi cant degradation in delity of colourin third and fourth generation duplicates because an increase in contrast occurswith every copy
42 Shutter speedsThe impact of camera motion (both due to the photographer and to the motion
of the spacecraft ) on image quality is determined by shutter speedmdash1250 to 1500second have been used because slower speeds record obvious blurring caused bythe rapid motion of the spacecraft relative to the surface of the Earth Generally1250 was used for ISO 64 lms (and slower) and after the switch to ISO 100 lmsa 1500 setting became standard New Hasselblad cameras that have been ownbeginning with STS-92 in October 2000 vary shutter speed using a focal plane shutterfor exposure bracketing (rather than varying aperture) and 1250 1500 and 11000second are used
Across an altitude range of 120ndash300 nautical miles (222ndash555 km) and orbitalinclinations of 285deg and 570deg the median relative ground velocity of orbitingspacecraft is 73 km s Otilde 1 At a nominal shutter speed of 1500 the expected blur dueto motion relative to the ground would be 146 m When cameras are handheld (asopposed to mounted in a bracket) blur can be reduced by physical compensationfor the motion (tracking) by the photographer many photographs show a level ofdetail indicating that blur at this level did not occur (eg gure 3(d )) Thus motionrelative to the ground is not an absolute barrier to ground resolution for handheldphotographs
43 Spacecraft windowsMost of the photographs in the NASA Astronaut Photography Database were
taken through window ports on the spacecraft used The transmittance of the windowport may be aVected by material inhomogeniety of the glass the number of layersof panes coatings used the quality of the surface polish environmentally inducedchanges in window materials (pressure loads thermal gradients) or deposited con-tamination Such degradation of the window cannot be corrected is diVerent foreach window and changes over time See Eppler et al (1996) and Scott (2000) fordiscussion of the spectral transmittance of the fused quartz window that is part ofthe US Laboratory Module (Destiny) of the International Space Station
44 Digitized images from lmWhen lm is scanned digitally the amount of information retained depends on
the spectral information extracted from the lm at the spatial limits of the scanner
J A Robinson et al4418
To date standard digitizing methodologies for astronaut photographs have not beenestablished and lm is digitized on a case-by-case basis using the equipment available
441 Digitizing and spatial resolutionLight (1993 1996) provided equations for determining the digitizing spatial
resolution needed to preserve spatial resolution of aerial photography based on thestatic resolving power (AWAR) for the system For lms where the manufacturer hasprovided data (Eastman Kodak 1998 K Teitelbaum personal communication) the resolving power of lms used for astronaut photography ranges from 32 lp mm Otilde 1to 100 lp mm Otilde 1 at low contrast (objectbackground ratio 161 table 3) The AWARfor the static case of the Hasselblad camera has been measured at high and lowcontrast (using lenses shown in table 1) with a maximum of approximately55 lp mm Otilde 1 (Fred Pearce unpublished data)
Based on the method of Light (1993 1996) the dimension of one spatial resolutionelement for a photograph with maximum static AWAR of 55 lp mm Otilde 1 would be18 mm lp Otilde 1 The acceptable range of spot size to preserve spatial information wouldthen be
18 mm
2 atilde 2aring scan spot size aring
18 mm
2(1)
and 6 mm aring scan spot size aring 9 mm Similarly for a more typical static AWAR of30 lp mm Otilde 1 (33 mm lpOtilde 1 low contrast Fred Pearce unpublished data) 11 mm aring scanspot sizearing 17 mm These scan spot sizes correspond to digitizing resolutions rangingfrom 4233ndash2822 ppi (pixels inch Otilde 1 ) for AWAR of 55 lp mm Otilde 1 and 2309ndash1494 ppi forAWAR of 30 lp mm Otilde 1 Scan spot sizes calculated for astronaut photography arecomparable to those calculated for the National Aerial Photography Program(9ndash13 mm Light 1996)
Widely available scanners that can digitize colour transparency lm currentlyhave a maximum spatial resolution of approximately 2400 ppi (106 mm pixel Otilde 1) Forexample we routinely use an Agfa Arcus II desktop scanner (see also Baltsavias1996) with 2400 ppi digitizing spatial resolution (2400 ppi optical resolution in onedirection and 1200 ppi interpolated to 2400 ppi in the other direction) and 2400 ppiwas used to calculate IFOV equivalents (tables 2 and 4) For some combinations oflens camera and contrast 2400 ppi will capture nearly all of the spatial informationcontained in the lm However for lm with higher resolving power thanEktachrome-64 for better lenses and for higher contrast targets digitizing at 2400 ppiwill not capture all of the spatial information in the lm
Improvements in digitizing technology will only produce real increases in IFOVto the limit of the AWAR of the photography system The incremental increase inspatial information above 3000 ppi (85 mm pixel Otilde 1 ) is not likely to outweigh thecosts of storing large images (Luman et al 1997) At 2400 ppi a Hasselblad frameis approximately 5200times5200 or 27 million pixels (table 2) while the same imagedigitized at 4000 ppi would contain 75 million pixels
Initial studies using astronaut photographs digitized at 2400 ppi (106 mm pixel Otilde 1 Webb et al 2000 Robinson et al 2000c) indicate that some GRD is lost comparedto photographic products Nevertheless the spatial resolution is still comparablewith other widely used data sources (Webb et al 2000) Robinson et al (2000c)found that digitizing at 21 mm pixel Otilde 1 provided information equivalent to 106 mmpixel Otilde 1 for identifying general urban area boundaries for six cities except for a single
Astronaut photography as digital data for remote sensing 4419
photo that required higher spatial resolution digitizing (it had been taken with ashorter lens and thus had less spatial resolution) Part of the equivalence observedin the urban areas study may be attributable to the fact that the atbed scannerused interpolates from 1200 ppi to 2400 ppi in one direction The appropriate digitiz-ing spatial resolution will in part depend on the scale of features of interest to theresearchers with a maximum upper limit set of a scan spot size of approximately6 mm (4233 ppi)
442 Digitizing and spectral resolutionWhen colour lm is digitized there will be loss of spectral resolution The three
lm emulsion layers (red green blue) have relatively distinct spectral responses butare fused together so that digitizers detect one colour signal and must convert itback into three (red green blue) digital channels Digital scanning is also subject tospectral calibration and reproduction errors Studies using digital astronaut photo-graphs to date have used 8 bits per channel However this is another parameterthat can be controlled when the lm is digitized We do not further address spectralresolution of digitally scanned images in this paper
45 External factors that in uence GRDAlthough not discussed in detail in this paper factors external to the spacecraft
and camera system (as listed by Moik (1980) for remote sensing in general) alsoimpact GRD These include atmospheric interference (due to scattering attenuationhaze) variable surface illumination (diVerences in terrain slope and orientation) andchange of terrain re ectance with viewing angle (bidirectional re ectance) For astro-naut photographs variable illumination is particularly important because orbits arenot sun-synchronous Photographs are illuminated by diVerent sun angles and imagesof a given location will have colour intensities that vary widely In addition theviewing angle has an eVect on the degree of object-to-backgroun d contrast andatmospheric interference
5 Estimating spatial resolution of astronaut photographsIn order to use astronaut photographs for digital remote sensing it is important
to be able to calculate the equivalent to an IFOVmdashthe ground area represented bya single pixel in a digitized orbital photograph The obliquity of most photographsmeans that pixel lsquosizesrsquo vary at diVerent places in an image Given a ground distanceD represented by a photograph in each direction (horizontal vertical ) an approxi-mate average pixel width (P the equivalent of IFOV) for the entire image can becalculated as follows
P=D
dS(2)
where D is the projected distance on the ground covered by the image in the samedirection as the pixel is measured d is the actual width of the image on the original lm (table 1) and S is the digitizing spatial resolution
Here we present three mathematical formulations for estimating the size of thefootprint or area on the ground covered by the image Example results from theapplication of all three formulations are given in table 5 The rst and simplestcalculation (formulation 1) gives an idea of the maximum spatial resolution attainableat a given altitude of orbit with a given lm format and a perfectly vertical (nadir)
J A Robinson et al4420
view downward Formulation 2 takes into account obliquity by calculating lookangle from the diVerence between the location on the ground represented at thecentre of the photograph and the nadir location of the spacecraft at the time thephotograph was taken ( gure 1) Formulation 3 describes an alternate solution tothe oblique look angle problem using coordinate-system transformations Thisformulation has been implemented in a documented spreadsheet and is availablefor download (httpeoljscnasagovsseopFootprintCalculatorxls) and is in theprocess of being implemented on a Web-based user interface to the AstronautPhotography Database (OYce of Earth Sciences 2000)
Although formulations 2 and 3 account for obliquity for the purposes of calcula-tion they treat the position of the spacecraft and position of the camera as one Inactuality astronauts are generally holding the cameras by hand (although camerasbracketed in the window are also used) and the selection of window position of theastronaut in the window and rotation of the camera relative to the movement ofthe spacecraft are not known Thus calculations using only the photo centre point(PC) and spacecraft nadir point (SN) give a locator ellipse and not the locations ofthe corners of the photograph A locator ellipse describes an estimated area on theground that is likely to be included in a speci c photograph regardless of the rotationof the lm plane about the camerarsquos optical axis ( gure 6)
Estimating the corner positions of the photo requires additional user input of asingle auxiliary pointmdasha location on the image that has a known location on theground Addition of this auxiliary point is an option available to users of thespreadsheet An example of the results of adding an auxiliary point is shown in gure 7 with comparisons of the various calculations in table 5
51 Formulation 1 Footprint for a nadir viewThe simplest way to estimate footprint size is to use the geometry of camera lens
and spacecraft altitude to calculate the scaling relationship between the image in the
Figure 6 Sketch representation of the locator ellipse an estimated area on the ground thatis likely to be included in a speci c photograph regardless of the rotation of the lmplane about the camerarsquos optical axis
Astronaut photography as digital data for remote sensing 4421
Figure 7 An example of the application of the Low Oblique Space Photo FootprintCalculator The photograph (STS093-704-50) of Limmen Bight Australia was takenfrom an altitude of 283 km using the Hasselblad camera with a 250 mm lens Mapused is 11 000 000 Operational Navigational Chart The distances shown equate toan average pixel size of 124 m for this photograph
lm and the area covered on the ground For a perfect nadir view the scalerelationship from the geometry of similar triangles is
d
D=
f
H(3)
where d=original image size D=distance of footprint on the ground f =focallength of lens and H=altitude Once D is known pixel size ( length=width) can becalculated from equation (2) These calculations represent the minimum footprintand minimum pixel size possible for a given camera system altitude and digitisingspatial resolution (table 2) Formulation 1 was used for calculating minimum pixelsizes shown in tables 2 and 4
By assuming digitizing at 2400 ppi (106 mm pixel Otilde 1 ) currently a spatial resolutioncommonly attainable from multipurpose colour transparency scanners (see sect441)this formulation was used to convert area covered to an IFOV equivalent for missionsof diVerent altitudes (table 2) Table 4 provides a comparison of IFOV of imagesfrom various satellites including the equivalent for astronaut photography Valuesin this table were derived by using formulation 1 to estimate the area covered becausea perfect nadir view represents the best possible spatial resolution and smallest eldof view that could be obtained
52 Formulation 2 Footprint for oblique views using simpli ed geometry and thegreat circle distance
A more realistic approach to determining the footprint of the photographaccounts for the fact that the camera is not usually pointing perfectly down at the
J A Robinson et al4422
Table 4 Comparison of pixel sizes (instantaneous elds of view) for satellite remote sensorswith pixel sizes for near vertical viewing photography from spacecraft Note that alldata returned from the automated satellites have the same eld of view while indicatedpixel sizes for astronaut photography are best-case for near vertical look angles See gure 5 for distributions of astronaut photographs of various look angles High-altitude aerial photography is also included for reference
Altitude PixelInstrument (km) width (mtimesm) Bands1
Landsat Multipectral Scanner2 (MSS 1974-) 880ndash940 79times56 4ndash5Landsat Thematic Mapper (TM 1982-) ~705 30times30 7Satellite pour lrsquoObservation de la Terre (SPOT 1986-) ~832
SPOT 1-3 Panchromatic 10times10 1SPOT 1-3 Multispectral (XS) 20times20 3
Astronaut Photography (1961-)Skylab3 (1973-1974 ) ~435
Multispectral Camera (6 camera stations) 17ndash19times17ndash19 3ndash8Earth Terrain Camera (hi-res colour) 875times875 3
Space Shuttle5 (1981-) 222ndash611Hasselblad 70 mm (250mm lens colour) vertical view 9ndash26times9ndash26 3Hasselblad 70 mm (100mm lens colour) vertical view 23ndash65times23ndash65 3Nikon 35 mm (400mm lens colour lm or ESC) vertical 5ndash16times5ndash16 3
High Altitude Aerial Photograph (1100 000) ~12 2times2 3
1Three bands listed for aerial photography obtainable by high-resolution colour digitizingFor high-resolution colour lm at 65 lp mmOtilde 1 (K Teitelbaum Eastman Kodak Co personalcommunication) the maximum information contained would be 3302 ppi
2Data from Simonett (1983Table 1-2)3Calculated as recommended by Jensen (19831596) the dividend of estimated ground
resolution at low contrast given in NASA (1974179-180) and 244Six lters plus colour and colour infrared lm (NASA 19747) 5Extracted from table 2
nadir point The look angle (the angle oV nadir that the camera is pointing) can becalculated trigonometrically by assuming a spherical earth and calculating the dis-tance between the coordinates of SN and PC ( gure 1) using the Great Circledistance haversine solution (Sinnott 1984 Snyder 198730ndash32 Chamberlain 1996)The diVerence between the spacecraft centre and nadir point latitudes Dlat=lat 2 shy lat 1 and the diVerence between the spacecraft centre and nadir pointlongitudes Dlon=lon 2 shy lon 1 enter the following equations
a=sin2ADlat
2 B+cos(lat 1) cos(lat 2) sin2ADlon
2 B (4)
c=2 arcsin(min[1 atilde a]) (5)
and oVset=R c with R the radius of a spherical Earth=6370 kmThe look angle (t ) is then given by
t=arctanAoffset
H B (6)
Assuming that the camera was positioned so that the imaginary line between thecentre and nadir points (the principal line) runs vertically through the centre of thephotograph the distance between the geometric centre of the photograph (principalpoint PP) and the top of the photograph is d2 ( gure 8(a)) The scale at any point
Astronaut photography as digital data for remote sensing 4423
(a) (b)
Figure 8 The geometric relationship between look angle lens focal length and the centrepoint and nadir points of a photograph (see also Estes and Simonett 1975) Note thatthe Earthrsquos surface and footprint represented in gure 1 are not shown here only arepresentation of the image area of the photograph Variables shown in the gurelook angle t focal length f PP centre of the image on the lm SN spacecraft nadird original image size (a) The special case where the camera is aligned squarely withthe isometric parallel In this case tilt occurs along a plane parallel with the centre ofthe photograph When the conditions are met calculations using Formulation 3 willgive the ground coordinates of the corner points of the photograph (b) The generalrelationship between these parameters and key photogrammetric points on the photo-graph For this more typical case coordinates of an additional point in the photographmust be input in order to adjust for twist of the camera and to nd the corner pointcoordinates
in the photograph varies as a function of the distance y along the principal linebetween the isocentre and the point according to the relationship
d
D=
f shy ysint
H(7)
(Wong 1980 equation (214) H amp the ground elevation h) Using gure 8(a) at thetop of the photo
y= f tanAt
2B+d
2(8)
and at the bottom of the photo
y= f tanAt
2Bshyd
2(9)
Thus for given PC and SN coordinates and assuming a photo orientation as in gure 8(a) and not gure 8(b) we can estimate a minimum D (using equations (7)and (8)) and a maximum D (using equations (7) and (9)) and then average the twoto determine the pixel size (P ) via equation (2)
J A Robinson et al4424
53 Formulation 3 T he L ow Oblique Space Photo Footprint CalculatorThe Low Oblique Space Photo Footprint Calculator was developed to provide
a more accurate estimation of the geographic coordinates for the footprint of a lowoblique photo of the Earthrsquos surface taken from a human-occupied spacecraft inorbit The calculator performs a series of three-dimensional coordinate transforma-tions to compute the location and orientation of the centre of the photo exposureplane relative to an earth referenced coordinate system The nominal camera focallength is then used to create a vector from the photorsquos perspective point througheach of eight points around the parameter of the image as de ned by the formatsize The geographical coordinates for the photo footprint are then computed byintersecting these photo vectors with a spherical earth model Although more sophist-icated projection algorithms are available no signi cant increase in the accuracy ofthe results would be produced by these algorithms due to inherent uncertainties inthe available input data (ie the spacecraft altitude photo centre location etc)
The calculations were initially implemented within a Microsoft Excel workbookwhich allowed one to embed the mathematical and graphical documentation nextto the actual calculations Thus interested users can inspect the mathematical pro-cessing A set of error traps was also built into the calculations to detect erroneousresults A summary of any errors generated is reported to the user with the resultsof the calculations Interested individuals are invited to download the Excel work-book from httpeoljscnasagovsseopFootprintCalculatorxls The calculations arecurrently being encoded in a high-level programming language and should soon beavailable alongside other background data provided for each photograph at OYceof Earth Sciences (2000)
For the purposes of these calculations a low oblique photograph was de ned as
one with the centre within 10 degrees of latitude and longitude of the spacecraftnadir point For the typical range of spacecraft altitudes to date this restricted thecalculations to photographs in which Earthrsquos horizon does not appear (the generalde nition of a low oblique photograph eg Campbell 199671 and gure 4)
531 Input data and resultsUpon opening the Excel workbook the user is presented with a program intro-
duction providing instructions for using the calculator The second worksheet tablsquoHow-To-Usersquo provides detailed step-by-step instructions for preparing the baselinedata for the program The third tab lsquoInput-Output rsquo contains the user input eldsand displays the results of the calculations The additional worksheets contain theactual calculations and program documentation Although users are welcome toreview these sheets an experienced user need only access the lsquoInput-Outputrsquo
spreadsheetTo begin a calculation the user enters the following information which is available
for each photo in the NASA Astronaut Photography Database (1) SN geographicalcoordinates of spacecraft nadir position at the time of photo (2) H spacecraftaltitude (3) PC the geographical coordinates of the centre of the photo (4) f nominal focal length and (5) d image format size The automatic implementationof the workbook on the web will automatically enter these values and completecalculations
For more accurate results the user may optionally enter the geographic co-ordinates and orientation for an auxiliary point on the photo which resolves thecamerarsquos rotation uncertainty about the optical axis The auxiliary point data must
Astronaut photography as digital data for remote sensing 4425
be computed by the user following the instruction contained in the lsquoHow-To-Usersquotab of the spreadsheet
After entering the input data the geographic coordinates of the photo footprint(ie four photo corner points four points at the bisector of each edge and the centreof the photo) are immediately displayed below the input elds along with any errormessages generated by the user input or by the calculations ( gure 7) Although
results are computed and displayed they should not be used when error messagesare produced by the program The program also computes the tilt angle for each ofthe photo vectors relative to the spacecraft nadir vector To the right of the photofootprint coordinates is displayed the arc distance along the surface of the sphere
between adjacent computed points
532 Calculation assumptions
The mathematical calculations implemented in the Low Oblique Space PhotoFootprint Calculator use the following assumptions
1 The SN location is used as exact even though the true value may vary by upto plusmn01 degree from the location provided with the photo
2 The spacecraft altitude is used as exact Although the determination of the
nadir point at the instant of a known spacecraft vector is relatively precise(plusmn115times10 Otilde 4 degrees) the propagator interpolates between sets of approxi-mately 10ndash40 known vectors per day and the time code recorded on the lmcan drift Thus the true value for SN may vary by up to plusmn01 degree fromthe value provided with the photo
3 The perspective centre of the camera is assumed to be at the given altitudeover the speci ed spacecraft nadir location at the time of photo exposure
4 The PC location is used as exact even though the true value may vary by upto plusmn05deg latitude and plusmn05deg longitude from the location provided with the
photo5 A spherical earth model is used with a nominal radius of 6 372 16154 m (a
common rst-order approximation for a spherical earth used in geodeticcomputations)
6 The nominal lens focal length of the camera lens is used in the computations(calibrated focal length values are not available)
7 The photo projection is based on the classic pin-hole camera model8 No correction for lens distortion or atmospheric re ection is made
9 If no auxiliary point data is provided the lsquoTop of the Imagersquo is orientedperpendicular to the vector from SN towards PC
533 T ransformation from Earth to photo coordinate systemsThe calculations begin by converting the geographic coordinates ( latitude and
longitude) of the SN and PC to a Rectangular Earth-Centred Coordinate System(R-Earth) de ned as shown in gure 9 (with the centre of the Earth at (0 0 0))
Using the vector from the Earthrsquos centre through SN and the spacecraft altitude thespacecraft location (SC) is also computed in R-Earth
For ease of computation a Rectangular Spacecraft-Centred Coordinate System(R-Spacecraft ) is de ned as shown in gure 9 The origin of R-Spacecraft is located
at SC with its +Z-axis aligned with vector from the centre of the Earth throughSN and its +X-axis aligned with the vector from SN to PC ( gure 9) The speci c
J A Robinson et al4426
Figure 9 Representation of the area included in a photograph as a geometric projectiononto the Earthrsquos surface Note that the focal length (the distance from the SpaceShuttle to the photo) is grossly overscale compared to the altitude (the distance fromthe Shuttle to the surface of the Earth
rotations and translation used to convert from the R-Earth to the R-Spacecraft arecomputed and documented in the spreadsheet
With the mathematical positions of the SC SN PC and the centre of the Earthcomputed in R-Spacecraft the program next computes the location of the camerarsquosprincipal point (PP) The principle point is the point of intersection of the opticalaxis of the lens with the image plane (ie the lm) It is nominally positioned at a
Astronaut photography as digital data for remote sensing 4427
distance equal to the focal length from the perspective centre of the camera (whichis assumed to be at SC) along the vector from PC through SC as shown in gure 9
A third coordinate system the Rectangular Photo Coordinate System (R-Photo)is created with its origin at PP its XndashY axial plane normal to the vector from PCthrough SC and its +X axis aligned with the +X axis of R-Spacecraft as shownin gure 9 The XndashY plane of this coordinate system represents the image plane ofthe photograph
534 Auxiliary point calculationsThe calculations above employ a critical assumption that all photos are taken
with the lsquotop of the imagersquo oriented perpendicular to the vector from the SN towardsPC as shown in gure 9 To avoid non-uniform solar heating of the external surfacemost orbiting spacecraft are slowly and continually rotated about one or more axesIn this condition a ight crew member taking a photo while oating in microgravitycould orient the photo with practically any orientation relative to the horizon (seealso gure 8(b)) Unfortunately since these photos are taken with conventionalhandheld cameras there is no other information available which can be used toresolve the photorsquos rotational ambiguity about the optical axis other then the photoitself This is why the above assumption is used and the footprint computed by thiscalculator is actually a lsquolocator ellipsersquo which estimates the area on the ground whatis likely to be included in a speci c photograph (see gure 6) This locator ellipse ismost accurate for square image formats and subject to additional distortion as thephotograph format becomes more rectangular
If the user wants a more precise calculation of the image footprint the photorsquosrotational ambiguity about the optical axis must be resolved This can be done inthe calculator by adding data for an auxiliary point Detailed instructions regardinghow to prepare and use auxiliary point data in the computations are included in thelsquoHow-To-Usersquo tab of the spreadsheet Basically the user determines which side ofthe photograph is top and then measures the angle between the line from PP to thetop of the photo and from PP to the auxiliary point on the photo ( gure 9)
If the user includes data for an auxiliary point a series of computations arecompleted to resolve the photo rotation ambiguity about the optical axis (ie the
+Z axis in R-Photo) A vector from the Auxiliary Point on the Earth (AE) throughthe photograph perspective centre ( located at SC) is intersected with the photo imageplane (XndashY plane of R-Photo) to compute the coordinates of the Auxiliary Point onthe photo (AP) in R-Photo A two-dimensional angle in the XndashY plane of R-Photofrom the shy X axis to a line from PP to AP is calculated as shown in gure 9 The
shy X axis is used as the origin of the angle since it represents the top of the photoonce it passes through the perspective centre The diVerence between the computedangle and the angle measured by the user on the photo resolves the ambiguity inthe rotation of the photo relative to the principal line ( gures 7 and 9) Thetransformations from R-Spacecraft and R-Photo are then modi ed to include anadditional rotation angle about the +Z axis in R-Photo
535 lsquoFootprint rsquo calculationsThe program next computes the coordinates of eight points about the perimeter
of the image format (ie located at the four photo corners plus a bisector pointalong each edge of the image) These points are identi ed in R-Photo based uponthe photograph format size and then converted to R-Spacecraft Since all computa-tions are done in orthogonal coordinate systems the R-Spacecraft to R-Photo
J A Robinson et al4428
rotation matrix is transposed to produce an R-Photo to R-Spacecraft rotation matrixOnce in R-Spacecraft a unit vector from each of the eight perimeter points throughthe photo perspective centre (the same point as SC) is computed This provides thecoordinates for points about the perimeter of the image format with their directionvectors in a common coordinate system with other key points needed to computethe photo footprint
The next step is to compute the point of intersection between the spherical earthmodel and each of the eight perimeter point vectors The scalar value for eachperimeter point unit vector is computed using two-dimensional planar trigonometryAn angle c is computed using the formula for the cosine of an angle between twoincluded vectors (the perimeter point unit vector and the vector from SC to thecentre of the Earth) Angle y is computed using the Law of Sines Angle e=180degrees shy c shy y The scalar value of the perimeter point vector is computed using eand the Law of Cosines The scalar value is then multiplied by the perimeter pointunit vector to produce the three-dimensional point of intersection of the vector withEarthrsquos surface in R-Spacecraft The process is repeated independently for each ofthe eight perimeter point vectors Aside from its mathematical simplicity the valueof arcsine (computed in step 2) will exceed the normal range for the sin of an anglewhen the perimeter point vectors fail to intersect with the surface of the earth Asimple test based on this principle allows the program to correctly handle obliquephotos which image a portion of the horizon (see results for high oblique photographsin table 5)
The nal step in the process converts the eight earth intersection points from theR-Spacecraft to R-Earth The results are then converted to the standard geographiccoordinate system and displayed on the lsquoInput-Outputrsquo page of the spreadsheet
54 ExamplesAll three formulations were applied to the photographs included in this paper
and the results are compared in table 5 Formulation 1 gives too small a value fordistance across the photograph (D) for all but the most nadir shots and thus servesas an indicator of the best theoretical case but is not a good measure for a speci cphotograph For example the photograph of Lake Eyre taken from 276 km altitude( gure 2(a)) and the photograph of Limmen Bight ( gure 7) were closest to beingnadir views (oVsetslt68 km or tlt15deg table 5) For these photographs D calculatedusing Formulation 1 was similar to the minimum D calculated using Formulations2 and 3 For almost all other more oblique photographs Formulation 1 gave asigni cant underestimate of the distance covered in the photograph For gure 5(A)(the picture of Houston taken with a 40 mm lens) Formulation 1 did not give anunderestimate for D This is because Formulation 1 does not account for curvatureof the Earth in any way With this large eld of view assuming a at Earth in atedthe value of D above the minimum from calculations that included Earth curvature
A major diVerence between Formulations 2 and 3 is the ability to estimate pixelsizes (P) in both directions (along the principal line and perpendicular to the principalline) For the more oblique photographs the vertical estimate of D and pixel sizes ismuch larger than in the horizontal direction (eg the low oblique and high obliquephotographs of Hawaii gure 4 table 5)
For the area of Limmen Bight ( gure 7) table 5 illustrates the improvement inthe estimate of distance and pixel size that can be obtained by re-estimating thelocation of the PC with greater accuracy Centre points in the catalogued data are
Astronaut photography as digital data for remote sensing 4429
plusmn05deg of latitude and longitude When the centre point was re-estimated to plusmn002degwe determined that the photograph was not taken as obliquely as rst thought(change in the estimate of the look angle t from 169 to 128deg table 5) When theauxiliary point was added to the calculations of Formula 3 the calculated look angleshrank further to 110deg indicating that this photograph was taken at very close toa nadir view Of course this improvement in accuracy could have also led to estimatesof greater obliquity and corresponding larger pixel sizes
We also tested the performance of the scale calculator with auxiliary point byestimating the corner and point locations on the photograph using a 11 000 000Operational Navigational Chart For this test we estimated our ability to readcoordinates from the map as plusmn002degand our error in nding the locations of thecorner points as plusmn015deg (this error varies among photographs depending on thedetail that can be matched between photo and map) For Limmen Bight ( gure 7)the mean diVerence between map estimates and calculator estimates for four pointswas 031deg (SD=018 n=8) For a photograph of San Francisco Bay (STS062-151-291) the mean diVerence between map estimates and calculator estimates forfour points was 0064deg (SD=018 n=8) For a photograph of San Francisco Bay(STS062-151-291) the mean diVerence between map estimates and calculator estim-ates for eight points was 0196deg (SD=0146 n=16) Thus in one case the calculatorestimates were better than our estimate of the error in locating corner points on themap It is reasonable to expect that for nadir-viewing photographs the calculatorused with an auxiliary point can estimate locations of the edges of a photograph towithin plusmn03deg
55 Empirical con rmation of spatial resolution estimatesAs stated previously a challenge to estimating system-AWAR for astronaut
photography of Earth is the lack of suitable targets Small features in an image cansometimes be used as a check on the size of objects that can be successfully resolvedgiving an approximate value for GRD Similarly the number of pixels that make upthose features in the digitized image can be used to make an independent calculationof pixel size We have successfully used features such as roads and airport runwaysto make estimates of spatial scale and resolution (eg Robinson et al 2000c) Whilerecognizing that the use of linear detail in an image is a poor approximation to abar target and that linear objects smaller than the resolving power can often bedetected (Charman 1965) few objects other than roads could be found to make anydirect estimates of GRD Thus roads and runways were used in the images ofHouston (where one can readily conduct ground veri cations and where a numberof higher-contrast concrete roadways were available) to obtain empirical estimatesof GRD and pixel size for comparison with table 5
In the all-digital ESC image of Houston ( gure 3(d )) we examined EllingtonField runway 4-22 (centre left of the image) which is 24384 mtimes457 m This runwayis approximately 6ndash7 pixels in width and 304ndash3094 pixels in length so pixelsrepresent an distance on the ground 7ndash8 m Using a lower contrast measure of astreet length between two intersections (2125 m=211 pixels) a pixel width of 101 mis estimated These results compare favourably with the minimum estimate of 81 mpixels using Formulation 1 (table 5) For an estimate of GRD that would be morecomparable to aerial photography of a line target the smallest street without treecover that could be clearly distinguished on the photograph was 792 m wide Thesmallest non-street object (a gap between stages of the Saturn rocket on display in
J A Robinson et al4430
Tab
le5
App
lica
tio
nof
met
hod
sfo
res
tim
ati
ng
spati
alre
solu
tio
no
fast
ron
aut
ph
oto
gra
ph
sin
clu
din
gF
orm
ula
tion
s12
and
3Im
pro
ved
cen
tre
po
int
esti
mat
esan
dauxilia
ryp
oin
tca
lcu
lati
ons
are
sho
wn
for
gure
7V
aria
ble
sas
use
din
the
text
fle
ns
foca
lle
ngt
hH
al
titu
de
PC
ph
oto
gra
ph
cen
tre
poin
tSN
sp
ace
craft
nadir
po
int
D
dis
tance
cover
edo
nth
egr
oun
d
Pd
ista
nce
on
the
grou
nd
repre
sen
ted
by
apix
el
OVse
td
ista
nce
bet
wee
nP
Cand
SNusi
ng
the
grea
tci
rcle
form
ula
t
look
angle
away
from
SN
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
2L
ake
Eyre
ST
S093
-717
-80
250
276
shy29
0137
0shy
28
413
71
607
11
767
413
7609
ndash64
212
013
760
9ndash64
7121
124
ST
S082
-754
-61
250
617
shy28
5137
0shy
28
113
50
135
726
1201
18
013
8ndash14
827
517
9138ndash15
3277
294
Fig
ure
3H
ou
sto
nS
TS
062
-97-
151
4029
829
5shy
95
530
5shy
95
4410
78
8112
20
5348ndash58
990
1205
351
ndash63
8951
10
36
ST
S094
-737
-28
100
294
295
shy
95
528
3shy
95
5162
31
1133
24
415
8ndash20
334
724
3158ndash20
7351
390
ST
S055
-71-
41250
302
295
shy
95
028
2shy
93
766
412
8192
32
5736
ndash84
715
232
273
9ndash85
7154
185
S73
E51
45300
2685
295
shy
95
024
78
0F
igu
re4
Haw
aii
ST
S069
-701
-65
250
370
195
shy
155
520
4shy
153
881
415
7204
28
9876
ndash98
917
928
688
0ndash10
9181
210
ST
S069
-701
-70
250
370
210
shy
157
519
2shy
150
781
415
7738
63
414
9ndash23
336
760
7148ndash55
6394
10
63
ST
S069
-729
-27
100
398
200
shy
156
620
5shy
148
2219
42
1867
65
432
8ndash13
10
158
62
4323+
311
N
A
Astronaut photography as digital data for remote sensing 4431
Tab
le5
(Cont
inued
)
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
7L
imm
enB
igh
tS
TS
093
-704
-50
250
283
shy15
0135
0shy
15
013
58
623
120
859
16
963
0ndash67
3125
16
863
0ndash67
612
613
2P
CR
e-es
tim
ated
shy14
733
1352
67
64
5128
623
ndash65
5123
128
623
ndash65
712
3127
Auxilia
ryP
oin
t611
062
3ndash65
112
312
5F
igu
re10
N
ear
Au
stin
T
XS
TS
095
-716
-46
250
543
30
5shy
975
28
6shy
1007
120
230
375
34
613
5ndash157
281
34
1136ndash16
128
636
0
1 All
pho
togr
aphs
exce
pt
on
eta
ken
wit
hH
ass
elbla
dca
mer
aso
dis
tan
ceacr
oss
the
image
d=
55m
m
for
S73
E51
45
taken
wit
hel
ectr
onic
still
cam
erad=
276
5m
mho
rizo
nta
l2 P
Cand
SN
are
giv
enin
the
form
(lati
tud
elo
ngit
ude)
deg
rees
w
ith
neg
ativ
en
um
ber
sre
pre
senti
ng
south
and
wes
tA
ccura
cyo
fP
Cis
plusmn0
5and
SN
isplusmn
01
as
reco
rded
inth
eA
stro
naut
Ph
oto
gra
phy
Data
base
ex
cep
tw
her
en
ote
dth
atce
ntr
ep
oin
tw
asre
-est
imate
dw
ith
grea
ter
accu
racy
3 C
alc
ula
ted
usi
ng
the
mea
nofth
em
inim
um
an
dm
axim
um
esti
mate
sof
DF
orm
ula
tion
2co
nsi
der
sD
inth
ed
irec
tion
per
pen
dic
ula
rto
the
Pri
nci
pal
Lin
eo
nly
4 C
alc
ula
ted
inho
rizo
nta
lan
dve
rtic
aldir
ecti
on
sb
ut
still
ass
um
ing
geom
etry
of
gure
6(B
)F
irst
nu
mb
eris
the
mea
no
fth
em
inim
um
Dat
the
bott
om
of
the
ph
oto
and
the
maxi
mum
Dat
the
top
of
the
ph
oto
(range
show
nin
the
tab
le)
and
isco
mpara
ble
toF
orm
ula
tio
n2
Sec
ond
num
ber
isth
em
ean
of
Des
tim
ated
alon
gth
eP
rin
cip
alline
and
alo
ng
the
ou
tsid
eed
ge
(range
not
show
n)
5 Alt
itu
de
isap
pro
xim
ate
for
mis
sion
as
exac
tti
me
pho
togr
ap
hw
asta
ken
and
corr
esp
on
din
gn
adir
data
are
no
tava
ilab
le
Lack
of
nad
irdata
pre
ven
tsap
plica
tion
of
Form
ula
tion
s2
an
d3
6 Au
xilliar
ypo
int
add
edw
as
shy14
80
lati
tud
e13
51
2lo
ngit
ude
shy46
5degro
tati
on
angle
in
stru
ctio
ns
for
esti
mati
ng
the
rota
tio
nan
gle
para
met
erare
conta
ined
as
par
tof
the
scal
eca
lcu
lato
rsp
read
shee
tdo
cum
enta
tio
n
J A Robinson et al4432
a park at Johnson Space Center) that could clearly be distinguished on the photo-graph was 853 m wide
For the photograph of Houston taken with a 250 mm lens ( gure 3(C)) anddigitized from second generation lm at 2400 ppi (106 mm pixel Otilde 1 ) Ellington Fieldrunway 4-22 is 3 pixels in width and 1613 pixels in length so pixels represent151ndash152 m on the ground These results compare favourably with the minimumestimate of 152 m using Formulation 2 and 154ndash185 m pixels using Formulation 3(table 5) For an estimate of GRD using an 8times8 inch print (1369 enlargement) and4times magni cation the smallest street that could clearly be distinguished was 822 mwide the same feature could barely be distinguished on the digitized image
We also made an empirical estimate of spatial resolution for lower contrastvegetation boundaries By clearing forest so that a pattern would be visible to landingaircraft a landowner outside Austin Texas (see also aerial photo in Lisheron 2000)created a target that is also useful for evaluating spatial resolution of astronautphotographs The forest was selectively cleared in order to spell the landownerrsquosname lsquoLUECKErsquo with the remaining trees ( gure 10) According to local surveyorswho planned the clearing the plan was to create letters that were 3100 fttimes1700 ft(9449 mtimes5182 m) Photographed at a high altitude relative to most Shuttle missions(543 km) with a 250 mm lens Formula 3 predicts that each pixel would represent anarea 286 mtimes360 m on the ground (table 5) When original lm was digitized at2400 ppi (106 mm pixel Otilde 1 ) letters correspond to 294times188 pixels for a comparablepixel size of 27ndash32 m
6 Summary and conclusionsAstronaut photographs can be an excellent source of data for remote sensing
applications Best-case resolutions are similar to that for Landsat or SPOT withpixels as small as lt10 m It was estimated that of images taken to date 504 hadlens and obliquity characteristics that make them potential candidates for remotesensing information Digitized astronaut photographs can be overlaid with othersatellite data using GIS (Eckardt et al 2000) or used to ll in gaps in time serieswhen other imagery is not available As a source of public-domain information theycan be very useful for scientists who do not have access to satellite imagery eitherbecause of the expense of image acquisition (to the end user) or the computersystems needed for processing satellite images These diVerences in image acquisitioncosts (to the user) are summarized in table 6
Searching of the complete database of NASA astronaut photography includinglow-spatial-resolution browse images is available via the Web (OYce of EarthSciences 2000) This provides nearly global access for identifying images that willcontribute to a speci c scienti c project For the most detailed studies digitalproducts posted to the web will not be of suYcient quality To date we have madeit a practice of digitizing small numbers of images when requested by scientists atno charge Such requests (including a description of the project involved) can bemade through the authors or using the contact information listed on the websiteInvestigators with needs for larger numbers of digital images have also been servedthrough collaborative agreements
In our experience scientists that nd the data most valuable are those in develop-ing countries those studying areas that have not usually been targets for the majorsatellites those having diYculty in nding low-cloud images those interested inconstructing time series or those interested in using a large number of images
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
Astronaut photography as digital data for remote sensing 4413
Tab
le2
Max
imu
mpo
ssib
led
igit
alsp
atia
lre
solu
tio
n(p
ixel
size
inm
)fo
rca
mer
asy
stem
scu
rren
tly
use
dby
ast
ron
auts
top
ho
togr
aph
eart
hb
ase
do
nth
ege
om
etry
of
alt
itud
ele
ns
and
lm
size
C
alc
ula
tion
sas
sum
eth
atco
lour
tran
spar
ency
lm
isdig
itiz
edat
2400
ppi
(pix
els
per
inch
10
6m
mpix
elOtilde
1 )
Min
imu
mgro
un
ddis
tance
repre
sente
dby
1p
ixel
(m)
As
afu
nct
ion
of
alti
tude1
Imag
esi
zeM
inim
um
alt
itud
eM
edia
nalt
itu
de
Maxi
mum
alt
itud
eC
amer
asy
stem
mm
timesm
mp
ixel
s2(2
22k
m)
(326
km
)(6
11
km
)
Lin
ho
f125
mm
90
mm
lens
105
times120
8978times
11
434
261
383
718
250
mm
lens
94
138
259
Has
selb
lad
70
mm
100
mm
lens
55times
55
5198
times519
823
534
564
725
0m
mle
ns
94
138
259
Nik
on
35m
m30
0m
mle
ns
36times
24
3308
times217
478
115
216
400
mm
lens
59
86
162
ESC
35m
m18
0m
mle
ns3
27
6times
18
530
69times
204
311
116
330
630
0m
mle
ns
67
98
184
400
mm
lens
50
74
138
1 Alt
itud
essh
ow
nar
em
inim
um
m
edia
nan
dm
axim
um
thro
ugh
ST
S-8
9(J
anuary
199
8N
=84
mis
sion
s)
2 Act
ual
pix
els
for
ES
C
oth
erw
ise
assu
med
dig
itiz
ati
on
at240
0pp
i(p
ixel
sp
erin
ch)
equ
alto
10
6m
mpix
elOtilde
1 3 T
he
mo
stco
mm
on
con
gura
tion
for
Eart
hph
oto
gra
ph
yas
part
of
the
Eart
hK
AM
pro
ject
J A Robinson et al4414
Figure 4 De nitions of look angle for photographs taken from low orbit around EarthThe example photographs of Hawaii were all taken from approximately the samealtitude (370ndash398 km) and with the same (250 mm) lens on a Hasselblad camera(STS069-701-69 STS069-701-70 and STS069-729-27 [100 mm lens])
astronaut photography due to light path properties of the lenses used A lack of atness of the lm at the moment of exposure (because cameras used in orbit donot incorporate a vacuum platen) also introduces slight degradations A numberof additional sources of image degradation that would aVect GRD of astronautphotographs are listed
41 Films and processing411 Colour reversal lms
Most lms used in NASA handheld cameras in orbit have been E-6 processcolour reversal lms (Ektachromes) that have a nominal speed of 64 to 100 ASAalthough many other lms have been own (table 3) Reversal lms are used becausedamage from radiation exposure while outside the atmospheric protection of Earthis more noticeable in negative lms than in positive lms (Slater 1996) The choiceof ASA has been to balance the coarser grains ( lower lm resolving power) of high-speed lms with the fact that slower lms require longer exposures and thus areaVected more by vehicle motion (approximately 73 km s Otilde 1 relative to Earthrsquos surfacefor the Space Shuttle) Extremely fast lms (gt400 ASA) are also more susceptibleto radiation damage in orbit (particularly during long-duration or high-altitudemissions Slater 1996) and have not traditionally been used for Earth photography The manufacturer stock numbers identifying the lm used are available for eachimage in the Astronaut Photography Database (OYce of Earth Sciences 2000)
412 Colour infrared lmsColour infrared lm (CIR) was used during an Earth-orbiting Apollo mission
(Colwell 1971) in the multispectral S-190A and the high-resolution S-190B camera
Astronaut photography as digital data for remote sensing 4415
Figure 5 (a) Distributions of astronaut photographs by look angle oV-nadir (calculated fromcentre point and nadir point using Formulation 2) Data included for all photographsfor which centre and nadir points are known with high oblique look angles (n=55 155) excluded (Total sample shown is 108 293 of 382 563 total records in thedatabase when compiled For records where nadir data was not available number ofobservations for qualitative look angles are near vertical 14 086 low oblique 115 834undetermined 89 195) (b) Altitude distribution for astronaut photographs of Earthtaken with the Hasselblad camera Data included for Hasselblad camera only focallength determined with high oblique look angles excluded (Total sample shown is159 046 of 206 971 Hasselblad records with known altitude and of 382 563 total recordsin the database when compiled For Hasselblad records with known altitude data notshown includes high oblique photographs using 100 mm and 250 mm lenses n=20 262and all look angles with other lenses n=27 663)
systems on Skylab (NASA 1974 Wilmarth et al 1977) and occasionally on Shuttlemissions and Shuttle-Mir missions (table 3) The CIR lm used is a three-layerAerochrome having one layer that is sensitive to re ected solar infrared energy toapproximately 900 nm (Eastman Kodak Company 1998) protective coatings onmost spacecraft windows also limit IR transmittance between 800 mm and 1000 nmThis layer is extremely sensitive to the temperature of the lm which createsunpredictable degradation of the IR signature under some space ight conditions
J A Robinson et al4416
Tab
le3
Det
ails
on
lm
suse
dfo
rast
ron
aut
pho
togr
aphy
of
Eart
h
Dat
ain
clud
ep
ho
togr
aphy
fro
mall
NA
SA
hum
an
spac
eig
ht
mis
sio
ns
thro
ugh
ST
S-9
6(J
un
e19
99
378
461
tota
lfr
ames
inth
edata
base
)F
ilm
sw
ith
lt50
00fr
am
esex
po
sed
and
lm
suse
dfo
r35
mm
cam
eras
only
are
no
tin
clud
ed
Res
olv
ing
Nu
mb
erof
Film
man
ufa
ctu
rer
AS
Ap
ow
er1
fram
esP
erio
dof
use
Cam
eras
Colo
ur
posi
tive
Kod
akA
eroch
rom
eII
(2447
2448
SO
-131S
O-3
68
)32
40ndash50
513
8196
8ndash19
85
Hass
elbla
dL
inh
of
Kod
akE
kta
chro
me
(5017
502
56
017
6118
SO
-121
64
50
113
932
196
5ndash19
68
Hass
elbla
dL
inh
of
Role
iex
SO
-217
Q
X-8
24
QX
-868
)19
81ndash1993
Mau
rer
Nik
on
Kod
akL
um
iere
(5046
5048
)100
105
743
199
4ndash19
98
Hass
elbla
dL
inho
fK
od
akE
lite
100
S(5
069
)100
41
486
199
6ndashp
rese
nt
Hass
elbla
d2
Fu
jiV
elvia
50C
S135
-36
32
13
882
1992ndash19
97H
ass
elbla
dN
iko
nK
od
akH
i-R
esolu
tio
nC
olo
r(S
O-2
42S
O-3
56
)10
096
74
1973ndash19
75H
ass
elbla
d
S19
0A
S190
B
Infr
are
dand
bla
ck-a
nd-w
hit
e
lms
Kod
akA
eroch
rom
eC
olo
rIn
frar
ed40
32
40
704
196
5ndashp
rese
nt2
Hass
elbla
dL
inh
of
Nik
on
S190
A
(8443
34
432443
)S
190B
Kod
akB
ampW
Infr
are
d(2
424
)32
10
553
197
3ndash19
74
S190
AK
od
akP
anato
mic
-XB
ampW
(SO
-022
)63
10
657
197
3ndash19
74
S190
A
1A
tlo
wco
ntr
ast
(ob
ject
back
grou
nd
rati
o16
1)
aspro
vid
edby
Ko
dak
and
list
edby
Sla
ter
etal
(1
983
256
)R
esolv
ing
pow
erw
as
dis
con
tin
ued
fro
mth
ete
chn
ical
test
sper
form
edb
yK
odak
inapp
roxim
atel
y19
93
an
dit
isn
ot
kn
ow
nif
curr
ent
lm
sh
ave
sim
ilar
reso
lvin
gpo
wer
too
lder
ver
sion
so
fsi
milar
lm
s(K
T
eite
lbaum
K
od
akp
erso
nal
com
mu
nic
atio
n)
Th
egra
nu
lari
tym
easu
reno
wpro
vid
edb
yK
od
ak
can
no
tb
ere
adily
con
ver
ted
toa
spat
ial
reso
luti
on
2 The
Lin
hof
was
use
dag
ain
on
ST
S-1
03in
Dec
emb
er19
99(d
ata
not
cata
logued
asof
the
pre
para
tion
of
this
tab
le)
usi
ng
Ko
dak
Elite
100S
lm
3 B
egin
nin
gin
July
1999
2443
colo
ur-
infr
are
d
lmpro
cess
was
chan
ged
from
EA
-5to
E-6
Astronaut photography as digital data for remote sensing 4417
413 Film duplicationThe original lm is archived permanently after producing about 20 duplicate
printing masters (second generation) Duplicate printing masters are disseminatedto regional archives and used to produce products (thirdndashfourth generation) for themedia the public and for scienti c use When prints slides transparencies anddigital products are produced for public distribution they are often colour adjustedto correspond to a more realistic look Digital products from second generationcopies can be requested by scienti c users (contact the authors for information onobtaining such products for a particular project) and these are recommended forremote sensing applications Care should be taken that the digital product acquiredfor remote sensing analysis has not been colour adjusted for presentation purposesBased on qualitative observations there is little visible increase in GRD in the thirdor fourth generation products There is a signi cant degradation in delity of colourin third and fourth generation duplicates because an increase in contrast occurswith every copy
42 Shutter speedsThe impact of camera motion (both due to the photographer and to the motion
of the spacecraft ) on image quality is determined by shutter speedmdash1250 to 1500second have been used because slower speeds record obvious blurring caused bythe rapid motion of the spacecraft relative to the surface of the Earth Generally1250 was used for ISO 64 lms (and slower) and after the switch to ISO 100 lmsa 1500 setting became standard New Hasselblad cameras that have been ownbeginning with STS-92 in October 2000 vary shutter speed using a focal plane shutterfor exposure bracketing (rather than varying aperture) and 1250 1500 and 11000second are used
Across an altitude range of 120ndash300 nautical miles (222ndash555 km) and orbitalinclinations of 285deg and 570deg the median relative ground velocity of orbitingspacecraft is 73 km s Otilde 1 At a nominal shutter speed of 1500 the expected blur dueto motion relative to the ground would be 146 m When cameras are handheld (asopposed to mounted in a bracket) blur can be reduced by physical compensationfor the motion (tracking) by the photographer many photographs show a level ofdetail indicating that blur at this level did not occur (eg gure 3(d )) Thus motionrelative to the ground is not an absolute barrier to ground resolution for handheldphotographs
43 Spacecraft windowsMost of the photographs in the NASA Astronaut Photography Database were
taken through window ports on the spacecraft used The transmittance of the windowport may be aVected by material inhomogeniety of the glass the number of layersof panes coatings used the quality of the surface polish environmentally inducedchanges in window materials (pressure loads thermal gradients) or deposited con-tamination Such degradation of the window cannot be corrected is diVerent foreach window and changes over time See Eppler et al (1996) and Scott (2000) fordiscussion of the spectral transmittance of the fused quartz window that is part ofthe US Laboratory Module (Destiny) of the International Space Station
44 Digitized images from lmWhen lm is scanned digitally the amount of information retained depends on
the spectral information extracted from the lm at the spatial limits of the scanner
J A Robinson et al4418
To date standard digitizing methodologies for astronaut photographs have not beenestablished and lm is digitized on a case-by-case basis using the equipment available
441 Digitizing and spatial resolutionLight (1993 1996) provided equations for determining the digitizing spatial
resolution needed to preserve spatial resolution of aerial photography based on thestatic resolving power (AWAR) for the system For lms where the manufacturer hasprovided data (Eastman Kodak 1998 K Teitelbaum personal communication) the resolving power of lms used for astronaut photography ranges from 32 lp mm Otilde 1to 100 lp mm Otilde 1 at low contrast (objectbackground ratio 161 table 3) The AWARfor the static case of the Hasselblad camera has been measured at high and lowcontrast (using lenses shown in table 1) with a maximum of approximately55 lp mm Otilde 1 (Fred Pearce unpublished data)
Based on the method of Light (1993 1996) the dimension of one spatial resolutionelement for a photograph with maximum static AWAR of 55 lp mm Otilde 1 would be18 mm lp Otilde 1 The acceptable range of spot size to preserve spatial information wouldthen be
18 mm
2 atilde 2aring scan spot size aring
18 mm
2(1)
and 6 mm aring scan spot size aring 9 mm Similarly for a more typical static AWAR of30 lp mm Otilde 1 (33 mm lpOtilde 1 low contrast Fred Pearce unpublished data) 11 mm aring scanspot sizearing 17 mm These scan spot sizes correspond to digitizing resolutions rangingfrom 4233ndash2822 ppi (pixels inch Otilde 1 ) for AWAR of 55 lp mm Otilde 1 and 2309ndash1494 ppi forAWAR of 30 lp mm Otilde 1 Scan spot sizes calculated for astronaut photography arecomparable to those calculated for the National Aerial Photography Program(9ndash13 mm Light 1996)
Widely available scanners that can digitize colour transparency lm currentlyhave a maximum spatial resolution of approximately 2400 ppi (106 mm pixel Otilde 1) Forexample we routinely use an Agfa Arcus II desktop scanner (see also Baltsavias1996) with 2400 ppi digitizing spatial resolution (2400 ppi optical resolution in onedirection and 1200 ppi interpolated to 2400 ppi in the other direction) and 2400 ppiwas used to calculate IFOV equivalents (tables 2 and 4) For some combinations oflens camera and contrast 2400 ppi will capture nearly all of the spatial informationcontained in the lm However for lm with higher resolving power thanEktachrome-64 for better lenses and for higher contrast targets digitizing at 2400 ppiwill not capture all of the spatial information in the lm
Improvements in digitizing technology will only produce real increases in IFOVto the limit of the AWAR of the photography system The incremental increase inspatial information above 3000 ppi (85 mm pixel Otilde 1 ) is not likely to outweigh thecosts of storing large images (Luman et al 1997) At 2400 ppi a Hasselblad frameis approximately 5200times5200 or 27 million pixels (table 2) while the same imagedigitized at 4000 ppi would contain 75 million pixels
Initial studies using astronaut photographs digitized at 2400 ppi (106 mm pixel Otilde 1 Webb et al 2000 Robinson et al 2000c) indicate that some GRD is lost comparedto photographic products Nevertheless the spatial resolution is still comparablewith other widely used data sources (Webb et al 2000) Robinson et al (2000c)found that digitizing at 21 mm pixel Otilde 1 provided information equivalent to 106 mmpixel Otilde 1 for identifying general urban area boundaries for six cities except for a single
Astronaut photography as digital data for remote sensing 4419
photo that required higher spatial resolution digitizing (it had been taken with ashorter lens and thus had less spatial resolution) Part of the equivalence observedin the urban areas study may be attributable to the fact that the atbed scannerused interpolates from 1200 ppi to 2400 ppi in one direction The appropriate digitiz-ing spatial resolution will in part depend on the scale of features of interest to theresearchers with a maximum upper limit set of a scan spot size of approximately6 mm (4233 ppi)
442 Digitizing and spectral resolutionWhen colour lm is digitized there will be loss of spectral resolution The three
lm emulsion layers (red green blue) have relatively distinct spectral responses butare fused together so that digitizers detect one colour signal and must convert itback into three (red green blue) digital channels Digital scanning is also subject tospectral calibration and reproduction errors Studies using digital astronaut photo-graphs to date have used 8 bits per channel However this is another parameterthat can be controlled when the lm is digitized We do not further address spectralresolution of digitally scanned images in this paper
45 External factors that in uence GRDAlthough not discussed in detail in this paper factors external to the spacecraft
and camera system (as listed by Moik (1980) for remote sensing in general) alsoimpact GRD These include atmospheric interference (due to scattering attenuationhaze) variable surface illumination (diVerences in terrain slope and orientation) andchange of terrain re ectance with viewing angle (bidirectional re ectance) For astro-naut photographs variable illumination is particularly important because orbits arenot sun-synchronous Photographs are illuminated by diVerent sun angles and imagesof a given location will have colour intensities that vary widely In addition theviewing angle has an eVect on the degree of object-to-backgroun d contrast andatmospheric interference
5 Estimating spatial resolution of astronaut photographsIn order to use astronaut photographs for digital remote sensing it is important
to be able to calculate the equivalent to an IFOVmdashthe ground area represented bya single pixel in a digitized orbital photograph The obliquity of most photographsmeans that pixel lsquosizesrsquo vary at diVerent places in an image Given a ground distanceD represented by a photograph in each direction (horizontal vertical ) an approxi-mate average pixel width (P the equivalent of IFOV) for the entire image can becalculated as follows
P=D
dS(2)
where D is the projected distance on the ground covered by the image in the samedirection as the pixel is measured d is the actual width of the image on the original lm (table 1) and S is the digitizing spatial resolution
Here we present three mathematical formulations for estimating the size of thefootprint or area on the ground covered by the image Example results from theapplication of all three formulations are given in table 5 The rst and simplestcalculation (formulation 1) gives an idea of the maximum spatial resolution attainableat a given altitude of orbit with a given lm format and a perfectly vertical (nadir)
J A Robinson et al4420
view downward Formulation 2 takes into account obliquity by calculating lookangle from the diVerence between the location on the ground represented at thecentre of the photograph and the nadir location of the spacecraft at the time thephotograph was taken ( gure 1) Formulation 3 describes an alternate solution tothe oblique look angle problem using coordinate-system transformations Thisformulation has been implemented in a documented spreadsheet and is availablefor download (httpeoljscnasagovsseopFootprintCalculatorxls) and is in theprocess of being implemented on a Web-based user interface to the AstronautPhotography Database (OYce of Earth Sciences 2000)
Although formulations 2 and 3 account for obliquity for the purposes of calcula-tion they treat the position of the spacecraft and position of the camera as one Inactuality astronauts are generally holding the cameras by hand (although camerasbracketed in the window are also used) and the selection of window position of theastronaut in the window and rotation of the camera relative to the movement ofthe spacecraft are not known Thus calculations using only the photo centre point(PC) and spacecraft nadir point (SN) give a locator ellipse and not the locations ofthe corners of the photograph A locator ellipse describes an estimated area on theground that is likely to be included in a speci c photograph regardless of the rotationof the lm plane about the camerarsquos optical axis ( gure 6)
Estimating the corner positions of the photo requires additional user input of asingle auxiliary pointmdasha location on the image that has a known location on theground Addition of this auxiliary point is an option available to users of thespreadsheet An example of the results of adding an auxiliary point is shown in gure 7 with comparisons of the various calculations in table 5
51 Formulation 1 Footprint for a nadir viewThe simplest way to estimate footprint size is to use the geometry of camera lens
and spacecraft altitude to calculate the scaling relationship between the image in the
Figure 6 Sketch representation of the locator ellipse an estimated area on the ground thatis likely to be included in a speci c photograph regardless of the rotation of the lmplane about the camerarsquos optical axis
Astronaut photography as digital data for remote sensing 4421
Figure 7 An example of the application of the Low Oblique Space Photo FootprintCalculator The photograph (STS093-704-50) of Limmen Bight Australia was takenfrom an altitude of 283 km using the Hasselblad camera with a 250 mm lens Mapused is 11 000 000 Operational Navigational Chart The distances shown equate toan average pixel size of 124 m for this photograph
lm and the area covered on the ground For a perfect nadir view the scalerelationship from the geometry of similar triangles is
d
D=
f
H(3)
where d=original image size D=distance of footprint on the ground f =focallength of lens and H=altitude Once D is known pixel size ( length=width) can becalculated from equation (2) These calculations represent the minimum footprintand minimum pixel size possible for a given camera system altitude and digitisingspatial resolution (table 2) Formulation 1 was used for calculating minimum pixelsizes shown in tables 2 and 4
By assuming digitizing at 2400 ppi (106 mm pixel Otilde 1 ) currently a spatial resolutioncommonly attainable from multipurpose colour transparency scanners (see sect441)this formulation was used to convert area covered to an IFOV equivalent for missionsof diVerent altitudes (table 2) Table 4 provides a comparison of IFOV of imagesfrom various satellites including the equivalent for astronaut photography Valuesin this table were derived by using formulation 1 to estimate the area covered becausea perfect nadir view represents the best possible spatial resolution and smallest eldof view that could be obtained
52 Formulation 2 Footprint for oblique views using simpli ed geometry and thegreat circle distance
A more realistic approach to determining the footprint of the photographaccounts for the fact that the camera is not usually pointing perfectly down at the
J A Robinson et al4422
Table 4 Comparison of pixel sizes (instantaneous elds of view) for satellite remote sensorswith pixel sizes for near vertical viewing photography from spacecraft Note that alldata returned from the automated satellites have the same eld of view while indicatedpixel sizes for astronaut photography are best-case for near vertical look angles See gure 5 for distributions of astronaut photographs of various look angles High-altitude aerial photography is also included for reference
Altitude PixelInstrument (km) width (mtimesm) Bands1
Landsat Multipectral Scanner2 (MSS 1974-) 880ndash940 79times56 4ndash5Landsat Thematic Mapper (TM 1982-) ~705 30times30 7Satellite pour lrsquoObservation de la Terre (SPOT 1986-) ~832
SPOT 1-3 Panchromatic 10times10 1SPOT 1-3 Multispectral (XS) 20times20 3
Astronaut Photography (1961-)Skylab3 (1973-1974 ) ~435
Multispectral Camera (6 camera stations) 17ndash19times17ndash19 3ndash8Earth Terrain Camera (hi-res colour) 875times875 3
Space Shuttle5 (1981-) 222ndash611Hasselblad 70 mm (250mm lens colour) vertical view 9ndash26times9ndash26 3Hasselblad 70 mm (100mm lens colour) vertical view 23ndash65times23ndash65 3Nikon 35 mm (400mm lens colour lm or ESC) vertical 5ndash16times5ndash16 3
High Altitude Aerial Photograph (1100 000) ~12 2times2 3
1Three bands listed for aerial photography obtainable by high-resolution colour digitizingFor high-resolution colour lm at 65 lp mmOtilde 1 (K Teitelbaum Eastman Kodak Co personalcommunication) the maximum information contained would be 3302 ppi
2Data from Simonett (1983Table 1-2)3Calculated as recommended by Jensen (19831596) the dividend of estimated ground
resolution at low contrast given in NASA (1974179-180) and 244Six lters plus colour and colour infrared lm (NASA 19747) 5Extracted from table 2
nadir point The look angle (the angle oV nadir that the camera is pointing) can becalculated trigonometrically by assuming a spherical earth and calculating the dis-tance between the coordinates of SN and PC ( gure 1) using the Great Circledistance haversine solution (Sinnott 1984 Snyder 198730ndash32 Chamberlain 1996)The diVerence between the spacecraft centre and nadir point latitudes Dlat=lat 2 shy lat 1 and the diVerence between the spacecraft centre and nadir pointlongitudes Dlon=lon 2 shy lon 1 enter the following equations
a=sin2ADlat
2 B+cos(lat 1) cos(lat 2) sin2ADlon
2 B (4)
c=2 arcsin(min[1 atilde a]) (5)
and oVset=R c with R the radius of a spherical Earth=6370 kmThe look angle (t ) is then given by
t=arctanAoffset
H B (6)
Assuming that the camera was positioned so that the imaginary line between thecentre and nadir points (the principal line) runs vertically through the centre of thephotograph the distance between the geometric centre of the photograph (principalpoint PP) and the top of the photograph is d2 ( gure 8(a)) The scale at any point
Astronaut photography as digital data for remote sensing 4423
(a) (b)
Figure 8 The geometric relationship between look angle lens focal length and the centrepoint and nadir points of a photograph (see also Estes and Simonett 1975) Note thatthe Earthrsquos surface and footprint represented in gure 1 are not shown here only arepresentation of the image area of the photograph Variables shown in the gurelook angle t focal length f PP centre of the image on the lm SN spacecraft nadird original image size (a) The special case where the camera is aligned squarely withthe isometric parallel In this case tilt occurs along a plane parallel with the centre ofthe photograph When the conditions are met calculations using Formulation 3 willgive the ground coordinates of the corner points of the photograph (b) The generalrelationship between these parameters and key photogrammetric points on the photo-graph For this more typical case coordinates of an additional point in the photographmust be input in order to adjust for twist of the camera and to nd the corner pointcoordinates
in the photograph varies as a function of the distance y along the principal linebetween the isocentre and the point according to the relationship
d
D=
f shy ysint
H(7)
(Wong 1980 equation (214) H amp the ground elevation h) Using gure 8(a) at thetop of the photo
y= f tanAt
2B+d
2(8)
and at the bottom of the photo
y= f tanAt
2Bshyd
2(9)
Thus for given PC and SN coordinates and assuming a photo orientation as in gure 8(a) and not gure 8(b) we can estimate a minimum D (using equations (7)and (8)) and a maximum D (using equations (7) and (9)) and then average the twoto determine the pixel size (P ) via equation (2)
J A Robinson et al4424
53 Formulation 3 T he L ow Oblique Space Photo Footprint CalculatorThe Low Oblique Space Photo Footprint Calculator was developed to provide
a more accurate estimation of the geographic coordinates for the footprint of a lowoblique photo of the Earthrsquos surface taken from a human-occupied spacecraft inorbit The calculator performs a series of three-dimensional coordinate transforma-tions to compute the location and orientation of the centre of the photo exposureplane relative to an earth referenced coordinate system The nominal camera focallength is then used to create a vector from the photorsquos perspective point througheach of eight points around the parameter of the image as de ned by the formatsize The geographical coordinates for the photo footprint are then computed byintersecting these photo vectors with a spherical earth model Although more sophist-icated projection algorithms are available no signi cant increase in the accuracy ofthe results would be produced by these algorithms due to inherent uncertainties inthe available input data (ie the spacecraft altitude photo centre location etc)
The calculations were initially implemented within a Microsoft Excel workbookwhich allowed one to embed the mathematical and graphical documentation nextto the actual calculations Thus interested users can inspect the mathematical pro-cessing A set of error traps was also built into the calculations to detect erroneousresults A summary of any errors generated is reported to the user with the resultsof the calculations Interested individuals are invited to download the Excel work-book from httpeoljscnasagovsseopFootprintCalculatorxls The calculations arecurrently being encoded in a high-level programming language and should soon beavailable alongside other background data provided for each photograph at OYceof Earth Sciences (2000)
For the purposes of these calculations a low oblique photograph was de ned as
one with the centre within 10 degrees of latitude and longitude of the spacecraftnadir point For the typical range of spacecraft altitudes to date this restricted thecalculations to photographs in which Earthrsquos horizon does not appear (the generalde nition of a low oblique photograph eg Campbell 199671 and gure 4)
531 Input data and resultsUpon opening the Excel workbook the user is presented with a program intro-
duction providing instructions for using the calculator The second worksheet tablsquoHow-To-Usersquo provides detailed step-by-step instructions for preparing the baselinedata for the program The third tab lsquoInput-Output rsquo contains the user input eldsand displays the results of the calculations The additional worksheets contain theactual calculations and program documentation Although users are welcome toreview these sheets an experienced user need only access the lsquoInput-Outputrsquo
spreadsheetTo begin a calculation the user enters the following information which is available
for each photo in the NASA Astronaut Photography Database (1) SN geographicalcoordinates of spacecraft nadir position at the time of photo (2) H spacecraftaltitude (3) PC the geographical coordinates of the centre of the photo (4) f nominal focal length and (5) d image format size The automatic implementationof the workbook on the web will automatically enter these values and completecalculations
For more accurate results the user may optionally enter the geographic co-ordinates and orientation for an auxiliary point on the photo which resolves thecamerarsquos rotation uncertainty about the optical axis The auxiliary point data must
Astronaut photography as digital data for remote sensing 4425
be computed by the user following the instruction contained in the lsquoHow-To-Usersquotab of the spreadsheet
After entering the input data the geographic coordinates of the photo footprint(ie four photo corner points four points at the bisector of each edge and the centreof the photo) are immediately displayed below the input elds along with any errormessages generated by the user input or by the calculations ( gure 7) Although
results are computed and displayed they should not be used when error messagesare produced by the program The program also computes the tilt angle for each ofthe photo vectors relative to the spacecraft nadir vector To the right of the photofootprint coordinates is displayed the arc distance along the surface of the sphere
between adjacent computed points
532 Calculation assumptions
The mathematical calculations implemented in the Low Oblique Space PhotoFootprint Calculator use the following assumptions
1 The SN location is used as exact even though the true value may vary by upto plusmn01 degree from the location provided with the photo
2 The spacecraft altitude is used as exact Although the determination of the
nadir point at the instant of a known spacecraft vector is relatively precise(plusmn115times10 Otilde 4 degrees) the propagator interpolates between sets of approxi-mately 10ndash40 known vectors per day and the time code recorded on the lmcan drift Thus the true value for SN may vary by up to plusmn01 degree fromthe value provided with the photo
3 The perspective centre of the camera is assumed to be at the given altitudeover the speci ed spacecraft nadir location at the time of photo exposure
4 The PC location is used as exact even though the true value may vary by upto plusmn05deg latitude and plusmn05deg longitude from the location provided with the
photo5 A spherical earth model is used with a nominal radius of 6 372 16154 m (a
common rst-order approximation for a spherical earth used in geodeticcomputations)
6 The nominal lens focal length of the camera lens is used in the computations(calibrated focal length values are not available)
7 The photo projection is based on the classic pin-hole camera model8 No correction for lens distortion or atmospheric re ection is made
9 If no auxiliary point data is provided the lsquoTop of the Imagersquo is orientedperpendicular to the vector from SN towards PC
533 T ransformation from Earth to photo coordinate systemsThe calculations begin by converting the geographic coordinates ( latitude and
longitude) of the SN and PC to a Rectangular Earth-Centred Coordinate System(R-Earth) de ned as shown in gure 9 (with the centre of the Earth at (0 0 0))
Using the vector from the Earthrsquos centre through SN and the spacecraft altitude thespacecraft location (SC) is also computed in R-Earth
For ease of computation a Rectangular Spacecraft-Centred Coordinate System(R-Spacecraft ) is de ned as shown in gure 9 The origin of R-Spacecraft is located
at SC with its +Z-axis aligned with vector from the centre of the Earth throughSN and its +X-axis aligned with the vector from SN to PC ( gure 9) The speci c
J A Robinson et al4426
Figure 9 Representation of the area included in a photograph as a geometric projectiononto the Earthrsquos surface Note that the focal length (the distance from the SpaceShuttle to the photo) is grossly overscale compared to the altitude (the distance fromthe Shuttle to the surface of the Earth
rotations and translation used to convert from the R-Earth to the R-Spacecraft arecomputed and documented in the spreadsheet
With the mathematical positions of the SC SN PC and the centre of the Earthcomputed in R-Spacecraft the program next computes the location of the camerarsquosprincipal point (PP) The principle point is the point of intersection of the opticalaxis of the lens with the image plane (ie the lm) It is nominally positioned at a
Astronaut photography as digital data for remote sensing 4427
distance equal to the focal length from the perspective centre of the camera (whichis assumed to be at SC) along the vector from PC through SC as shown in gure 9
A third coordinate system the Rectangular Photo Coordinate System (R-Photo)is created with its origin at PP its XndashY axial plane normal to the vector from PCthrough SC and its +X axis aligned with the +X axis of R-Spacecraft as shownin gure 9 The XndashY plane of this coordinate system represents the image plane ofthe photograph
534 Auxiliary point calculationsThe calculations above employ a critical assumption that all photos are taken
with the lsquotop of the imagersquo oriented perpendicular to the vector from the SN towardsPC as shown in gure 9 To avoid non-uniform solar heating of the external surfacemost orbiting spacecraft are slowly and continually rotated about one or more axesIn this condition a ight crew member taking a photo while oating in microgravitycould orient the photo with practically any orientation relative to the horizon (seealso gure 8(b)) Unfortunately since these photos are taken with conventionalhandheld cameras there is no other information available which can be used toresolve the photorsquos rotational ambiguity about the optical axis other then the photoitself This is why the above assumption is used and the footprint computed by thiscalculator is actually a lsquolocator ellipsersquo which estimates the area on the ground whatis likely to be included in a speci c photograph (see gure 6) This locator ellipse ismost accurate for square image formats and subject to additional distortion as thephotograph format becomes more rectangular
If the user wants a more precise calculation of the image footprint the photorsquosrotational ambiguity about the optical axis must be resolved This can be done inthe calculator by adding data for an auxiliary point Detailed instructions regardinghow to prepare and use auxiliary point data in the computations are included in thelsquoHow-To-Usersquo tab of the spreadsheet Basically the user determines which side ofthe photograph is top and then measures the angle between the line from PP to thetop of the photo and from PP to the auxiliary point on the photo ( gure 9)
If the user includes data for an auxiliary point a series of computations arecompleted to resolve the photo rotation ambiguity about the optical axis (ie the
+Z axis in R-Photo) A vector from the Auxiliary Point on the Earth (AE) throughthe photograph perspective centre ( located at SC) is intersected with the photo imageplane (XndashY plane of R-Photo) to compute the coordinates of the Auxiliary Point onthe photo (AP) in R-Photo A two-dimensional angle in the XndashY plane of R-Photofrom the shy X axis to a line from PP to AP is calculated as shown in gure 9 The
shy X axis is used as the origin of the angle since it represents the top of the photoonce it passes through the perspective centre The diVerence between the computedangle and the angle measured by the user on the photo resolves the ambiguity inthe rotation of the photo relative to the principal line ( gures 7 and 9) Thetransformations from R-Spacecraft and R-Photo are then modi ed to include anadditional rotation angle about the +Z axis in R-Photo
535 lsquoFootprint rsquo calculationsThe program next computes the coordinates of eight points about the perimeter
of the image format (ie located at the four photo corners plus a bisector pointalong each edge of the image) These points are identi ed in R-Photo based uponthe photograph format size and then converted to R-Spacecraft Since all computa-tions are done in orthogonal coordinate systems the R-Spacecraft to R-Photo
J A Robinson et al4428
rotation matrix is transposed to produce an R-Photo to R-Spacecraft rotation matrixOnce in R-Spacecraft a unit vector from each of the eight perimeter points throughthe photo perspective centre (the same point as SC) is computed This provides thecoordinates for points about the perimeter of the image format with their directionvectors in a common coordinate system with other key points needed to computethe photo footprint
The next step is to compute the point of intersection between the spherical earthmodel and each of the eight perimeter point vectors The scalar value for eachperimeter point unit vector is computed using two-dimensional planar trigonometryAn angle c is computed using the formula for the cosine of an angle between twoincluded vectors (the perimeter point unit vector and the vector from SC to thecentre of the Earth) Angle y is computed using the Law of Sines Angle e=180degrees shy c shy y The scalar value of the perimeter point vector is computed using eand the Law of Cosines The scalar value is then multiplied by the perimeter pointunit vector to produce the three-dimensional point of intersection of the vector withEarthrsquos surface in R-Spacecraft The process is repeated independently for each ofthe eight perimeter point vectors Aside from its mathematical simplicity the valueof arcsine (computed in step 2) will exceed the normal range for the sin of an anglewhen the perimeter point vectors fail to intersect with the surface of the earth Asimple test based on this principle allows the program to correctly handle obliquephotos which image a portion of the horizon (see results for high oblique photographsin table 5)
The nal step in the process converts the eight earth intersection points from theR-Spacecraft to R-Earth The results are then converted to the standard geographiccoordinate system and displayed on the lsquoInput-Outputrsquo page of the spreadsheet
54 ExamplesAll three formulations were applied to the photographs included in this paper
and the results are compared in table 5 Formulation 1 gives too small a value fordistance across the photograph (D) for all but the most nadir shots and thus servesas an indicator of the best theoretical case but is not a good measure for a speci cphotograph For example the photograph of Lake Eyre taken from 276 km altitude( gure 2(a)) and the photograph of Limmen Bight ( gure 7) were closest to beingnadir views (oVsetslt68 km or tlt15deg table 5) For these photographs D calculatedusing Formulation 1 was similar to the minimum D calculated using Formulations2 and 3 For almost all other more oblique photographs Formulation 1 gave asigni cant underestimate of the distance covered in the photograph For gure 5(A)(the picture of Houston taken with a 40 mm lens) Formulation 1 did not give anunderestimate for D This is because Formulation 1 does not account for curvatureof the Earth in any way With this large eld of view assuming a at Earth in atedthe value of D above the minimum from calculations that included Earth curvature
A major diVerence between Formulations 2 and 3 is the ability to estimate pixelsizes (P) in both directions (along the principal line and perpendicular to the principalline) For the more oblique photographs the vertical estimate of D and pixel sizes ismuch larger than in the horizontal direction (eg the low oblique and high obliquephotographs of Hawaii gure 4 table 5)
For the area of Limmen Bight ( gure 7) table 5 illustrates the improvement inthe estimate of distance and pixel size that can be obtained by re-estimating thelocation of the PC with greater accuracy Centre points in the catalogued data are
Astronaut photography as digital data for remote sensing 4429
plusmn05deg of latitude and longitude When the centre point was re-estimated to plusmn002degwe determined that the photograph was not taken as obliquely as rst thought(change in the estimate of the look angle t from 169 to 128deg table 5) When theauxiliary point was added to the calculations of Formula 3 the calculated look angleshrank further to 110deg indicating that this photograph was taken at very close toa nadir view Of course this improvement in accuracy could have also led to estimatesof greater obliquity and corresponding larger pixel sizes
We also tested the performance of the scale calculator with auxiliary point byestimating the corner and point locations on the photograph using a 11 000 000Operational Navigational Chart For this test we estimated our ability to readcoordinates from the map as plusmn002degand our error in nding the locations of thecorner points as plusmn015deg (this error varies among photographs depending on thedetail that can be matched between photo and map) For Limmen Bight ( gure 7)the mean diVerence between map estimates and calculator estimates for four pointswas 031deg (SD=018 n=8) For a photograph of San Francisco Bay (STS062-151-291) the mean diVerence between map estimates and calculator estimates forfour points was 0064deg (SD=018 n=8) For a photograph of San Francisco Bay(STS062-151-291) the mean diVerence between map estimates and calculator estim-ates for eight points was 0196deg (SD=0146 n=16) Thus in one case the calculatorestimates were better than our estimate of the error in locating corner points on themap It is reasonable to expect that for nadir-viewing photographs the calculatorused with an auxiliary point can estimate locations of the edges of a photograph towithin plusmn03deg
55 Empirical con rmation of spatial resolution estimatesAs stated previously a challenge to estimating system-AWAR for astronaut
photography of Earth is the lack of suitable targets Small features in an image cansometimes be used as a check on the size of objects that can be successfully resolvedgiving an approximate value for GRD Similarly the number of pixels that make upthose features in the digitized image can be used to make an independent calculationof pixel size We have successfully used features such as roads and airport runwaysto make estimates of spatial scale and resolution (eg Robinson et al 2000c) Whilerecognizing that the use of linear detail in an image is a poor approximation to abar target and that linear objects smaller than the resolving power can often bedetected (Charman 1965) few objects other than roads could be found to make anydirect estimates of GRD Thus roads and runways were used in the images ofHouston (where one can readily conduct ground veri cations and where a numberof higher-contrast concrete roadways were available) to obtain empirical estimatesof GRD and pixel size for comparison with table 5
In the all-digital ESC image of Houston ( gure 3(d )) we examined EllingtonField runway 4-22 (centre left of the image) which is 24384 mtimes457 m This runwayis approximately 6ndash7 pixels in width and 304ndash3094 pixels in length so pixelsrepresent an distance on the ground 7ndash8 m Using a lower contrast measure of astreet length between two intersections (2125 m=211 pixels) a pixel width of 101 mis estimated These results compare favourably with the minimum estimate of 81 mpixels using Formulation 1 (table 5) For an estimate of GRD that would be morecomparable to aerial photography of a line target the smallest street without treecover that could be clearly distinguished on the photograph was 792 m wide Thesmallest non-street object (a gap between stages of the Saturn rocket on display in
J A Robinson et al4430
Tab
le5
App
lica
tio
nof
met
hod
sfo
res
tim
ati
ng
spati
alre
solu
tio
no
fast
ron
aut
ph
oto
gra
ph
sin
clu
din
gF
orm
ula
tion
s12
and
3Im
pro
ved
cen
tre
po
int
esti
mat
esan
dauxilia
ryp
oin
tca
lcu
lati
ons
are
sho
wn
for
gure
7V
aria
ble
sas
use
din
the
text
fle
ns
foca
lle
ngt
hH
al
titu
de
PC
ph
oto
gra
ph
cen
tre
poin
tSN
sp
ace
craft
nadir
po
int
D
dis
tance
cover
edo
nth
egr
oun
d
Pd
ista
nce
on
the
grou
nd
repre
sen
ted
by
apix
el
OVse
td
ista
nce
bet
wee
nP
Cand
SNusi
ng
the
grea
tci
rcle
form
ula
t
look
angle
away
from
SN
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
2L
ake
Eyre
ST
S093
-717
-80
250
276
shy29
0137
0shy
28
413
71
607
11
767
413
7609
ndash64
212
013
760
9ndash64
7121
124
ST
S082
-754
-61
250
617
shy28
5137
0shy
28
113
50
135
726
1201
18
013
8ndash14
827
517
9138ndash15
3277
294
Fig
ure
3H
ou
sto
nS
TS
062
-97-
151
4029
829
5shy
95
530
5shy
95
4410
78
8112
20
5348ndash58
990
1205
351
ndash63
8951
10
36
ST
S094
-737
-28
100
294
295
shy
95
528
3shy
95
5162
31
1133
24
415
8ndash20
334
724
3158ndash20
7351
390
ST
S055
-71-
41250
302
295
shy
95
028
2shy
93
766
412
8192
32
5736
ndash84
715
232
273
9ndash85
7154
185
S73
E51
45300
2685
295
shy
95
024
78
0F
igu
re4
Haw
aii
ST
S069
-701
-65
250
370
195
shy
155
520
4shy
153
881
415
7204
28
9876
ndash98
917
928
688
0ndash10
9181
210
ST
S069
-701
-70
250
370
210
shy
157
519
2shy
150
781
415
7738
63
414
9ndash23
336
760
7148ndash55
6394
10
63
ST
S069
-729
-27
100
398
200
shy
156
620
5shy
148
2219
42
1867
65
432
8ndash13
10
158
62
4323+
311
N
A
Astronaut photography as digital data for remote sensing 4431
Tab
le5
(Cont
inued
)
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
7L
imm
enB
igh
tS
TS
093
-704
-50
250
283
shy15
0135
0shy
15
013
58
623
120
859
16
963
0ndash67
3125
16
863
0ndash67
612
613
2P
CR
e-es
tim
ated
shy14
733
1352
67
64
5128
623
ndash65
5123
128
623
ndash65
712
3127
Auxilia
ryP
oin
t611
062
3ndash65
112
312
5F
igu
re10
N
ear
Au
stin
T
XS
TS
095
-716
-46
250
543
30
5shy
975
28
6shy
1007
120
230
375
34
613
5ndash157
281
34
1136ndash16
128
636
0
1 All
pho
togr
aphs
exce
pt
on
eta
ken
wit
hH
ass
elbla
dca
mer
aso
dis
tan
ceacr
oss
the
image
d=
55m
m
for
S73
E51
45
taken
wit
hel
ectr
onic
still
cam
erad=
276
5m
mho
rizo
nta
l2 P
Cand
SN
are
giv
enin
the
form
(lati
tud
elo
ngit
ude)
deg
rees
w
ith
neg
ativ
en
um
ber
sre
pre
senti
ng
south
and
wes
tA
ccura
cyo
fP
Cis
plusmn0
5and
SN
isplusmn
01
as
reco
rded
inth
eA
stro
naut
Ph
oto
gra
phy
Data
base
ex
cep
tw
her
en
ote
dth
atce
ntr
ep
oin
tw
asre
-est
imate
dw
ith
grea
ter
accu
racy
3 C
alc
ula
ted
usi
ng
the
mea
nofth
em
inim
um
an
dm
axim
um
esti
mate
sof
DF
orm
ula
tion
2co
nsi
der
sD
inth
ed
irec
tion
per
pen
dic
ula
rto
the
Pri
nci
pal
Lin
eo
nly
4 C
alc
ula
ted
inho
rizo
nta
lan
dve
rtic
aldir
ecti
on
sb
ut
still
ass
um
ing
geom
etry
of
gure
6(B
)F
irst
nu
mb
eris
the
mea
no
fth
em
inim
um
Dat
the
bott
om
of
the
ph
oto
and
the
maxi
mum
Dat
the
top
of
the
ph
oto
(range
show
nin
the
tab
le)
and
isco
mpara
ble
toF
orm
ula
tio
n2
Sec
ond
num
ber
isth
em
ean
of
Des
tim
ated
alon
gth
eP
rin
cip
alline
and
alo
ng
the
ou
tsid
eed
ge
(range
not
show
n)
5 Alt
itu
de
isap
pro
xim
ate
for
mis
sion
as
exac
tti
me
pho
togr
ap
hw
asta
ken
and
corr
esp
on
din
gn
adir
data
are
no
tava
ilab
le
Lack
of
nad
irdata
pre
ven
tsap
plica
tion
of
Form
ula
tion
s2
an
d3
6 Au
xilliar
ypo
int
add
edw
as
shy14
80
lati
tud
e13
51
2lo
ngit
ude
shy46
5degro
tati
on
angle
in
stru
ctio
ns
for
esti
mati
ng
the
rota
tio
nan
gle
para
met
erare
conta
ined
as
par
tof
the
scal
eca
lcu
lato
rsp
read
shee
tdo
cum
enta
tio
n
J A Robinson et al4432
a park at Johnson Space Center) that could clearly be distinguished on the photo-graph was 853 m wide
For the photograph of Houston taken with a 250 mm lens ( gure 3(C)) anddigitized from second generation lm at 2400 ppi (106 mm pixel Otilde 1 ) Ellington Fieldrunway 4-22 is 3 pixels in width and 1613 pixels in length so pixels represent151ndash152 m on the ground These results compare favourably with the minimumestimate of 152 m using Formulation 2 and 154ndash185 m pixels using Formulation 3(table 5) For an estimate of GRD using an 8times8 inch print (1369 enlargement) and4times magni cation the smallest street that could clearly be distinguished was 822 mwide the same feature could barely be distinguished on the digitized image
We also made an empirical estimate of spatial resolution for lower contrastvegetation boundaries By clearing forest so that a pattern would be visible to landingaircraft a landowner outside Austin Texas (see also aerial photo in Lisheron 2000)created a target that is also useful for evaluating spatial resolution of astronautphotographs The forest was selectively cleared in order to spell the landownerrsquosname lsquoLUECKErsquo with the remaining trees ( gure 10) According to local surveyorswho planned the clearing the plan was to create letters that were 3100 fttimes1700 ft(9449 mtimes5182 m) Photographed at a high altitude relative to most Shuttle missions(543 km) with a 250 mm lens Formula 3 predicts that each pixel would represent anarea 286 mtimes360 m on the ground (table 5) When original lm was digitized at2400 ppi (106 mm pixel Otilde 1 ) letters correspond to 294times188 pixels for a comparablepixel size of 27ndash32 m
6 Summary and conclusionsAstronaut photographs can be an excellent source of data for remote sensing
applications Best-case resolutions are similar to that for Landsat or SPOT withpixels as small as lt10 m It was estimated that of images taken to date 504 hadlens and obliquity characteristics that make them potential candidates for remotesensing information Digitized astronaut photographs can be overlaid with othersatellite data using GIS (Eckardt et al 2000) or used to ll in gaps in time serieswhen other imagery is not available As a source of public-domain information theycan be very useful for scientists who do not have access to satellite imagery eitherbecause of the expense of image acquisition (to the end user) or the computersystems needed for processing satellite images These diVerences in image acquisitioncosts (to the user) are summarized in table 6
Searching of the complete database of NASA astronaut photography includinglow-spatial-resolution browse images is available via the Web (OYce of EarthSciences 2000) This provides nearly global access for identifying images that willcontribute to a speci c scienti c project For the most detailed studies digitalproducts posted to the web will not be of suYcient quality To date we have madeit a practice of digitizing small numbers of images when requested by scientists atno charge Such requests (including a description of the project involved) can bemade through the authors or using the contact information listed on the websiteInvestigators with needs for larger numbers of digital images have also been servedthrough collaborative agreements
In our experience scientists that nd the data most valuable are those in develop-ing countries those studying areas that have not usually been targets for the majorsatellites those having diYculty in nding low-cloud images those interested inconstructing time series or those interested in using a large number of images
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
J A Robinson et al4414
Figure 4 De nitions of look angle for photographs taken from low orbit around EarthThe example photographs of Hawaii were all taken from approximately the samealtitude (370ndash398 km) and with the same (250 mm) lens on a Hasselblad camera(STS069-701-69 STS069-701-70 and STS069-729-27 [100 mm lens])
astronaut photography due to light path properties of the lenses used A lack of atness of the lm at the moment of exposure (because cameras used in orbit donot incorporate a vacuum platen) also introduces slight degradations A numberof additional sources of image degradation that would aVect GRD of astronautphotographs are listed
41 Films and processing411 Colour reversal lms
Most lms used in NASA handheld cameras in orbit have been E-6 processcolour reversal lms (Ektachromes) that have a nominal speed of 64 to 100 ASAalthough many other lms have been own (table 3) Reversal lms are used becausedamage from radiation exposure while outside the atmospheric protection of Earthis more noticeable in negative lms than in positive lms (Slater 1996) The choiceof ASA has been to balance the coarser grains ( lower lm resolving power) of high-speed lms with the fact that slower lms require longer exposures and thus areaVected more by vehicle motion (approximately 73 km s Otilde 1 relative to Earthrsquos surfacefor the Space Shuttle) Extremely fast lms (gt400 ASA) are also more susceptibleto radiation damage in orbit (particularly during long-duration or high-altitudemissions Slater 1996) and have not traditionally been used for Earth photography The manufacturer stock numbers identifying the lm used are available for eachimage in the Astronaut Photography Database (OYce of Earth Sciences 2000)
412 Colour infrared lmsColour infrared lm (CIR) was used during an Earth-orbiting Apollo mission
(Colwell 1971) in the multispectral S-190A and the high-resolution S-190B camera
Astronaut photography as digital data for remote sensing 4415
Figure 5 (a) Distributions of astronaut photographs by look angle oV-nadir (calculated fromcentre point and nadir point using Formulation 2) Data included for all photographsfor which centre and nadir points are known with high oblique look angles (n=55 155) excluded (Total sample shown is 108 293 of 382 563 total records in thedatabase when compiled For records where nadir data was not available number ofobservations for qualitative look angles are near vertical 14 086 low oblique 115 834undetermined 89 195) (b) Altitude distribution for astronaut photographs of Earthtaken with the Hasselblad camera Data included for Hasselblad camera only focallength determined with high oblique look angles excluded (Total sample shown is159 046 of 206 971 Hasselblad records with known altitude and of 382 563 total recordsin the database when compiled For Hasselblad records with known altitude data notshown includes high oblique photographs using 100 mm and 250 mm lenses n=20 262and all look angles with other lenses n=27 663)
systems on Skylab (NASA 1974 Wilmarth et al 1977) and occasionally on Shuttlemissions and Shuttle-Mir missions (table 3) The CIR lm used is a three-layerAerochrome having one layer that is sensitive to re ected solar infrared energy toapproximately 900 nm (Eastman Kodak Company 1998) protective coatings onmost spacecraft windows also limit IR transmittance between 800 mm and 1000 nmThis layer is extremely sensitive to the temperature of the lm which createsunpredictable degradation of the IR signature under some space ight conditions
J A Robinson et al4416
Tab
le3
Det
ails
on
lm
suse
dfo
rast
ron
aut
pho
togr
aphy
of
Eart
h
Dat
ain
clud
ep
ho
togr
aphy
fro
mall
NA
SA
hum
an
spac
eig
ht
mis
sio
ns
thro
ugh
ST
S-9
6(J
un
e19
99
378
461
tota
lfr
ames
inth
edata
base
)F
ilm
sw
ith
lt50
00fr
am
esex
po
sed
and
lm
suse
dfo
r35
mm
cam
eras
only
are
no
tin
clud
ed
Res
olv
ing
Nu
mb
erof
Film
man
ufa
ctu
rer
AS
Ap
ow
er1
fram
esP
erio
dof
use
Cam
eras
Colo
ur
posi
tive
Kod
akA
eroch
rom
eII
(2447
2448
SO
-131S
O-3
68
)32
40ndash50
513
8196
8ndash19
85
Hass
elbla
dL
inh
of
Kod
akE
kta
chro
me
(5017
502
56
017
6118
SO
-121
64
50
113
932
196
5ndash19
68
Hass
elbla
dL
inh
of
Role
iex
SO
-217
Q
X-8
24
QX
-868
)19
81ndash1993
Mau
rer
Nik
on
Kod
akL
um
iere
(5046
5048
)100
105
743
199
4ndash19
98
Hass
elbla
dL
inho
fK
od
akE
lite
100
S(5
069
)100
41
486
199
6ndashp
rese
nt
Hass
elbla
d2
Fu
jiV
elvia
50C
S135
-36
32
13
882
1992ndash19
97H
ass
elbla
dN
iko
nK
od
akH
i-R
esolu
tio
nC
olo
r(S
O-2
42S
O-3
56
)10
096
74
1973ndash19
75H
ass
elbla
d
S19
0A
S190
B
Infr
are
dand
bla
ck-a
nd-w
hit
e
lms
Kod
akA
eroch
rom
eC
olo
rIn
frar
ed40
32
40
704
196
5ndashp
rese
nt2
Hass
elbla
dL
inh
of
Nik
on
S190
A
(8443
34
432443
)S
190B
Kod
akB
ampW
Infr
are
d(2
424
)32
10
553
197
3ndash19
74
S190
AK
od
akP
anato
mic
-XB
ampW
(SO
-022
)63
10
657
197
3ndash19
74
S190
A
1A
tlo
wco
ntr
ast
(ob
ject
back
grou
nd
rati
o16
1)
aspro
vid
edby
Ko
dak
and
list
edby
Sla
ter
etal
(1
983
256
)R
esolv
ing
pow
erw
as
dis
con
tin
ued
fro
mth
ete
chn
ical
test
sper
form
edb
yK
odak
inapp
roxim
atel
y19
93
an
dit
isn
ot
kn
ow
nif
curr
ent
lm
sh
ave
sim
ilar
reso
lvin
gpo
wer
too
lder
ver
sion
so
fsi
milar
lm
s(K
T
eite
lbaum
K
od
akp
erso
nal
com
mu
nic
atio
n)
Th
egra
nu
lari
tym
easu
reno
wpro
vid
edb
yK
od
ak
can
no
tb
ere
adily
con
ver
ted
toa
spat
ial
reso
luti
on
2 The
Lin
hof
was
use
dag
ain
on
ST
S-1
03in
Dec
emb
er19
99(d
ata
not
cata
logued
asof
the
pre
para
tion
of
this
tab
le)
usi
ng
Ko
dak
Elite
100S
lm
3 B
egin
nin
gin
July
1999
2443
colo
ur-
infr
are
d
lmpro
cess
was
chan
ged
from
EA
-5to
E-6
Astronaut photography as digital data for remote sensing 4417
413 Film duplicationThe original lm is archived permanently after producing about 20 duplicate
printing masters (second generation) Duplicate printing masters are disseminatedto regional archives and used to produce products (thirdndashfourth generation) for themedia the public and for scienti c use When prints slides transparencies anddigital products are produced for public distribution they are often colour adjustedto correspond to a more realistic look Digital products from second generationcopies can be requested by scienti c users (contact the authors for information onobtaining such products for a particular project) and these are recommended forremote sensing applications Care should be taken that the digital product acquiredfor remote sensing analysis has not been colour adjusted for presentation purposesBased on qualitative observations there is little visible increase in GRD in the thirdor fourth generation products There is a signi cant degradation in delity of colourin third and fourth generation duplicates because an increase in contrast occurswith every copy
42 Shutter speedsThe impact of camera motion (both due to the photographer and to the motion
of the spacecraft ) on image quality is determined by shutter speedmdash1250 to 1500second have been used because slower speeds record obvious blurring caused bythe rapid motion of the spacecraft relative to the surface of the Earth Generally1250 was used for ISO 64 lms (and slower) and after the switch to ISO 100 lmsa 1500 setting became standard New Hasselblad cameras that have been ownbeginning with STS-92 in October 2000 vary shutter speed using a focal plane shutterfor exposure bracketing (rather than varying aperture) and 1250 1500 and 11000second are used
Across an altitude range of 120ndash300 nautical miles (222ndash555 km) and orbitalinclinations of 285deg and 570deg the median relative ground velocity of orbitingspacecraft is 73 km s Otilde 1 At a nominal shutter speed of 1500 the expected blur dueto motion relative to the ground would be 146 m When cameras are handheld (asopposed to mounted in a bracket) blur can be reduced by physical compensationfor the motion (tracking) by the photographer many photographs show a level ofdetail indicating that blur at this level did not occur (eg gure 3(d )) Thus motionrelative to the ground is not an absolute barrier to ground resolution for handheldphotographs
43 Spacecraft windowsMost of the photographs in the NASA Astronaut Photography Database were
taken through window ports on the spacecraft used The transmittance of the windowport may be aVected by material inhomogeniety of the glass the number of layersof panes coatings used the quality of the surface polish environmentally inducedchanges in window materials (pressure loads thermal gradients) or deposited con-tamination Such degradation of the window cannot be corrected is diVerent foreach window and changes over time See Eppler et al (1996) and Scott (2000) fordiscussion of the spectral transmittance of the fused quartz window that is part ofthe US Laboratory Module (Destiny) of the International Space Station
44 Digitized images from lmWhen lm is scanned digitally the amount of information retained depends on
the spectral information extracted from the lm at the spatial limits of the scanner
J A Robinson et al4418
To date standard digitizing methodologies for astronaut photographs have not beenestablished and lm is digitized on a case-by-case basis using the equipment available
441 Digitizing and spatial resolutionLight (1993 1996) provided equations for determining the digitizing spatial
resolution needed to preserve spatial resolution of aerial photography based on thestatic resolving power (AWAR) for the system For lms where the manufacturer hasprovided data (Eastman Kodak 1998 K Teitelbaum personal communication) the resolving power of lms used for astronaut photography ranges from 32 lp mm Otilde 1to 100 lp mm Otilde 1 at low contrast (objectbackground ratio 161 table 3) The AWARfor the static case of the Hasselblad camera has been measured at high and lowcontrast (using lenses shown in table 1) with a maximum of approximately55 lp mm Otilde 1 (Fred Pearce unpublished data)
Based on the method of Light (1993 1996) the dimension of one spatial resolutionelement for a photograph with maximum static AWAR of 55 lp mm Otilde 1 would be18 mm lp Otilde 1 The acceptable range of spot size to preserve spatial information wouldthen be
18 mm
2 atilde 2aring scan spot size aring
18 mm
2(1)
and 6 mm aring scan spot size aring 9 mm Similarly for a more typical static AWAR of30 lp mm Otilde 1 (33 mm lpOtilde 1 low contrast Fred Pearce unpublished data) 11 mm aring scanspot sizearing 17 mm These scan spot sizes correspond to digitizing resolutions rangingfrom 4233ndash2822 ppi (pixels inch Otilde 1 ) for AWAR of 55 lp mm Otilde 1 and 2309ndash1494 ppi forAWAR of 30 lp mm Otilde 1 Scan spot sizes calculated for astronaut photography arecomparable to those calculated for the National Aerial Photography Program(9ndash13 mm Light 1996)
Widely available scanners that can digitize colour transparency lm currentlyhave a maximum spatial resolution of approximately 2400 ppi (106 mm pixel Otilde 1) Forexample we routinely use an Agfa Arcus II desktop scanner (see also Baltsavias1996) with 2400 ppi digitizing spatial resolution (2400 ppi optical resolution in onedirection and 1200 ppi interpolated to 2400 ppi in the other direction) and 2400 ppiwas used to calculate IFOV equivalents (tables 2 and 4) For some combinations oflens camera and contrast 2400 ppi will capture nearly all of the spatial informationcontained in the lm However for lm with higher resolving power thanEktachrome-64 for better lenses and for higher contrast targets digitizing at 2400 ppiwill not capture all of the spatial information in the lm
Improvements in digitizing technology will only produce real increases in IFOVto the limit of the AWAR of the photography system The incremental increase inspatial information above 3000 ppi (85 mm pixel Otilde 1 ) is not likely to outweigh thecosts of storing large images (Luman et al 1997) At 2400 ppi a Hasselblad frameis approximately 5200times5200 or 27 million pixels (table 2) while the same imagedigitized at 4000 ppi would contain 75 million pixels
Initial studies using astronaut photographs digitized at 2400 ppi (106 mm pixel Otilde 1 Webb et al 2000 Robinson et al 2000c) indicate that some GRD is lost comparedto photographic products Nevertheless the spatial resolution is still comparablewith other widely used data sources (Webb et al 2000) Robinson et al (2000c)found that digitizing at 21 mm pixel Otilde 1 provided information equivalent to 106 mmpixel Otilde 1 for identifying general urban area boundaries for six cities except for a single
Astronaut photography as digital data for remote sensing 4419
photo that required higher spatial resolution digitizing (it had been taken with ashorter lens and thus had less spatial resolution) Part of the equivalence observedin the urban areas study may be attributable to the fact that the atbed scannerused interpolates from 1200 ppi to 2400 ppi in one direction The appropriate digitiz-ing spatial resolution will in part depend on the scale of features of interest to theresearchers with a maximum upper limit set of a scan spot size of approximately6 mm (4233 ppi)
442 Digitizing and spectral resolutionWhen colour lm is digitized there will be loss of spectral resolution The three
lm emulsion layers (red green blue) have relatively distinct spectral responses butare fused together so that digitizers detect one colour signal and must convert itback into three (red green blue) digital channels Digital scanning is also subject tospectral calibration and reproduction errors Studies using digital astronaut photo-graphs to date have used 8 bits per channel However this is another parameterthat can be controlled when the lm is digitized We do not further address spectralresolution of digitally scanned images in this paper
45 External factors that in uence GRDAlthough not discussed in detail in this paper factors external to the spacecraft
and camera system (as listed by Moik (1980) for remote sensing in general) alsoimpact GRD These include atmospheric interference (due to scattering attenuationhaze) variable surface illumination (diVerences in terrain slope and orientation) andchange of terrain re ectance with viewing angle (bidirectional re ectance) For astro-naut photographs variable illumination is particularly important because orbits arenot sun-synchronous Photographs are illuminated by diVerent sun angles and imagesof a given location will have colour intensities that vary widely In addition theviewing angle has an eVect on the degree of object-to-backgroun d contrast andatmospheric interference
5 Estimating spatial resolution of astronaut photographsIn order to use astronaut photographs for digital remote sensing it is important
to be able to calculate the equivalent to an IFOVmdashthe ground area represented bya single pixel in a digitized orbital photograph The obliquity of most photographsmeans that pixel lsquosizesrsquo vary at diVerent places in an image Given a ground distanceD represented by a photograph in each direction (horizontal vertical ) an approxi-mate average pixel width (P the equivalent of IFOV) for the entire image can becalculated as follows
P=D
dS(2)
where D is the projected distance on the ground covered by the image in the samedirection as the pixel is measured d is the actual width of the image on the original lm (table 1) and S is the digitizing spatial resolution
Here we present three mathematical formulations for estimating the size of thefootprint or area on the ground covered by the image Example results from theapplication of all three formulations are given in table 5 The rst and simplestcalculation (formulation 1) gives an idea of the maximum spatial resolution attainableat a given altitude of orbit with a given lm format and a perfectly vertical (nadir)
J A Robinson et al4420
view downward Formulation 2 takes into account obliquity by calculating lookangle from the diVerence between the location on the ground represented at thecentre of the photograph and the nadir location of the spacecraft at the time thephotograph was taken ( gure 1) Formulation 3 describes an alternate solution tothe oblique look angle problem using coordinate-system transformations Thisformulation has been implemented in a documented spreadsheet and is availablefor download (httpeoljscnasagovsseopFootprintCalculatorxls) and is in theprocess of being implemented on a Web-based user interface to the AstronautPhotography Database (OYce of Earth Sciences 2000)
Although formulations 2 and 3 account for obliquity for the purposes of calcula-tion they treat the position of the spacecraft and position of the camera as one Inactuality astronauts are generally holding the cameras by hand (although camerasbracketed in the window are also used) and the selection of window position of theastronaut in the window and rotation of the camera relative to the movement ofthe spacecraft are not known Thus calculations using only the photo centre point(PC) and spacecraft nadir point (SN) give a locator ellipse and not the locations ofthe corners of the photograph A locator ellipse describes an estimated area on theground that is likely to be included in a speci c photograph regardless of the rotationof the lm plane about the camerarsquos optical axis ( gure 6)
Estimating the corner positions of the photo requires additional user input of asingle auxiliary pointmdasha location on the image that has a known location on theground Addition of this auxiliary point is an option available to users of thespreadsheet An example of the results of adding an auxiliary point is shown in gure 7 with comparisons of the various calculations in table 5
51 Formulation 1 Footprint for a nadir viewThe simplest way to estimate footprint size is to use the geometry of camera lens
and spacecraft altitude to calculate the scaling relationship between the image in the
Figure 6 Sketch representation of the locator ellipse an estimated area on the ground thatis likely to be included in a speci c photograph regardless of the rotation of the lmplane about the camerarsquos optical axis
Astronaut photography as digital data for remote sensing 4421
Figure 7 An example of the application of the Low Oblique Space Photo FootprintCalculator The photograph (STS093-704-50) of Limmen Bight Australia was takenfrom an altitude of 283 km using the Hasselblad camera with a 250 mm lens Mapused is 11 000 000 Operational Navigational Chart The distances shown equate toan average pixel size of 124 m for this photograph
lm and the area covered on the ground For a perfect nadir view the scalerelationship from the geometry of similar triangles is
d
D=
f
H(3)
where d=original image size D=distance of footprint on the ground f =focallength of lens and H=altitude Once D is known pixel size ( length=width) can becalculated from equation (2) These calculations represent the minimum footprintand minimum pixel size possible for a given camera system altitude and digitisingspatial resolution (table 2) Formulation 1 was used for calculating minimum pixelsizes shown in tables 2 and 4
By assuming digitizing at 2400 ppi (106 mm pixel Otilde 1 ) currently a spatial resolutioncommonly attainable from multipurpose colour transparency scanners (see sect441)this formulation was used to convert area covered to an IFOV equivalent for missionsof diVerent altitudes (table 2) Table 4 provides a comparison of IFOV of imagesfrom various satellites including the equivalent for astronaut photography Valuesin this table were derived by using formulation 1 to estimate the area covered becausea perfect nadir view represents the best possible spatial resolution and smallest eldof view that could be obtained
52 Formulation 2 Footprint for oblique views using simpli ed geometry and thegreat circle distance
A more realistic approach to determining the footprint of the photographaccounts for the fact that the camera is not usually pointing perfectly down at the
J A Robinson et al4422
Table 4 Comparison of pixel sizes (instantaneous elds of view) for satellite remote sensorswith pixel sizes for near vertical viewing photography from spacecraft Note that alldata returned from the automated satellites have the same eld of view while indicatedpixel sizes for astronaut photography are best-case for near vertical look angles See gure 5 for distributions of astronaut photographs of various look angles High-altitude aerial photography is also included for reference
Altitude PixelInstrument (km) width (mtimesm) Bands1
Landsat Multipectral Scanner2 (MSS 1974-) 880ndash940 79times56 4ndash5Landsat Thematic Mapper (TM 1982-) ~705 30times30 7Satellite pour lrsquoObservation de la Terre (SPOT 1986-) ~832
SPOT 1-3 Panchromatic 10times10 1SPOT 1-3 Multispectral (XS) 20times20 3
Astronaut Photography (1961-)Skylab3 (1973-1974 ) ~435
Multispectral Camera (6 camera stations) 17ndash19times17ndash19 3ndash8Earth Terrain Camera (hi-res colour) 875times875 3
Space Shuttle5 (1981-) 222ndash611Hasselblad 70 mm (250mm lens colour) vertical view 9ndash26times9ndash26 3Hasselblad 70 mm (100mm lens colour) vertical view 23ndash65times23ndash65 3Nikon 35 mm (400mm lens colour lm or ESC) vertical 5ndash16times5ndash16 3
High Altitude Aerial Photograph (1100 000) ~12 2times2 3
1Three bands listed for aerial photography obtainable by high-resolution colour digitizingFor high-resolution colour lm at 65 lp mmOtilde 1 (K Teitelbaum Eastman Kodak Co personalcommunication) the maximum information contained would be 3302 ppi
2Data from Simonett (1983Table 1-2)3Calculated as recommended by Jensen (19831596) the dividend of estimated ground
resolution at low contrast given in NASA (1974179-180) and 244Six lters plus colour and colour infrared lm (NASA 19747) 5Extracted from table 2
nadir point The look angle (the angle oV nadir that the camera is pointing) can becalculated trigonometrically by assuming a spherical earth and calculating the dis-tance between the coordinates of SN and PC ( gure 1) using the Great Circledistance haversine solution (Sinnott 1984 Snyder 198730ndash32 Chamberlain 1996)The diVerence between the spacecraft centre and nadir point latitudes Dlat=lat 2 shy lat 1 and the diVerence between the spacecraft centre and nadir pointlongitudes Dlon=lon 2 shy lon 1 enter the following equations
a=sin2ADlat
2 B+cos(lat 1) cos(lat 2) sin2ADlon
2 B (4)
c=2 arcsin(min[1 atilde a]) (5)
and oVset=R c with R the radius of a spherical Earth=6370 kmThe look angle (t ) is then given by
t=arctanAoffset
H B (6)
Assuming that the camera was positioned so that the imaginary line between thecentre and nadir points (the principal line) runs vertically through the centre of thephotograph the distance between the geometric centre of the photograph (principalpoint PP) and the top of the photograph is d2 ( gure 8(a)) The scale at any point
Astronaut photography as digital data for remote sensing 4423
(a) (b)
Figure 8 The geometric relationship between look angle lens focal length and the centrepoint and nadir points of a photograph (see also Estes and Simonett 1975) Note thatthe Earthrsquos surface and footprint represented in gure 1 are not shown here only arepresentation of the image area of the photograph Variables shown in the gurelook angle t focal length f PP centre of the image on the lm SN spacecraft nadird original image size (a) The special case where the camera is aligned squarely withthe isometric parallel In this case tilt occurs along a plane parallel with the centre ofthe photograph When the conditions are met calculations using Formulation 3 willgive the ground coordinates of the corner points of the photograph (b) The generalrelationship between these parameters and key photogrammetric points on the photo-graph For this more typical case coordinates of an additional point in the photographmust be input in order to adjust for twist of the camera and to nd the corner pointcoordinates
in the photograph varies as a function of the distance y along the principal linebetween the isocentre and the point according to the relationship
d
D=
f shy ysint
H(7)
(Wong 1980 equation (214) H amp the ground elevation h) Using gure 8(a) at thetop of the photo
y= f tanAt
2B+d
2(8)
and at the bottom of the photo
y= f tanAt
2Bshyd
2(9)
Thus for given PC and SN coordinates and assuming a photo orientation as in gure 8(a) and not gure 8(b) we can estimate a minimum D (using equations (7)and (8)) and a maximum D (using equations (7) and (9)) and then average the twoto determine the pixel size (P ) via equation (2)
J A Robinson et al4424
53 Formulation 3 T he L ow Oblique Space Photo Footprint CalculatorThe Low Oblique Space Photo Footprint Calculator was developed to provide
a more accurate estimation of the geographic coordinates for the footprint of a lowoblique photo of the Earthrsquos surface taken from a human-occupied spacecraft inorbit The calculator performs a series of three-dimensional coordinate transforma-tions to compute the location and orientation of the centre of the photo exposureplane relative to an earth referenced coordinate system The nominal camera focallength is then used to create a vector from the photorsquos perspective point througheach of eight points around the parameter of the image as de ned by the formatsize The geographical coordinates for the photo footprint are then computed byintersecting these photo vectors with a spherical earth model Although more sophist-icated projection algorithms are available no signi cant increase in the accuracy ofthe results would be produced by these algorithms due to inherent uncertainties inthe available input data (ie the spacecraft altitude photo centre location etc)
The calculations were initially implemented within a Microsoft Excel workbookwhich allowed one to embed the mathematical and graphical documentation nextto the actual calculations Thus interested users can inspect the mathematical pro-cessing A set of error traps was also built into the calculations to detect erroneousresults A summary of any errors generated is reported to the user with the resultsof the calculations Interested individuals are invited to download the Excel work-book from httpeoljscnasagovsseopFootprintCalculatorxls The calculations arecurrently being encoded in a high-level programming language and should soon beavailable alongside other background data provided for each photograph at OYceof Earth Sciences (2000)
For the purposes of these calculations a low oblique photograph was de ned as
one with the centre within 10 degrees of latitude and longitude of the spacecraftnadir point For the typical range of spacecraft altitudes to date this restricted thecalculations to photographs in which Earthrsquos horizon does not appear (the generalde nition of a low oblique photograph eg Campbell 199671 and gure 4)
531 Input data and resultsUpon opening the Excel workbook the user is presented with a program intro-
duction providing instructions for using the calculator The second worksheet tablsquoHow-To-Usersquo provides detailed step-by-step instructions for preparing the baselinedata for the program The third tab lsquoInput-Output rsquo contains the user input eldsand displays the results of the calculations The additional worksheets contain theactual calculations and program documentation Although users are welcome toreview these sheets an experienced user need only access the lsquoInput-Outputrsquo
spreadsheetTo begin a calculation the user enters the following information which is available
for each photo in the NASA Astronaut Photography Database (1) SN geographicalcoordinates of spacecraft nadir position at the time of photo (2) H spacecraftaltitude (3) PC the geographical coordinates of the centre of the photo (4) f nominal focal length and (5) d image format size The automatic implementationof the workbook on the web will automatically enter these values and completecalculations
For more accurate results the user may optionally enter the geographic co-ordinates and orientation for an auxiliary point on the photo which resolves thecamerarsquos rotation uncertainty about the optical axis The auxiliary point data must
Astronaut photography as digital data for remote sensing 4425
be computed by the user following the instruction contained in the lsquoHow-To-Usersquotab of the spreadsheet
After entering the input data the geographic coordinates of the photo footprint(ie four photo corner points four points at the bisector of each edge and the centreof the photo) are immediately displayed below the input elds along with any errormessages generated by the user input or by the calculations ( gure 7) Although
results are computed and displayed they should not be used when error messagesare produced by the program The program also computes the tilt angle for each ofthe photo vectors relative to the spacecraft nadir vector To the right of the photofootprint coordinates is displayed the arc distance along the surface of the sphere
between adjacent computed points
532 Calculation assumptions
The mathematical calculations implemented in the Low Oblique Space PhotoFootprint Calculator use the following assumptions
1 The SN location is used as exact even though the true value may vary by upto plusmn01 degree from the location provided with the photo
2 The spacecraft altitude is used as exact Although the determination of the
nadir point at the instant of a known spacecraft vector is relatively precise(plusmn115times10 Otilde 4 degrees) the propagator interpolates between sets of approxi-mately 10ndash40 known vectors per day and the time code recorded on the lmcan drift Thus the true value for SN may vary by up to plusmn01 degree fromthe value provided with the photo
3 The perspective centre of the camera is assumed to be at the given altitudeover the speci ed spacecraft nadir location at the time of photo exposure
4 The PC location is used as exact even though the true value may vary by upto plusmn05deg latitude and plusmn05deg longitude from the location provided with the
photo5 A spherical earth model is used with a nominal radius of 6 372 16154 m (a
common rst-order approximation for a spherical earth used in geodeticcomputations)
6 The nominal lens focal length of the camera lens is used in the computations(calibrated focal length values are not available)
7 The photo projection is based on the classic pin-hole camera model8 No correction for lens distortion or atmospheric re ection is made
9 If no auxiliary point data is provided the lsquoTop of the Imagersquo is orientedperpendicular to the vector from SN towards PC
533 T ransformation from Earth to photo coordinate systemsThe calculations begin by converting the geographic coordinates ( latitude and
longitude) of the SN and PC to a Rectangular Earth-Centred Coordinate System(R-Earth) de ned as shown in gure 9 (with the centre of the Earth at (0 0 0))
Using the vector from the Earthrsquos centre through SN and the spacecraft altitude thespacecraft location (SC) is also computed in R-Earth
For ease of computation a Rectangular Spacecraft-Centred Coordinate System(R-Spacecraft ) is de ned as shown in gure 9 The origin of R-Spacecraft is located
at SC with its +Z-axis aligned with vector from the centre of the Earth throughSN and its +X-axis aligned with the vector from SN to PC ( gure 9) The speci c
J A Robinson et al4426
Figure 9 Representation of the area included in a photograph as a geometric projectiononto the Earthrsquos surface Note that the focal length (the distance from the SpaceShuttle to the photo) is grossly overscale compared to the altitude (the distance fromthe Shuttle to the surface of the Earth
rotations and translation used to convert from the R-Earth to the R-Spacecraft arecomputed and documented in the spreadsheet
With the mathematical positions of the SC SN PC and the centre of the Earthcomputed in R-Spacecraft the program next computes the location of the camerarsquosprincipal point (PP) The principle point is the point of intersection of the opticalaxis of the lens with the image plane (ie the lm) It is nominally positioned at a
Astronaut photography as digital data for remote sensing 4427
distance equal to the focal length from the perspective centre of the camera (whichis assumed to be at SC) along the vector from PC through SC as shown in gure 9
A third coordinate system the Rectangular Photo Coordinate System (R-Photo)is created with its origin at PP its XndashY axial plane normal to the vector from PCthrough SC and its +X axis aligned with the +X axis of R-Spacecraft as shownin gure 9 The XndashY plane of this coordinate system represents the image plane ofthe photograph
534 Auxiliary point calculationsThe calculations above employ a critical assumption that all photos are taken
with the lsquotop of the imagersquo oriented perpendicular to the vector from the SN towardsPC as shown in gure 9 To avoid non-uniform solar heating of the external surfacemost orbiting spacecraft are slowly and continually rotated about one or more axesIn this condition a ight crew member taking a photo while oating in microgravitycould orient the photo with practically any orientation relative to the horizon (seealso gure 8(b)) Unfortunately since these photos are taken with conventionalhandheld cameras there is no other information available which can be used toresolve the photorsquos rotational ambiguity about the optical axis other then the photoitself This is why the above assumption is used and the footprint computed by thiscalculator is actually a lsquolocator ellipsersquo which estimates the area on the ground whatis likely to be included in a speci c photograph (see gure 6) This locator ellipse ismost accurate for square image formats and subject to additional distortion as thephotograph format becomes more rectangular
If the user wants a more precise calculation of the image footprint the photorsquosrotational ambiguity about the optical axis must be resolved This can be done inthe calculator by adding data for an auxiliary point Detailed instructions regardinghow to prepare and use auxiliary point data in the computations are included in thelsquoHow-To-Usersquo tab of the spreadsheet Basically the user determines which side ofthe photograph is top and then measures the angle between the line from PP to thetop of the photo and from PP to the auxiliary point on the photo ( gure 9)
If the user includes data for an auxiliary point a series of computations arecompleted to resolve the photo rotation ambiguity about the optical axis (ie the
+Z axis in R-Photo) A vector from the Auxiliary Point on the Earth (AE) throughthe photograph perspective centre ( located at SC) is intersected with the photo imageplane (XndashY plane of R-Photo) to compute the coordinates of the Auxiliary Point onthe photo (AP) in R-Photo A two-dimensional angle in the XndashY plane of R-Photofrom the shy X axis to a line from PP to AP is calculated as shown in gure 9 The
shy X axis is used as the origin of the angle since it represents the top of the photoonce it passes through the perspective centre The diVerence between the computedangle and the angle measured by the user on the photo resolves the ambiguity inthe rotation of the photo relative to the principal line ( gures 7 and 9) Thetransformations from R-Spacecraft and R-Photo are then modi ed to include anadditional rotation angle about the +Z axis in R-Photo
535 lsquoFootprint rsquo calculationsThe program next computes the coordinates of eight points about the perimeter
of the image format (ie located at the four photo corners plus a bisector pointalong each edge of the image) These points are identi ed in R-Photo based uponthe photograph format size and then converted to R-Spacecraft Since all computa-tions are done in orthogonal coordinate systems the R-Spacecraft to R-Photo
J A Robinson et al4428
rotation matrix is transposed to produce an R-Photo to R-Spacecraft rotation matrixOnce in R-Spacecraft a unit vector from each of the eight perimeter points throughthe photo perspective centre (the same point as SC) is computed This provides thecoordinates for points about the perimeter of the image format with their directionvectors in a common coordinate system with other key points needed to computethe photo footprint
The next step is to compute the point of intersection between the spherical earthmodel and each of the eight perimeter point vectors The scalar value for eachperimeter point unit vector is computed using two-dimensional planar trigonometryAn angle c is computed using the formula for the cosine of an angle between twoincluded vectors (the perimeter point unit vector and the vector from SC to thecentre of the Earth) Angle y is computed using the Law of Sines Angle e=180degrees shy c shy y The scalar value of the perimeter point vector is computed using eand the Law of Cosines The scalar value is then multiplied by the perimeter pointunit vector to produce the three-dimensional point of intersection of the vector withEarthrsquos surface in R-Spacecraft The process is repeated independently for each ofthe eight perimeter point vectors Aside from its mathematical simplicity the valueof arcsine (computed in step 2) will exceed the normal range for the sin of an anglewhen the perimeter point vectors fail to intersect with the surface of the earth Asimple test based on this principle allows the program to correctly handle obliquephotos which image a portion of the horizon (see results for high oblique photographsin table 5)
The nal step in the process converts the eight earth intersection points from theR-Spacecraft to R-Earth The results are then converted to the standard geographiccoordinate system and displayed on the lsquoInput-Outputrsquo page of the spreadsheet
54 ExamplesAll three formulations were applied to the photographs included in this paper
and the results are compared in table 5 Formulation 1 gives too small a value fordistance across the photograph (D) for all but the most nadir shots and thus servesas an indicator of the best theoretical case but is not a good measure for a speci cphotograph For example the photograph of Lake Eyre taken from 276 km altitude( gure 2(a)) and the photograph of Limmen Bight ( gure 7) were closest to beingnadir views (oVsetslt68 km or tlt15deg table 5) For these photographs D calculatedusing Formulation 1 was similar to the minimum D calculated using Formulations2 and 3 For almost all other more oblique photographs Formulation 1 gave asigni cant underestimate of the distance covered in the photograph For gure 5(A)(the picture of Houston taken with a 40 mm lens) Formulation 1 did not give anunderestimate for D This is because Formulation 1 does not account for curvatureof the Earth in any way With this large eld of view assuming a at Earth in atedthe value of D above the minimum from calculations that included Earth curvature
A major diVerence between Formulations 2 and 3 is the ability to estimate pixelsizes (P) in both directions (along the principal line and perpendicular to the principalline) For the more oblique photographs the vertical estimate of D and pixel sizes ismuch larger than in the horizontal direction (eg the low oblique and high obliquephotographs of Hawaii gure 4 table 5)
For the area of Limmen Bight ( gure 7) table 5 illustrates the improvement inthe estimate of distance and pixel size that can be obtained by re-estimating thelocation of the PC with greater accuracy Centre points in the catalogued data are
Astronaut photography as digital data for remote sensing 4429
plusmn05deg of latitude and longitude When the centre point was re-estimated to plusmn002degwe determined that the photograph was not taken as obliquely as rst thought(change in the estimate of the look angle t from 169 to 128deg table 5) When theauxiliary point was added to the calculations of Formula 3 the calculated look angleshrank further to 110deg indicating that this photograph was taken at very close toa nadir view Of course this improvement in accuracy could have also led to estimatesof greater obliquity and corresponding larger pixel sizes
We also tested the performance of the scale calculator with auxiliary point byestimating the corner and point locations on the photograph using a 11 000 000Operational Navigational Chart For this test we estimated our ability to readcoordinates from the map as plusmn002degand our error in nding the locations of thecorner points as plusmn015deg (this error varies among photographs depending on thedetail that can be matched between photo and map) For Limmen Bight ( gure 7)the mean diVerence between map estimates and calculator estimates for four pointswas 031deg (SD=018 n=8) For a photograph of San Francisco Bay (STS062-151-291) the mean diVerence between map estimates and calculator estimates forfour points was 0064deg (SD=018 n=8) For a photograph of San Francisco Bay(STS062-151-291) the mean diVerence between map estimates and calculator estim-ates for eight points was 0196deg (SD=0146 n=16) Thus in one case the calculatorestimates were better than our estimate of the error in locating corner points on themap It is reasonable to expect that for nadir-viewing photographs the calculatorused with an auxiliary point can estimate locations of the edges of a photograph towithin plusmn03deg
55 Empirical con rmation of spatial resolution estimatesAs stated previously a challenge to estimating system-AWAR for astronaut
photography of Earth is the lack of suitable targets Small features in an image cansometimes be used as a check on the size of objects that can be successfully resolvedgiving an approximate value for GRD Similarly the number of pixels that make upthose features in the digitized image can be used to make an independent calculationof pixel size We have successfully used features such as roads and airport runwaysto make estimates of spatial scale and resolution (eg Robinson et al 2000c) Whilerecognizing that the use of linear detail in an image is a poor approximation to abar target and that linear objects smaller than the resolving power can often bedetected (Charman 1965) few objects other than roads could be found to make anydirect estimates of GRD Thus roads and runways were used in the images ofHouston (where one can readily conduct ground veri cations and where a numberof higher-contrast concrete roadways were available) to obtain empirical estimatesof GRD and pixel size for comparison with table 5
In the all-digital ESC image of Houston ( gure 3(d )) we examined EllingtonField runway 4-22 (centre left of the image) which is 24384 mtimes457 m This runwayis approximately 6ndash7 pixels in width and 304ndash3094 pixels in length so pixelsrepresent an distance on the ground 7ndash8 m Using a lower contrast measure of astreet length between two intersections (2125 m=211 pixels) a pixel width of 101 mis estimated These results compare favourably with the minimum estimate of 81 mpixels using Formulation 1 (table 5) For an estimate of GRD that would be morecomparable to aerial photography of a line target the smallest street without treecover that could be clearly distinguished on the photograph was 792 m wide Thesmallest non-street object (a gap between stages of the Saturn rocket on display in
J A Robinson et al4430
Tab
le5
App
lica
tio
nof
met
hod
sfo
res
tim
ati
ng
spati
alre
solu
tio
no
fast
ron
aut
ph
oto
gra
ph
sin
clu
din
gF
orm
ula
tion
s12
and
3Im
pro
ved
cen
tre
po
int
esti
mat
esan
dauxilia
ryp
oin
tca
lcu
lati
ons
are
sho
wn
for
gure
7V
aria
ble
sas
use
din
the
text
fle
ns
foca
lle
ngt
hH
al
titu
de
PC
ph
oto
gra
ph
cen
tre
poin
tSN
sp
ace
craft
nadir
po
int
D
dis
tance
cover
edo
nth
egr
oun
d
Pd
ista
nce
on
the
grou
nd
repre
sen
ted
by
apix
el
OVse
td
ista
nce
bet
wee
nP
Cand
SNusi
ng
the
grea
tci
rcle
form
ula
t
look
angle
away
from
SN
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
2L
ake
Eyre
ST
S093
-717
-80
250
276
shy29
0137
0shy
28
413
71
607
11
767
413
7609
ndash64
212
013
760
9ndash64
7121
124
ST
S082
-754
-61
250
617
shy28
5137
0shy
28
113
50
135
726
1201
18
013
8ndash14
827
517
9138ndash15
3277
294
Fig
ure
3H
ou
sto
nS
TS
062
-97-
151
4029
829
5shy
95
530
5shy
95
4410
78
8112
20
5348ndash58
990
1205
351
ndash63
8951
10
36
ST
S094
-737
-28
100
294
295
shy
95
528
3shy
95
5162
31
1133
24
415
8ndash20
334
724
3158ndash20
7351
390
ST
S055
-71-
41250
302
295
shy
95
028
2shy
93
766
412
8192
32
5736
ndash84
715
232
273
9ndash85
7154
185
S73
E51
45300
2685
295
shy
95
024
78
0F
igu
re4
Haw
aii
ST
S069
-701
-65
250
370
195
shy
155
520
4shy
153
881
415
7204
28
9876
ndash98
917
928
688
0ndash10
9181
210
ST
S069
-701
-70
250
370
210
shy
157
519
2shy
150
781
415
7738
63
414
9ndash23
336
760
7148ndash55
6394
10
63
ST
S069
-729
-27
100
398
200
shy
156
620
5shy
148
2219
42
1867
65
432
8ndash13
10
158
62
4323+
311
N
A
Astronaut photography as digital data for remote sensing 4431
Tab
le5
(Cont
inued
)
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
7L
imm
enB
igh
tS
TS
093
-704
-50
250
283
shy15
0135
0shy
15
013
58
623
120
859
16
963
0ndash67
3125
16
863
0ndash67
612
613
2P
CR
e-es
tim
ated
shy14
733
1352
67
64
5128
623
ndash65
5123
128
623
ndash65
712
3127
Auxilia
ryP
oin
t611
062
3ndash65
112
312
5F
igu
re10
N
ear
Au
stin
T
XS
TS
095
-716
-46
250
543
30
5shy
975
28
6shy
1007
120
230
375
34
613
5ndash157
281
34
1136ndash16
128
636
0
1 All
pho
togr
aphs
exce
pt
on
eta
ken
wit
hH
ass
elbla
dca
mer
aso
dis
tan
ceacr
oss
the
image
d=
55m
m
for
S73
E51
45
taken
wit
hel
ectr
onic
still
cam
erad=
276
5m
mho
rizo
nta
l2 P
Cand
SN
are
giv
enin
the
form
(lati
tud
elo
ngit
ude)
deg
rees
w
ith
neg
ativ
en
um
ber
sre
pre
senti
ng
south
and
wes
tA
ccura
cyo
fP
Cis
plusmn0
5and
SN
isplusmn
01
as
reco
rded
inth
eA
stro
naut
Ph
oto
gra
phy
Data
base
ex
cep
tw
her
en
ote
dth
atce
ntr
ep
oin
tw
asre
-est
imate
dw
ith
grea
ter
accu
racy
3 C
alc
ula
ted
usi
ng
the
mea
nofth
em
inim
um
an
dm
axim
um
esti
mate
sof
DF
orm
ula
tion
2co
nsi
der
sD
inth
ed
irec
tion
per
pen
dic
ula
rto
the
Pri
nci
pal
Lin
eo
nly
4 C
alc
ula
ted
inho
rizo
nta
lan
dve
rtic
aldir
ecti
on
sb
ut
still
ass
um
ing
geom
etry
of
gure
6(B
)F
irst
nu
mb
eris
the
mea
no
fth
em
inim
um
Dat
the
bott
om
of
the
ph
oto
and
the
maxi
mum
Dat
the
top
of
the
ph
oto
(range
show
nin
the
tab
le)
and
isco
mpara
ble
toF
orm
ula
tio
n2
Sec
ond
num
ber
isth
em
ean
of
Des
tim
ated
alon
gth
eP
rin
cip
alline
and
alo
ng
the
ou
tsid
eed
ge
(range
not
show
n)
5 Alt
itu
de
isap
pro
xim
ate
for
mis
sion
as
exac
tti
me
pho
togr
ap
hw
asta
ken
and
corr
esp
on
din
gn
adir
data
are
no
tava
ilab
le
Lack
of
nad
irdata
pre
ven
tsap
plica
tion
of
Form
ula
tion
s2
an
d3
6 Au
xilliar
ypo
int
add
edw
as
shy14
80
lati
tud
e13
51
2lo
ngit
ude
shy46
5degro
tati
on
angle
in
stru
ctio
ns
for
esti
mati
ng
the
rota
tio
nan
gle
para
met
erare
conta
ined
as
par
tof
the
scal
eca
lcu
lato
rsp
read
shee
tdo
cum
enta
tio
n
J A Robinson et al4432
a park at Johnson Space Center) that could clearly be distinguished on the photo-graph was 853 m wide
For the photograph of Houston taken with a 250 mm lens ( gure 3(C)) anddigitized from second generation lm at 2400 ppi (106 mm pixel Otilde 1 ) Ellington Fieldrunway 4-22 is 3 pixels in width and 1613 pixels in length so pixels represent151ndash152 m on the ground These results compare favourably with the minimumestimate of 152 m using Formulation 2 and 154ndash185 m pixels using Formulation 3(table 5) For an estimate of GRD using an 8times8 inch print (1369 enlargement) and4times magni cation the smallest street that could clearly be distinguished was 822 mwide the same feature could barely be distinguished on the digitized image
We also made an empirical estimate of spatial resolution for lower contrastvegetation boundaries By clearing forest so that a pattern would be visible to landingaircraft a landowner outside Austin Texas (see also aerial photo in Lisheron 2000)created a target that is also useful for evaluating spatial resolution of astronautphotographs The forest was selectively cleared in order to spell the landownerrsquosname lsquoLUECKErsquo with the remaining trees ( gure 10) According to local surveyorswho planned the clearing the plan was to create letters that were 3100 fttimes1700 ft(9449 mtimes5182 m) Photographed at a high altitude relative to most Shuttle missions(543 km) with a 250 mm lens Formula 3 predicts that each pixel would represent anarea 286 mtimes360 m on the ground (table 5) When original lm was digitized at2400 ppi (106 mm pixel Otilde 1 ) letters correspond to 294times188 pixels for a comparablepixel size of 27ndash32 m
6 Summary and conclusionsAstronaut photographs can be an excellent source of data for remote sensing
applications Best-case resolutions are similar to that for Landsat or SPOT withpixels as small as lt10 m It was estimated that of images taken to date 504 hadlens and obliquity characteristics that make them potential candidates for remotesensing information Digitized astronaut photographs can be overlaid with othersatellite data using GIS (Eckardt et al 2000) or used to ll in gaps in time serieswhen other imagery is not available As a source of public-domain information theycan be very useful for scientists who do not have access to satellite imagery eitherbecause of the expense of image acquisition (to the end user) or the computersystems needed for processing satellite images These diVerences in image acquisitioncosts (to the user) are summarized in table 6
Searching of the complete database of NASA astronaut photography includinglow-spatial-resolution browse images is available via the Web (OYce of EarthSciences 2000) This provides nearly global access for identifying images that willcontribute to a speci c scienti c project For the most detailed studies digitalproducts posted to the web will not be of suYcient quality To date we have madeit a practice of digitizing small numbers of images when requested by scientists atno charge Such requests (including a description of the project involved) can bemade through the authors or using the contact information listed on the websiteInvestigators with needs for larger numbers of digital images have also been servedthrough collaborative agreements
In our experience scientists that nd the data most valuable are those in develop-ing countries those studying areas that have not usually been targets for the majorsatellites those having diYculty in nding low-cloud images those interested inconstructing time series or those interested in using a large number of images
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
Astronaut photography as digital data for remote sensing 4415
Figure 5 (a) Distributions of astronaut photographs by look angle oV-nadir (calculated fromcentre point and nadir point using Formulation 2) Data included for all photographsfor which centre and nadir points are known with high oblique look angles (n=55 155) excluded (Total sample shown is 108 293 of 382 563 total records in thedatabase when compiled For records where nadir data was not available number ofobservations for qualitative look angles are near vertical 14 086 low oblique 115 834undetermined 89 195) (b) Altitude distribution for astronaut photographs of Earthtaken with the Hasselblad camera Data included for Hasselblad camera only focallength determined with high oblique look angles excluded (Total sample shown is159 046 of 206 971 Hasselblad records with known altitude and of 382 563 total recordsin the database when compiled For Hasselblad records with known altitude data notshown includes high oblique photographs using 100 mm and 250 mm lenses n=20 262and all look angles with other lenses n=27 663)
systems on Skylab (NASA 1974 Wilmarth et al 1977) and occasionally on Shuttlemissions and Shuttle-Mir missions (table 3) The CIR lm used is a three-layerAerochrome having one layer that is sensitive to re ected solar infrared energy toapproximately 900 nm (Eastman Kodak Company 1998) protective coatings onmost spacecraft windows also limit IR transmittance between 800 mm and 1000 nmThis layer is extremely sensitive to the temperature of the lm which createsunpredictable degradation of the IR signature under some space ight conditions
J A Robinson et al4416
Tab
le3
Det
ails
on
lm
suse
dfo
rast
ron
aut
pho
togr
aphy
of
Eart
h
Dat
ain
clud
ep
ho
togr
aphy
fro
mall
NA
SA
hum
an
spac
eig
ht
mis
sio
ns
thro
ugh
ST
S-9
6(J
un
e19
99
378
461
tota
lfr
ames
inth
edata
base
)F
ilm
sw
ith
lt50
00fr
am
esex
po
sed
and
lm
suse
dfo
r35
mm
cam
eras
only
are
no
tin
clud
ed
Res
olv
ing
Nu
mb
erof
Film
man
ufa
ctu
rer
AS
Ap
ow
er1
fram
esP
erio
dof
use
Cam
eras
Colo
ur
posi
tive
Kod
akA
eroch
rom
eII
(2447
2448
SO
-131S
O-3
68
)32
40ndash50
513
8196
8ndash19
85
Hass
elbla
dL
inh
of
Kod
akE
kta
chro
me
(5017
502
56
017
6118
SO
-121
64
50
113
932
196
5ndash19
68
Hass
elbla
dL
inh
of
Role
iex
SO
-217
Q
X-8
24
QX
-868
)19
81ndash1993
Mau
rer
Nik
on
Kod
akL
um
iere
(5046
5048
)100
105
743
199
4ndash19
98
Hass
elbla
dL
inho
fK
od
akE
lite
100
S(5
069
)100
41
486
199
6ndashp
rese
nt
Hass
elbla
d2
Fu
jiV
elvia
50C
S135
-36
32
13
882
1992ndash19
97H
ass
elbla
dN
iko
nK
od
akH
i-R
esolu
tio
nC
olo
r(S
O-2
42S
O-3
56
)10
096
74
1973ndash19
75H
ass
elbla
d
S19
0A
S190
B
Infr
are
dand
bla
ck-a
nd-w
hit
e
lms
Kod
akA
eroch
rom
eC
olo
rIn
frar
ed40
32
40
704
196
5ndashp
rese
nt2
Hass
elbla
dL
inh
of
Nik
on
S190
A
(8443
34
432443
)S
190B
Kod
akB
ampW
Infr
are
d(2
424
)32
10
553
197
3ndash19
74
S190
AK
od
akP
anato
mic
-XB
ampW
(SO
-022
)63
10
657
197
3ndash19
74
S190
A
1A
tlo
wco
ntr
ast
(ob
ject
back
grou
nd
rati
o16
1)
aspro
vid
edby
Ko
dak
and
list
edby
Sla
ter
etal
(1
983
256
)R
esolv
ing
pow
erw
as
dis
con
tin
ued
fro
mth
ete
chn
ical
test
sper
form
edb
yK
odak
inapp
roxim
atel
y19
93
an
dit
isn
ot
kn
ow
nif
curr
ent
lm
sh
ave
sim
ilar
reso
lvin
gpo
wer
too
lder
ver
sion
so
fsi
milar
lm
s(K
T
eite
lbaum
K
od
akp
erso
nal
com
mu
nic
atio
n)
Th
egra
nu
lari
tym
easu
reno
wpro
vid
edb
yK
od
ak
can
no
tb
ere
adily
con
ver
ted
toa
spat
ial
reso
luti
on
2 The
Lin
hof
was
use
dag
ain
on
ST
S-1
03in
Dec
emb
er19
99(d
ata
not
cata
logued
asof
the
pre
para
tion
of
this
tab
le)
usi
ng
Ko
dak
Elite
100S
lm
3 B
egin
nin
gin
July
1999
2443
colo
ur-
infr
are
d
lmpro
cess
was
chan
ged
from
EA
-5to
E-6
Astronaut photography as digital data for remote sensing 4417
413 Film duplicationThe original lm is archived permanently after producing about 20 duplicate
printing masters (second generation) Duplicate printing masters are disseminatedto regional archives and used to produce products (thirdndashfourth generation) for themedia the public and for scienti c use When prints slides transparencies anddigital products are produced for public distribution they are often colour adjustedto correspond to a more realistic look Digital products from second generationcopies can be requested by scienti c users (contact the authors for information onobtaining such products for a particular project) and these are recommended forremote sensing applications Care should be taken that the digital product acquiredfor remote sensing analysis has not been colour adjusted for presentation purposesBased on qualitative observations there is little visible increase in GRD in the thirdor fourth generation products There is a signi cant degradation in delity of colourin third and fourth generation duplicates because an increase in contrast occurswith every copy
42 Shutter speedsThe impact of camera motion (both due to the photographer and to the motion
of the spacecraft ) on image quality is determined by shutter speedmdash1250 to 1500second have been used because slower speeds record obvious blurring caused bythe rapid motion of the spacecraft relative to the surface of the Earth Generally1250 was used for ISO 64 lms (and slower) and after the switch to ISO 100 lmsa 1500 setting became standard New Hasselblad cameras that have been ownbeginning with STS-92 in October 2000 vary shutter speed using a focal plane shutterfor exposure bracketing (rather than varying aperture) and 1250 1500 and 11000second are used
Across an altitude range of 120ndash300 nautical miles (222ndash555 km) and orbitalinclinations of 285deg and 570deg the median relative ground velocity of orbitingspacecraft is 73 km s Otilde 1 At a nominal shutter speed of 1500 the expected blur dueto motion relative to the ground would be 146 m When cameras are handheld (asopposed to mounted in a bracket) blur can be reduced by physical compensationfor the motion (tracking) by the photographer many photographs show a level ofdetail indicating that blur at this level did not occur (eg gure 3(d )) Thus motionrelative to the ground is not an absolute barrier to ground resolution for handheldphotographs
43 Spacecraft windowsMost of the photographs in the NASA Astronaut Photography Database were
taken through window ports on the spacecraft used The transmittance of the windowport may be aVected by material inhomogeniety of the glass the number of layersof panes coatings used the quality of the surface polish environmentally inducedchanges in window materials (pressure loads thermal gradients) or deposited con-tamination Such degradation of the window cannot be corrected is diVerent foreach window and changes over time See Eppler et al (1996) and Scott (2000) fordiscussion of the spectral transmittance of the fused quartz window that is part ofthe US Laboratory Module (Destiny) of the International Space Station
44 Digitized images from lmWhen lm is scanned digitally the amount of information retained depends on
the spectral information extracted from the lm at the spatial limits of the scanner
J A Robinson et al4418
To date standard digitizing methodologies for astronaut photographs have not beenestablished and lm is digitized on a case-by-case basis using the equipment available
441 Digitizing and spatial resolutionLight (1993 1996) provided equations for determining the digitizing spatial
resolution needed to preserve spatial resolution of aerial photography based on thestatic resolving power (AWAR) for the system For lms where the manufacturer hasprovided data (Eastman Kodak 1998 K Teitelbaum personal communication) the resolving power of lms used for astronaut photography ranges from 32 lp mm Otilde 1to 100 lp mm Otilde 1 at low contrast (objectbackground ratio 161 table 3) The AWARfor the static case of the Hasselblad camera has been measured at high and lowcontrast (using lenses shown in table 1) with a maximum of approximately55 lp mm Otilde 1 (Fred Pearce unpublished data)
Based on the method of Light (1993 1996) the dimension of one spatial resolutionelement for a photograph with maximum static AWAR of 55 lp mm Otilde 1 would be18 mm lp Otilde 1 The acceptable range of spot size to preserve spatial information wouldthen be
18 mm
2 atilde 2aring scan spot size aring
18 mm
2(1)
and 6 mm aring scan spot size aring 9 mm Similarly for a more typical static AWAR of30 lp mm Otilde 1 (33 mm lpOtilde 1 low contrast Fred Pearce unpublished data) 11 mm aring scanspot sizearing 17 mm These scan spot sizes correspond to digitizing resolutions rangingfrom 4233ndash2822 ppi (pixels inch Otilde 1 ) for AWAR of 55 lp mm Otilde 1 and 2309ndash1494 ppi forAWAR of 30 lp mm Otilde 1 Scan spot sizes calculated for astronaut photography arecomparable to those calculated for the National Aerial Photography Program(9ndash13 mm Light 1996)
Widely available scanners that can digitize colour transparency lm currentlyhave a maximum spatial resolution of approximately 2400 ppi (106 mm pixel Otilde 1) Forexample we routinely use an Agfa Arcus II desktop scanner (see also Baltsavias1996) with 2400 ppi digitizing spatial resolution (2400 ppi optical resolution in onedirection and 1200 ppi interpolated to 2400 ppi in the other direction) and 2400 ppiwas used to calculate IFOV equivalents (tables 2 and 4) For some combinations oflens camera and contrast 2400 ppi will capture nearly all of the spatial informationcontained in the lm However for lm with higher resolving power thanEktachrome-64 for better lenses and for higher contrast targets digitizing at 2400 ppiwill not capture all of the spatial information in the lm
Improvements in digitizing technology will only produce real increases in IFOVto the limit of the AWAR of the photography system The incremental increase inspatial information above 3000 ppi (85 mm pixel Otilde 1 ) is not likely to outweigh thecosts of storing large images (Luman et al 1997) At 2400 ppi a Hasselblad frameis approximately 5200times5200 or 27 million pixels (table 2) while the same imagedigitized at 4000 ppi would contain 75 million pixels
Initial studies using astronaut photographs digitized at 2400 ppi (106 mm pixel Otilde 1 Webb et al 2000 Robinson et al 2000c) indicate that some GRD is lost comparedto photographic products Nevertheless the spatial resolution is still comparablewith other widely used data sources (Webb et al 2000) Robinson et al (2000c)found that digitizing at 21 mm pixel Otilde 1 provided information equivalent to 106 mmpixel Otilde 1 for identifying general urban area boundaries for six cities except for a single
Astronaut photography as digital data for remote sensing 4419
photo that required higher spatial resolution digitizing (it had been taken with ashorter lens and thus had less spatial resolution) Part of the equivalence observedin the urban areas study may be attributable to the fact that the atbed scannerused interpolates from 1200 ppi to 2400 ppi in one direction The appropriate digitiz-ing spatial resolution will in part depend on the scale of features of interest to theresearchers with a maximum upper limit set of a scan spot size of approximately6 mm (4233 ppi)
442 Digitizing and spectral resolutionWhen colour lm is digitized there will be loss of spectral resolution The three
lm emulsion layers (red green blue) have relatively distinct spectral responses butare fused together so that digitizers detect one colour signal and must convert itback into three (red green blue) digital channels Digital scanning is also subject tospectral calibration and reproduction errors Studies using digital astronaut photo-graphs to date have used 8 bits per channel However this is another parameterthat can be controlled when the lm is digitized We do not further address spectralresolution of digitally scanned images in this paper
45 External factors that in uence GRDAlthough not discussed in detail in this paper factors external to the spacecraft
and camera system (as listed by Moik (1980) for remote sensing in general) alsoimpact GRD These include atmospheric interference (due to scattering attenuationhaze) variable surface illumination (diVerences in terrain slope and orientation) andchange of terrain re ectance with viewing angle (bidirectional re ectance) For astro-naut photographs variable illumination is particularly important because orbits arenot sun-synchronous Photographs are illuminated by diVerent sun angles and imagesof a given location will have colour intensities that vary widely In addition theviewing angle has an eVect on the degree of object-to-backgroun d contrast andatmospheric interference
5 Estimating spatial resolution of astronaut photographsIn order to use astronaut photographs for digital remote sensing it is important
to be able to calculate the equivalent to an IFOVmdashthe ground area represented bya single pixel in a digitized orbital photograph The obliquity of most photographsmeans that pixel lsquosizesrsquo vary at diVerent places in an image Given a ground distanceD represented by a photograph in each direction (horizontal vertical ) an approxi-mate average pixel width (P the equivalent of IFOV) for the entire image can becalculated as follows
P=D
dS(2)
where D is the projected distance on the ground covered by the image in the samedirection as the pixel is measured d is the actual width of the image on the original lm (table 1) and S is the digitizing spatial resolution
Here we present three mathematical formulations for estimating the size of thefootprint or area on the ground covered by the image Example results from theapplication of all three formulations are given in table 5 The rst and simplestcalculation (formulation 1) gives an idea of the maximum spatial resolution attainableat a given altitude of orbit with a given lm format and a perfectly vertical (nadir)
J A Robinson et al4420
view downward Formulation 2 takes into account obliquity by calculating lookangle from the diVerence between the location on the ground represented at thecentre of the photograph and the nadir location of the spacecraft at the time thephotograph was taken ( gure 1) Formulation 3 describes an alternate solution tothe oblique look angle problem using coordinate-system transformations Thisformulation has been implemented in a documented spreadsheet and is availablefor download (httpeoljscnasagovsseopFootprintCalculatorxls) and is in theprocess of being implemented on a Web-based user interface to the AstronautPhotography Database (OYce of Earth Sciences 2000)
Although formulations 2 and 3 account for obliquity for the purposes of calcula-tion they treat the position of the spacecraft and position of the camera as one Inactuality astronauts are generally holding the cameras by hand (although camerasbracketed in the window are also used) and the selection of window position of theastronaut in the window and rotation of the camera relative to the movement ofthe spacecraft are not known Thus calculations using only the photo centre point(PC) and spacecraft nadir point (SN) give a locator ellipse and not the locations ofthe corners of the photograph A locator ellipse describes an estimated area on theground that is likely to be included in a speci c photograph regardless of the rotationof the lm plane about the camerarsquos optical axis ( gure 6)
Estimating the corner positions of the photo requires additional user input of asingle auxiliary pointmdasha location on the image that has a known location on theground Addition of this auxiliary point is an option available to users of thespreadsheet An example of the results of adding an auxiliary point is shown in gure 7 with comparisons of the various calculations in table 5
51 Formulation 1 Footprint for a nadir viewThe simplest way to estimate footprint size is to use the geometry of camera lens
and spacecraft altitude to calculate the scaling relationship between the image in the
Figure 6 Sketch representation of the locator ellipse an estimated area on the ground thatis likely to be included in a speci c photograph regardless of the rotation of the lmplane about the camerarsquos optical axis
Astronaut photography as digital data for remote sensing 4421
Figure 7 An example of the application of the Low Oblique Space Photo FootprintCalculator The photograph (STS093-704-50) of Limmen Bight Australia was takenfrom an altitude of 283 km using the Hasselblad camera with a 250 mm lens Mapused is 11 000 000 Operational Navigational Chart The distances shown equate toan average pixel size of 124 m for this photograph
lm and the area covered on the ground For a perfect nadir view the scalerelationship from the geometry of similar triangles is
d
D=
f
H(3)
where d=original image size D=distance of footprint on the ground f =focallength of lens and H=altitude Once D is known pixel size ( length=width) can becalculated from equation (2) These calculations represent the minimum footprintand minimum pixel size possible for a given camera system altitude and digitisingspatial resolution (table 2) Formulation 1 was used for calculating minimum pixelsizes shown in tables 2 and 4
By assuming digitizing at 2400 ppi (106 mm pixel Otilde 1 ) currently a spatial resolutioncommonly attainable from multipurpose colour transparency scanners (see sect441)this formulation was used to convert area covered to an IFOV equivalent for missionsof diVerent altitudes (table 2) Table 4 provides a comparison of IFOV of imagesfrom various satellites including the equivalent for astronaut photography Valuesin this table were derived by using formulation 1 to estimate the area covered becausea perfect nadir view represents the best possible spatial resolution and smallest eldof view that could be obtained
52 Formulation 2 Footprint for oblique views using simpli ed geometry and thegreat circle distance
A more realistic approach to determining the footprint of the photographaccounts for the fact that the camera is not usually pointing perfectly down at the
J A Robinson et al4422
Table 4 Comparison of pixel sizes (instantaneous elds of view) for satellite remote sensorswith pixel sizes for near vertical viewing photography from spacecraft Note that alldata returned from the automated satellites have the same eld of view while indicatedpixel sizes for astronaut photography are best-case for near vertical look angles See gure 5 for distributions of astronaut photographs of various look angles High-altitude aerial photography is also included for reference
Altitude PixelInstrument (km) width (mtimesm) Bands1
Landsat Multipectral Scanner2 (MSS 1974-) 880ndash940 79times56 4ndash5Landsat Thematic Mapper (TM 1982-) ~705 30times30 7Satellite pour lrsquoObservation de la Terre (SPOT 1986-) ~832
SPOT 1-3 Panchromatic 10times10 1SPOT 1-3 Multispectral (XS) 20times20 3
Astronaut Photography (1961-)Skylab3 (1973-1974 ) ~435
Multispectral Camera (6 camera stations) 17ndash19times17ndash19 3ndash8Earth Terrain Camera (hi-res colour) 875times875 3
Space Shuttle5 (1981-) 222ndash611Hasselblad 70 mm (250mm lens colour) vertical view 9ndash26times9ndash26 3Hasselblad 70 mm (100mm lens colour) vertical view 23ndash65times23ndash65 3Nikon 35 mm (400mm lens colour lm or ESC) vertical 5ndash16times5ndash16 3
High Altitude Aerial Photograph (1100 000) ~12 2times2 3
1Three bands listed for aerial photography obtainable by high-resolution colour digitizingFor high-resolution colour lm at 65 lp mmOtilde 1 (K Teitelbaum Eastman Kodak Co personalcommunication) the maximum information contained would be 3302 ppi
2Data from Simonett (1983Table 1-2)3Calculated as recommended by Jensen (19831596) the dividend of estimated ground
resolution at low contrast given in NASA (1974179-180) and 244Six lters plus colour and colour infrared lm (NASA 19747) 5Extracted from table 2
nadir point The look angle (the angle oV nadir that the camera is pointing) can becalculated trigonometrically by assuming a spherical earth and calculating the dis-tance between the coordinates of SN and PC ( gure 1) using the Great Circledistance haversine solution (Sinnott 1984 Snyder 198730ndash32 Chamberlain 1996)The diVerence between the spacecraft centre and nadir point latitudes Dlat=lat 2 shy lat 1 and the diVerence between the spacecraft centre and nadir pointlongitudes Dlon=lon 2 shy lon 1 enter the following equations
a=sin2ADlat
2 B+cos(lat 1) cos(lat 2) sin2ADlon
2 B (4)
c=2 arcsin(min[1 atilde a]) (5)
and oVset=R c with R the radius of a spherical Earth=6370 kmThe look angle (t ) is then given by
t=arctanAoffset
H B (6)
Assuming that the camera was positioned so that the imaginary line between thecentre and nadir points (the principal line) runs vertically through the centre of thephotograph the distance between the geometric centre of the photograph (principalpoint PP) and the top of the photograph is d2 ( gure 8(a)) The scale at any point
Astronaut photography as digital data for remote sensing 4423
(a) (b)
Figure 8 The geometric relationship between look angle lens focal length and the centrepoint and nadir points of a photograph (see also Estes and Simonett 1975) Note thatthe Earthrsquos surface and footprint represented in gure 1 are not shown here only arepresentation of the image area of the photograph Variables shown in the gurelook angle t focal length f PP centre of the image on the lm SN spacecraft nadird original image size (a) The special case where the camera is aligned squarely withthe isometric parallel In this case tilt occurs along a plane parallel with the centre ofthe photograph When the conditions are met calculations using Formulation 3 willgive the ground coordinates of the corner points of the photograph (b) The generalrelationship between these parameters and key photogrammetric points on the photo-graph For this more typical case coordinates of an additional point in the photographmust be input in order to adjust for twist of the camera and to nd the corner pointcoordinates
in the photograph varies as a function of the distance y along the principal linebetween the isocentre and the point according to the relationship
d
D=
f shy ysint
H(7)
(Wong 1980 equation (214) H amp the ground elevation h) Using gure 8(a) at thetop of the photo
y= f tanAt
2B+d
2(8)
and at the bottom of the photo
y= f tanAt
2Bshyd
2(9)
Thus for given PC and SN coordinates and assuming a photo orientation as in gure 8(a) and not gure 8(b) we can estimate a minimum D (using equations (7)and (8)) and a maximum D (using equations (7) and (9)) and then average the twoto determine the pixel size (P ) via equation (2)
J A Robinson et al4424
53 Formulation 3 T he L ow Oblique Space Photo Footprint CalculatorThe Low Oblique Space Photo Footprint Calculator was developed to provide
a more accurate estimation of the geographic coordinates for the footprint of a lowoblique photo of the Earthrsquos surface taken from a human-occupied spacecraft inorbit The calculator performs a series of three-dimensional coordinate transforma-tions to compute the location and orientation of the centre of the photo exposureplane relative to an earth referenced coordinate system The nominal camera focallength is then used to create a vector from the photorsquos perspective point througheach of eight points around the parameter of the image as de ned by the formatsize The geographical coordinates for the photo footprint are then computed byintersecting these photo vectors with a spherical earth model Although more sophist-icated projection algorithms are available no signi cant increase in the accuracy ofthe results would be produced by these algorithms due to inherent uncertainties inthe available input data (ie the spacecraft altitude photo centre location etc)
The calculations were initially implemented within a Microsoft Excel workbookwhich allowed one to embed the mathematical and graphical documentation nextto the actual calculations Thus interested users can inspect the mathematical pro-cessing A set of error traps was also built into the calculations to detect erroneousresults A summary of any errors generated is reported to the user with the resultsof the calculations Interested individuals are invited to download the Excel work-book from httpeoljscnasagovsseopFootprintCalculatorxls The calculations arecurrently being encoded in a high-level programming language and should soon beavailable alongside other background data provided for each photograph at OYceof Earth Sciences (2000)
For the purposes of these calculations a low oblique photograph was de ned as
one with the centre within 10 degrees of latitude and longitude of the spacecraftnadir point For the typical range of spacecraft altitudes to date this restricted thecalculations to photographs in which Earthrsquos horizon does not appear (the generalde nition of a low oblique photograph eg Campbell 199671 and gure 4)
531 Input data and resultsUpon opening the Excel workbook the user is presented with a program intro-
duction providing instructions for using the calculator The second worksheet tablsquoHow-To-Usersquo provides detailed step-by-step instructions for preparing the baselinedata for the program The third tab lsquoInput-Output rsquo contains the user input eldsand displays the results of the calculations The additional worksheets contain theactual calculations and program documentation Although users are welcome toreview these sheets an experienced user need only access the lsquoInput-Outputrsquo
spreadsheetTo begin a calculation the user enters the following information which is available
for each photo in the NASA Astronaut Photography Database (1) SN geographicalcoordinates of spacecraft nadir position at the time of photo (2) H spacecraftaltitude (3) PC the geographical coordinates of the centre of the photo (4) f nominal focal length and (5) d image format size The automatic implementationof the workbook on the web will automatically enter these values and completecalculations
For more accurate results the user may optionally enter the geographic co-ordinates and orientation for an auxiliary point on the photo which resolves thecamerarsquos rotation uncertainty about the optical axis The auxiliary point data must
Astronaut photography as digital data for remote sensing 4425
be computed by the user following the instruction contained in the lsquoHow-To-Usersquotab of the spreadsheet
After entering the input data the geographic coordinates of the photo footprint(ie four photo corner points four points at the bisector of each edge and the centreof the photo) are immediately displayed below the input elds along with any errormessages generated by the user input or by the calculations ( gure 7) Although
results are computed and displayed they should not be used when error messagesare produced by the program The program also computes the tilt angle for each ofthe photo vectors relative to the spacecraft nadir vector To the right of the photofootprint coordinates is displayed the arc distance along the surface of the sphere
between adjacent computed points
532 Calculation assumptions
The mathematical calculations implemented in the Low Oblique Space PhotoFootprint Calculator use the following assumptions
1 The SN location is used as exact even though the true value may vary by upto plusmn01 degree from the location provided with the photo
2 The spacecraft altitude is used as exact Although the determination of the
nadir point at the instant of a known spacecraft vector is relatively precise(plusmn115times10 Otilde 4 degrees) the propagator interpolates between sets of approxi-mately 10ndash40 known vectors per day and the time code recorded on the lmcan drift Thus the true value for SN may vary by up to plusmn01 degree fromthe value provided with the photo
3 The perspective centre of the camera is assumed to be at the given altitudeover the speci ed spacecraft nadir location at the time of photo exposure
4 The PC location is used as exact even though the true value may vary by upto plusmn05deg latitude and plusmn05deg longitude from the location provided with the
photo5 A spherical earth model is used with a nominal radius of 6 372 16154 m (a
common rst-order approximation for a spherical earth used in geodeticcomputations)
6 The nominal lens focal length of the camera lens is used in the computations(calibrated focal length values are not available)
7 The photo projection is based on the classic pin-hole camera model8 No correction for lens distortion or atmospheric re ection is made
9 If no auxiliary point data is provided the lsquoTop of the Imagersquo is orientedperpendicular to the vector from SN towards PC
533 T ransformation from Earth to photo coordinate systemsThe calculations begin by converting the geographic coordinates ( latitude and
longitude) of the SN and PC to a Rectangular Earth-Centred Coordinate System(R-Earth) de ned as shown in gure 9 (with the centre of the Earth at (0 0 0))
Using the vector from the Earthrsquos centre through SN and the spacecraft altitude thespacecraft location (SC) is also computed in R-Earth
For ease of computation a Rectangular Spacecraft-Centred Coordinate System(R-Spacecraft ) is de ned as shown in gure 9 The origin of R-Spacecraft is located
at SC with its +Z-axis aligned with vector from the centre of the Earth throughSN and its +X-axis aligned with the vector from SN to PC ( gure 9) The speci c
J A Robinson et al4426
Figure 9 Representation of the area included in a photograph as a geometric projectiononto the Earthrsquos surface Note that the focal length (the distance from the SpaceShuttle to the photo) is grossly overscale compared to the altitude (the distance fromthe Shuttle to the surface of the Earth
rotations and translation used to convert from the R-Earth to the R-Spacecraft arecomputed and documented in the spreadsheet
With the mathematical positions of the SC SN PC and the centre of the Earthcomputed in R-Spacecraft the program next computes the location of the camerarsquosprincipal point (PP) The principle point is the point of intersection of the opticalaxis of the lens with the image plane (ie the lm) It is nominally positioned at a
Astronaut photography as digital data for remote sensing 4427
distance equal to the focal length from the perspective centre of the camera (whichis assumed to be at SC) along the vector from PC through SC as shown in gure 9
A third coordinate system the Rectangular Photo Coordinate System (R-Photo)is created with its origin at PP its XndashY axial plane normal to the vector from PCthrough SC and its +X axis aligned with the +X axis of R-Spacecraft as shownin gure 9 The XndashY plane of this coordinate system represents the image plane ofthe photograph
534 Auxiliary point calculationsThe calculations above employ a critical assumption that all photos are taken
with the lsquotop of the imagersquo oriented perpendicular to the vector from the SN towardsPC as shown in gure 9 To avoid non-uniform solar heating of the external surfacemost orbiting spacecraft are slowly and continually rotated about one or more axesIn this condition a ight crew member taking a photo while oating in microgravitycould orient the photo with practically any orientation relative to the horizon (seealso gure 8(b)) Unfortunately since these photos are taken with conventionalhandheld cameras there is no other information available which can be used toresolve the photorsquos rotational ambiguity about the optical axis other then the photoitself This is why the above assumption is used and the footprint computed by thiscalculator is actually a lsquolocator ellipsersquo which estimates the area on the ground whatis likely to be included in a speci c photograph (see gure 6) This locator ellipse ismost accurate for square image formats and subject to additional distortion as thephotograph format becomes more rectangular
If the user wants a more precise calculation of the image footprint the photorsquosrotational ambiguity about the optical axis must be resolved This can be done inthe calculator by adding data for an auxiliary point Detailed instructions regardinghow to prepare and use auxiliary point data in the computations are included in thelsquoHow-To-Usersquo tab of the spreadsheet Basically the user determines which side ofthe photograph is top and then measures the angle between the line from PP to thetop of the photo and from PP to the auxiliary point on the photo ( gure 9)
If the user includes data for an auxiliary point a series of computations arecompleted to resolve the photo rotation ambiguity about the optical axis (ie the
+Z axis in R-Photo) A vector from the Auxiliary Point on the Earth (AE) throughthe photograph perspective centre ( located at SC) is intersected with the photo imageplane (XndashY plane of R-Photo) to compute the coordinates of the Auxiliary Point onthe photo (AP) in R-Photo A two-dimensional angle in the XndashY plane of R-Photofrom the shy X axis to a line from PP to AP is calculated as shown in gure 9 The
shy X axis is used as the origin of the angle since it represents the top of the photoonce it passes through the perspective centre The diVerence between the computedangle and the angle measured by the user on the photo resolves the ambiguity inthe rotation of the photo relative to the principal line ( gures 7 and 9) Thetransformations from R-Spacecraft and R-Photo are then modi ed to include anadditional rotation angle about the +Z axis in R-Photo
535 lsquoFootprint rsquo calculationsThe program next computes the coordinates of eight points about the perimeter
of the image format (ie located at the four photo corners plus a bisector pointalong each edge of the image) These points are identi ed in R-Photo based uponthe photograph format size and then converted to R-Spacecraft Since all computa-tions are done in orthogonal coordinate systems the R-Spacecraft to R-Photo
J A Robinson et al4428
rotation matrix is transposed to produce an R-Photo to R-Spacecraft rotation matrixOnce in R-Spacecraft a unit vector from each of the eight perimeter points throughthe photo perspective centre (the same point as SC) is computed This provides thecoordinates for points about the perimeter of the image format with their directionvectors in a common coordinate system with other key points needed to computethe photo footprint
The next step is to compute the point of intersection between the spherical earthmodel and each of the eight perimeter point vectors The scalar value for eachperimeter point unit vector is computed using two-dimensional planar trigonometryAn angle c is computed using the formula for the cosine of an angle between twoincluded vectors (the perimeter point unit vector and the vector from SC to thecentre of the Earth) Angle y is computed using the Law of Sines Angle e=180degrees shy c shy y The scalar value of the perimeter point vector is computed using eand the Law of Cosines The scalar value is then multiplied by the perimeter pointunit vector to produce the three-dimensional point of intersection of the vector withEarthrsquos surface in R-Spacecraft The process is repeated independently for each ofthe eight perimeter point vectors Aside from its mathematical simplicity the valueof arcsine (computed in step 2) will exceed the normal range for the sin of an anglewhen the perimeter point vectors fail to intersect with the surface of the earth Asimple test based on this principle allows the program to correctly handle obliquephotos which image a portion of the horizon (see results for high oblique photographsin table 5)
The nal step in the process converts the eight earth intersection points from theR-Spacecraft to R-Earth The results are then converted to the standard geographiccoordinate system and displayed on the lsquoInput-Outputrsquo page of the spreadsheet
54 ExamplesAll three formulations were applied to the photographs included in this paper
and the results are compared in table 5 Formulation 1 gives too small a value fordistance across the photograph (D) for all but the most nadir shots and thus servesas an indicator of the best theoretical case but is not a good measure for a speci cphotograph For example the photograph of Lake Eyre taken from 276 km altitude( gure 2(a)) and the photograph of Limmen Bight ( gure 7) were closest to beingnadir views (oVsetslt68 km or tlt15deg table 5) For these photographs D calculatedusing Formulation 1 was similar to the minimum D calculated using Formulations2 and 3 For almost all other more oblique photographs Formulation 1 gave asigni cant underestimate of the distance covered in the photograph For gure 5(A)(the picture of Houston taken with a 40 mm lens) Formulation 1 did not give anunderestimate for D This is because Formulation 1 does not account for curvatureof the Earth in any way With this large eld of view assuming a at Earth in atedthe value of D above the minimum from calculations that included Earth curvature
A major diVerence between Formulations 2 and 3 is the ability to estimate pixelsizes (P) in both directions (along the principal line and perpendicular to the principalline) For the more oblique photographs the vertical estimate of D and pixel sizes ismuch larger than in the horizontal direction (eg the low oblique and high obliquephotographs of Hawaii gure 4 table 5)
For the area of Limmen Bight ( gure 7) table 5 illustrates the improvement inthe estimate of distance and pixel size that can be obtained by re-estimating thelocation of the PC with greater accuracy Centre points in the catalogued data are
Astronaut photography as digital data for remote sensing 4429
plusmn05deg of latitude and longitude When the centre point was re-estimated to plusmn002degwe determined that the photograph was not taken as obliquely as rst thought(change in the estimate of the look angle t from 169 to 128deg table 5) When theauxiliary point was added to the calculations of Formula 3 the calculated look angleshrank further to 110deg indicating that this photograph was taken at very close toa nadir view Of course this improvement in accuracy could have also led to estimatesof greater obliquity and corresponding larger pixel sizes
We also tested the performance of the scale calculator with auxiliary point byestimating the corner and point locations on the photograph using a 11 000 000Operational Navigational Chart For this test we estimated our ability to readcoordinates from the map as plusmn002degand our error in nding the locations of thecorner points as plusmn015deg (this error varies among photographs depending on thedetail that can be matched between photo and map) For Limmen Bight ( gure 7)the mean diVerence between map estimates and calculator estimates for four pointswas 031deg (SD=018 n=8) For a photograph of San Francisco Bay (STS062-151-291) the mean diVerence between map estimates and calculator estimates forfour points was 0064deg (SD=018 n=8) For a photograph of San Francisco Bay(STS062-151-291) the mean diVerence between map estimates and calculator estim-ates for eight points was 0196deg (SD=0146 n=16) Thus in one case the calculatorestimates were better than our estimate of the error in locating corner points on themap It is reasonable to expect that for nadir-viewing photographs the calculatorused with an auxiliary point can estimate locations of the edges of a photograph towithin plusmn03deg
55 Empirical con rmation of spatial resolution estimatesAs stated previously a challenge to estimating system-AWAR for astronaut
photography of Earth is the lack of suitable targets Small features in an image cansometimes be used as a check on the size of objects that can be successfully resolvedgiving an approximate value for GRD Similarly the number of pixels that make upthose features in the digitized image can be used to make an independent calculationof pixel size We have successfully used features such as roads and airport runwaysto make estimates of spatial scale and resolution (eg Robinson et al 2000c) Whilerecognizing that the use of linear detail in an image is a poor approximation to abar target and that linear objects smaller than the resolving power can often bedetected (Charman 1965) few objects other than roads could be found to make anydirect estimates of GRD Thus roads and runways were used in the images ofHouston (where one can readily conduct ground veri cations and where a numberof higher-contrast concrete roadways were available) to obtain empirical estimatesof GRD and pixel size for comparison with table 5
In the all-digital ESC image of Houston ( gure 3(d )) we examined EllingtonField runway 4-22 (centre left of the image) which is 24384 mtimes457 m This runwayis approximately 6ndash7 pixels in width and 304ndash3094 pixels in length so pixelsrepresent an distance on the ground 7ndash8 m Using a lower contrast measure of astreet length between two intersections (2125 m=211 pixels) a pixel width of 101 mis estimated These results compare favourably with the minimum estimate of 81 mpixels using Formulation 1 (table 5) For an estimate of GRD that would be morecomparable to aerial photography of a line target the smallest street without treecover that could be clearly distinguished on the photograph was 792 m wide Thesmallest non-street object (a gap between stages of the Saturn rocket on display in
J A Robinson et al4430
Tab
le5
App
lica
tio
nof
met
hod
sfo
res
tim
ati
ng
spati
alre
solu
tio
no
fast
ron
aut
ph
oto
gra
ph
sin
clu
din
gF
orm
ula
tion
s12
and
3Im
pro
ved
cen
tre
po
int
esti
mat
esan
dauxilia
ryp
oin
tca
lcu
lati
ons
are
sho
wn
for
gure
7V
aria
ble
sas
use
din
the
text
fle
ns
foca
lle
ngt
hH
al
titu
de
PC
ph
oto
gra
ph
cen
tre
poin
tSN
sp
ace
craft
nadir
po
int
D
dis
tance
cover
edo
nth
egr
oun
d
Pd
ista
nce
on
the
grou
nd
repre
sen
ted
by
apix
el
OVse
td
ista
nce
bet
wee
nP
Cand
SNusi
ng
the
grea
tci
rcle
form
ula
t
look
angle
away
from
SN
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
2L
ake
Eyre
ST
S093
-717
-80
250
276
shy29
0137
0shy
28
413
71
607
11
767
413
7609
ndash64
212
013
760
9ndash64
7121
124
ST
S082
-754
-61
250
617
shy28
5137
0shy
28
113
50
135
726
1201
18
013
8ndash14
827
517
9138ndash15
3277
294
Fig
ure
3H
ou
sto
nS
TS
062
-97-
151
4029
829
5shy
95
530
5shy
95
4410
78
8112
20
5348ndash58
990
1205
351
ndash63
8951
10
36
ST
S094
-737
-28
100
294
295
shy
95
528
3shy
95
5162
31
1133
24
415
8ndash20
334
724
3158ndash20
7351
390
ST
S055
-71-
41250
302
295
shy
95
028
2shy
93
766
412
8192
32
5736
ndash84
715
232
273
9ndash85
7154
185
S73
E51
45300
2685
295
shy
95
024
78
0F
igu
re4
Haw
aii
ST
S069
-701
-65
250
370
195
shy
155
520
4shy
153
881
415
7204
28
9876
ndash98
917
928
688
0ndash10
9181
210
ST
S069
-701
-70
250
370
210
shy
157
519
2shy
150
781
415
7738
63
414
9ndash23
336
760
7148ndash55
6394
10
63
ST
S069
-729
-27
100
398
200
shy
156
620
5shy
148
2219
42
1867
65
432
8ndash13
10
158
62
4323+
311
N
A
Astronaut photography as digital data for remote sensing 4431
Tab
le5
(Cont
inued
)
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
7L
imm
enB
igh
tS
TS
093
-704
-50
250
283
shy15
0135
0shy
15
013
58
623
120
859
16
963
0ndash67
3125
16
863
0ndash67
612
613
2P
CR
e-es
tim
ated
shy14
733
1352
67
64
5128
623
ndash65
5123
128
623
ndash65
712
3127
Auxilia
ryP
oin
t611
062
3ndash65
112
312
5F
igu
re10
N
ear
Au
stin
T
XS
TS
095
-716
-46
250
543
30
5shy
975
28
6shy
1007
120
230
375
34
613
5ndash157
281
34
1136ndash16
128
636
0
1 All
pho
togr
aphs
exce
pt
on
eta
ken
wit
hH
ass
elbla
dca
mer
aso
dis
tan
ceacr
oss
the
image
d=
55m
m
for
S73
E51
45
taken
wit
hel
ectr
onic
still
cam
erad=
276
5m
mho
rizo
nta
l2 P
Cand
SN
are
giv
enin
the
form
(lati
tud
elo
ngit
ude)
deg
rees
w
ith
neg
ativ
en
um
ber
sre
pre
senti
ng
south
and
wes
tA
ccura
cyo
fP
Cis
plusmn0
5and
SN
isplusmn
01
as
reco
rded
inth
eA
stro
naut
Ph
oto
gra
phy
Data
base
ex
cep
tw
her
en
ote
dth
atce
ntr
ep
oin
tw
asre
-est
imate
dw
ith
grea
ter
accu
racy
3 C
alc
ula
ted
usi
ng
the
mea
nofth
em
inim
um
an
dm
axim
um
esti
mate
sof
DF
orm
ula
tion
2co
nsi
der
sD
inth
ed
irec
tion
per
pen
dic
ula
rto
the
Pri
nci
pal
Lin
eo
nly
4 C
alc
ula
ted
inho
rizo
nta
lan
dve
rtic
aldir
ecti
on
sb
ut
still
ass
um
ing
geom
etry
of
gure
6(B
)F
irst
nu
mb
eris
the
mea
no
fth
em
inim
um
Dat
the
bott
om
of
the
ph
oto
and
the
maxi
mum
Dat
the
top
of
the
ph
oto
(range
show
nin
the
tab
le)
and
isco
mpara
ble
toF
orm
ula
tio
n2
Sec
ond
num
ber
isth
em
ean
of
Des
tim
ated
alon
gth
eP
rin
cip
alline
and
alo
ng
the
ou
tsid
eed
ge
(range
not
show
n)
5 Alt
itu
de
isap
pro
xim
ate
for
mis
sion
as
exac
tti
me
pho
togr
ap
hw
asta
ken
and
corr
esp
on
din
gn
adir
data
are
no
tava
ilab
le
Lack
of
nad
irdata
pre
ven
tsap
plica
tion
of
Form
ula
tion
s2
an
d3
6 Au
xilliar
ypo
int
add
edw
as
shy14
80
lati
tud
e13
51
2lo
ngit
ude
shy46
5degro
tati
on
angle
in
stru
ctio
ns
for
esti
mati
ng
the
rota
tio
nan
gle
para
met
erare
conta
ined
as
par
tof
the
scal
eca
lcu
lato
rsp
read
shee
tdo
cum
enta
tio
n
J A Robinson et al4432
a park at Johnson Space Center) that could clearly be distinguished on the photo-graph was 853 m wide
For the photograph of Houston taken with a 250 mm lens ( gure 3(C)) anddigitized from second generation lm at 2400 ppi (106 mm pixel Otilde 1 ) Ellington Fieldrunway 4-22 is 3 pixels in width and 1613 pixels in length so pixels represent151ndash152 m on the ground These results compare favourably with the minimumestimate of 152 m using Formulation 2 and 154ndash185 m pixels using Formulation 3(table 5) For an estimate of GRD using an 8times8 inch print (1369 enlargement) and4times magni cation the smallest street that could clearly be distinguished was 822 mwide the same feature could barely be distinguished on the digitized image
We also made an empirical estimate of spatial resolution for lower contrastvegetation boundaries By clearing forest so that a pattern would be visible to landingaircraft a landowner outside Austin Texas (see also aerial photo in Lisheron 2000)created a target that is also useful for evaluating spatial resolution of astronautphotographs The forest was selectively cleared in order to spell the landownerrsquosname lsquoLUECKErsquo with the remaining trees ( gure 10) According to local surveyorswho planned the clearing the plan was to create letters that were 3100 fttimes1700 ft(9449 mtimes5182 m) Photographed at a high altitude relative to most Shuttle missions(543 km) with a 250 mm lens Formula 3 predicts that each pixel would represent anarea 286 mtimes360 m on the ground (table 5) When original lm was digitized at2400 ppi (106 mm pixel Otilde 1 ) letters correspond to 294times188 pixels for a comparablepixel size of 27ndash32 m
6 Summary and conclusionsAstronaut photographs can be an excellent source of data for remote sensing
applications Best-case resolutions are similar to that for Landsat or SPOT withpixels as small as lt10 m It was estimated that of images taken to date 504 hadlens and obliquity characteristics that make them potential candidates for remotesensing information Digitized astronaut photographs can be overlaid with othersatellite data using GIS (Eckardt et al 2000) or used to ll in gaps in time serieswhen other imagery is not available As a source of public-domain information theycan be very useful for scientists who do not have access to satellite imagery eitherbecause of the expense of image acquisition (to the end user) or the computersystems needed for processing satellite images These diVerences in image acquisitioncosts (to the user) are summarized in table 6
Searching of the complete database of NASA astronaut photography includinglow-spatial-resolution browse images is available via the Web (OYce of EarthSciences 2000) This provides nearly global access for identifying images that willcontribute to a speci c scienti c project For the most detailed studies digitalproducts posted to the web will not be of suYcient quality To date we have madeit a practice of digitizing small numbers of images when requested by scientists atno charge Such requests (including a description of the project involved) can bemade through the authors or using the contact information listed on the websiteInvestigators with needs for larger numbers of digital images have also been servedthrough collaborative agreements
In our experience scientists that nd the data most valuable are those in develop-ing countries those studying areas that have not usually been targets for the majorsatellites those having diYculty in nding low-cloud images those interested inconstructing time series or those interested in using a large number of images
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
J A Robinson et al4416
Tab
le3
Det
ails
on
lm
suse
dfo
rast
ron
aut
pho
togr
aphy
of
Eart
h
Dat
ain
clud
ep
ho
togr
aphy
fro
mall
NA
SA
hum
an
spac
eig
ht
mis
sio
ns
thro
ugh
ST
S-9
6(J
un
e19
99
378
461
tota
lfr
ames
inth
edata
base
)F
ilm
sw
ith
lt50
00fr
am
esex
po
sed
and
lm
suse
dfo
r35
mm
cam
eras
only
are
no
tin
clud
ed
Res
olv
ing
Nu
mb
erof
Film
man
ufa
ctu
rer
AS
Ap
ow
er1
fram
esP
erio
dof
use
Cam
eras
Colo
ur
posi
tive
Kod
akA
eroch
rom
eII
(2447
2448
SO
-131S
O-3
68
)32
40ndash50
513
8196
8ndash19
85
Hass
elbla
dL
inh
of
Kod
akE
kta
chro
me
(5017
502
56
017
6118
SO
-121
64
50
113
932
196
5ndash19
68
Hass
elbla
dL
inh
of
Role
iex
SO
-217
Q
X-8
24
QX
-868
)19
81ndash1993
Mau
rer
Nik
on
Kod
akL
um
iere
(5046
5048
)100
105
743
199
4ndash19
98
Hass
elbla
dL
inho
fK
od
akE
lite
100
S(5
069
)100
41
486
199
6ndashp
rese
nt
Hass
elbla
d2
Fu
jiV
elvia
50C
S135
-36
32
13
882
1992ndash19
97H
ass
elbla
dN
iko
nK
od
akH
i-R
esolu
tio
nC
olo
r(S
O-2
42S
O-3
56
)10
096
74
1973ndash19
75H
ass
elbla
d
S19
0A
S190
B
Infr
are
dand
bla
ck-a
nd-w
hit
e
lms
Kod
akA
eroch
rom
eC
olo
rIn
frar
ed40
32
40
704
196
5ndashp
rese
nt2
Hass
elbla
dL
inh
of
Nik
on
S190
A
(8443
34
432443
)S
190B
Kod
akB
ampW
Infr
are
d(2
424
)32
10
553
197
3ndash19
74
S190
AK
od
akP
anato
mic
-XB
ampW
(SO
-022
)63
10
657
197
3ndash19
74
S190
A
1A
tlo
wco
ntr
ast
(ob
ject
back
grou
nd
rati
o16
1)
aspro
vid
edby
Ko
dak
and
list
edby
Sla
ter
etal
(1
983
256
)R
esolv
ing
pow
erw
as
dis
con
tin
ued
fro
mth
ete
chn
ical
test
sper
form
edb
yK
odak
inapp
roxim
atel
y19
93
an
dit
isn
ot
kn
ow
nif
curr
ent
lm
sh
ave
sim
ilar
reso
lvin
gpo
wer
too
lder
ver
sion
so
fsi
milar
lm
s(K
T
eite
lbaum
K
od
akp
erso
nal
com
mu
nic
atio
n)
Th
egra
nu
lari
tym
easu
reno
wpro
vid
edb
yK
od
ak
can
no
tb
ere
adily
con
ver
ted
toa
spat
ial
reso
luti
on
2 The
Lin
hof
was
use
dag
ain
on
ST
S-1
03in
Dec
emb
er19
99(d
ata
not
cata
logued
asof
the
pre
para
tion
of
this
tab
le)
usi
ng
Ko
dak
Elite
100S
lm
3 B
egin
nin
gin
July
1999
2443
colo
ur-
infr
are
d
lmpro
cess
was
chan
ged
from
EA
-5to
E-6
Astronaut photography as digital data for remote sensing 4417
413 Film duplicationThe original lm is archived permanently after producing about 20 duplicate
printing masters (second generation) Duplicate printing masters are disseminatedto regional archives and used to produce products (thirdndashfourth generation) for themedia the public and for scienti c use When prints slides transparencies anddigital products are produced for public distribution they are often colour adjustedto correspond to a more realistic look Digital products from second generationcopies can be requested by scienti c users (contact the authors for information onobtaining such products for a particular project) and these are recommended forremote sensing applications Care should be taken that the digital product acquiredfor remote sensing analysis has not been colour adjusted for presentation purposesBased on qualitative observations there is little visible increase in GRD in the thirdor fourth generation products There is a signi cant degradation in delity of colourin third and fourth generation duplicates because an increase in contrast occurswith every copy
42 Shutter speedsThe impact of camera motion (both due to the photographer and to the motion
of the spacecraft ) on image quality is determined by shutter speedmdash1250 to 1500second have been used because slower speeds record obvious blurring caused bythe rapid motion of the spacecraft relative to the surface of the Earth Generally1250 was used for ISO 64 lms (and slower) and after the switch to ISO 100 lmsa 1500 setting became standard New Hasselblad cameras that have been ownbeginning with STS-92 in October 2000 vary shutter speed using a focal plane shutterfor exposure bracketing (rather than varying aperture) and 1250 1500 and 11000second are used
Across an altitude range of 120ndash300 nautical miles (222ndash555 km) and orbitalinclinations of 285deg and 570deg the median relative ground velocity of orbitingspacecraft is 73 km s Otilde 1 At a nominal shutter speed of 1500 the expected blur dueto motion relative to the ground would be 146 m When cameras are handheld (asopposed to mounted in a bracket) blur can be reduced by physical compensationfor the motion (tracking) by the photographer many photographs show a level ofdetail indicating that blur at this level did not occur (eg gure 3(d )) Thus motionrelative to the ground is not an absolute barrier to ground resolution for handheldphotographs
43 Spacecraft windowsMost of the photographs in the NASA Astronaut Photography Database were
taken through window ports on the spacecraft used The transmittance of the windowport may be aVected by material inhomogeniety of the glass the number of layersof panes coatings used the quality of the surface polish environmentally inducedchanges in window materials (pressure loads thermal gradients) or deposited con-tamination Such degradation of the window cannot be corrected is diVerent foreach window and changes over time See Eppler et al (1996) and Scott (2000) fordiscussion of the spectral transmittance of the fused quartz window that is part ofthe US Laboratory Module (Destiny) of the International Space Station
44 Digitized images from lmWhen lm is scanned digitally the amount of information retained depends on
the spectral information extracted from the lm at the spatial limits of the scanner
J A Robinson et al4418
To date standard digitizing methodologies for astronaut photographs have not beenestablished and lm is digitized on a case-by-case basis using the equipment available
441 Digitizing and spatial resolutionLight (1993 1996) provided equations for determining the digitizing spatial
resolution needed to preserve spatial resolution of aerial photography based on thestatic resolving power (AWAR) for the system For lms where the manufacturer hasprovided data (Eastman Kodak 1998 K Teitelbaum personal communication) the resolving power of lms used for astronaut photography ranges from 32 lp mm Otilde 1to 100 lp mm Otilde 1 at low contrast (objectbackground ratio 161 table 3) The AWARfor the static case of the Hasselblad camera has been measured at high and lowcontrast (using lenses shown in table 1) with a maximum of approximately55 lp mm Otilde 1 (Fred Pearce unpublished data)
Based on the method of Light (1993 1996) the dimension of one spatial resolutionelement for a photograph with maximum static AWAR of 55 lp mm Otilde 1 would be18 mm lp Otilde 1 The acceptable range of spot size to preserve spatial information wouldthen be
18 mm
2 atilde 2aring scan spot size aring
18 mm
2(1)
and 6 mm aring scan spot size aring 9 mm Similarly for a more typical static AWAR of30 lp mm Otilde 1 (33 mm lpOtilde 1 low contrast Fred Pearce unpublished data) 11 mm aring scanspot sizearing 17 mm These scan spot sizes correspond to digitizing resolutions rangingfrom 4233ndash2822 ppi (pixels inch Otilde 1 ) for AWAR of 55 lp mm Otilde 1 and 2309ndash1494 ppi forAWAR of 30 lp mm Otilde 1 Scan spot sizes calculated for astronaut photography arecomparable to those calculated for the National Aerial Photography Program(9ndash13 mm Light 1996)
Widely available scanners that can digitize colour transparency lm currentlyhave a maximum spatial resolution of approximately 2400 ppi (106 mm pixel Otilde 1) Forexample we routinely use an Agfa Arcus II desktop scanner (see also Baltsavias1996) with 2400 ppi digitizing spatial resolution (2400 ppi optical resolution in onedirection and 1200 ppi interpolated to 2400 ppi in the other direction) and 2400 ppiwas used to calculate IFOV equivalents (tables 2 and 4) For some combinations oflens camera and contrast 2400 ppi will capture nearly all of the spatial informationcontained in the lm However for lm with higher resolving power thanEktachrome-64 for better lenses and for higher contrast targets digitizing at 2400 ppiwill not capture all of the spatial information in the lm
Improvements in digitizing technology will only produce real increases in IFOVto the limit of the AWAR of the photography system The incremental increase inspatial information above 3000 ppi (85 mm pixel Otilde 1 ) is not likely to outweigh thecosts of storing large images (Luman et al 1997) At 2400 ppi a Hasselblad frameis approximately 5200times5200 or 27 million pixels (table 2) while the same imagedigitized at 4000 ppi would contain 75 million pixels
Initial studies using astronaut photographs digitized at 2400 ppi (106 mm pixel Otilde 1 Webb et al 2000 Robinson et al 2000c) indicate that some GRD is lost comparedto photographic products Nevertheless the spatial resolution is still comparablewith other widely used data sources (Webb et al 2000) Robinson et al (2000c)found that digitizing at 21 mm pixel Otilde 1 provided information equivalent to 106 mmpixel Otilde 1 for identifying general urban area boundaries for six cities except for a single
Astronaut photography as digital data for remote sensing 4419
photo that required higher spatial resolution digitizing (it had been taken with ashorter lens and thus had less spatial resolution) Part of the equivalence observedin the urban areas study may be attributable to the fact that the atbed scannerused interpolates from 1200 ppi to 2400 ppi in one direction The appropriate digitiz-ing spatial resolution will in part depend on the scale of features of interest to theresearchers with a maximum upper limit set of a scan spot size of approximately6 mm (4233 ppi)
442 Digitizing and spectral resolutionWhen colour lm is digitized there will be loss of spectral resolution The three
lm emulsion layers (red green blue) have relatively distinct spectral responses butare fused together so that digitizers detect one colour signal and must convert itback into three (red green blue) digital channels Digital scanning is also subject tospectral calibration and reproduction errors Studies using digital astronaut photo-graphs to date have used 8 bits per channel However this is another parameterthat can be controlled when the lm is digitized We do not further address spectralresolution of digitally scanned images in this paper
45 External factors that in uence GRDAlthough not discussed in detail in this paper factors external to the spacecraft
and camera system (as listed by Moik (1980) for remote sensing in general) alsoimpact GRD These include atmospheric interference (due to scattering attenuationhaze) variable surface illumination (diVerences in terrain slope and orientation) andchange of terrain re ectance with viewing angle (bidirectional re ectance) For astro-naut photographs variable illumination is particularly important because orbits arenot sun-synchronous Photographs are illuminated by diVerent sun angles and imagesof a given location will have colour intensities that vary widely In addition theviewing angle has an eVect on the degree of object-to-backgroun d contrast andatmospheric interference
5 Estimating spatial resolution of astronaut photographsIn order to use astronaut photographs for digital remote sensing it is important
to be able to calculate the equivalent to an IFOVmdashthe ground area represented bya single pixel in a digitized orbital photograph The obliquity of most photographsmeans that pixel lsquosizesrsquo vary at diVerent places in an image Given a ground distanceD represented by a photograph in each direction (horizontal vertical ) an approxi-mate average pixel width (P the equivalent of IFOV) for the entire image can becalculated as follows
P=D
dS(2)
where D is the projected distance on the ground covered by the image in the samedirection as the pixel is measured d is the actual width of the image on the original lm (table 1) and S is the digitizing spatial resolution
Here we present three mathematical formulations for estimating the size of thefootprint or area on the ground covered by the image Example results from theapplication of all three formulations are given in table 5 The rst and simplestcalculation (formulation 1) gives an idea of the maximum spatial resolution attainableat a given altitude of orbit with a given lm format and a perfectly vertical (nadir)
J A Robinson et al4420
view downward Formulation 2 takes into account obliquity by calculating lookangle from the diVerence between the location on the ground represented at thecentre of the photograph and the nadir location of the spacecraft at the time thephotograph was taken ( gure 1) Formulation 3 describes an alternate solution tothe oblique look angle problem using coordinate-system transformations Thisformulation has been implemented in a documented spreadsheet and is availablefor download (httpeoljscnasagovsseopFootprintCalculatorxls) and is in theprocess of being implemented on a Web-based user interface to the AstronautPhotography Database (OYce of Earth Sciences 2000)
Although formulations 2 and 3 account for obliquity for the purposes of calcula-tion they treat the position of the spacecraft and position of the camera as one Inactuality astronauts are generally holding the cameras by hand (although camerasbracketed in the window are also used) and the selection of window position of theastronaut in the window and rotation of the camera relative to the movement ofthe spacecraft are not known Thus calculations using only the photo centre point(PC) and spacecraft nadir point (SN) give a locator ellipse and not the locations ofthe corners of the photograph A locator ellipse describes an estimated area on theground that is likely to be included in a speci c photograph regardless of the rotationof the lm plane about the camerarsquos optical axis ( gure 6)
Estimating the corner positions of the photo requires additional user input of asingle auxiliary pointmdasha location on the image that has a known location on theground Addition of this auxiliary point is an option available to users of thespreadsheet An example of the results of adding an auxiliary point is shown in gure 7 with comparisons of the various calculations in table 5
51 Formulation 1 Footprint for a nadir viewThe simplest way to estimate footprint size is to use the geometry of camera lens
and spacecraft altitude to calculate the scaling relationship between the image in the
Figure 6 Sketch representation of the locator ellipse an estimated area on the ground thatis likely to be included in a speci c photograph regardless of the rotation of the lmplane about the camerarsquos optical axis
Astronaut photography as digital data for remote sensing 4421
Figure 7 An example of the application of the Low Oblique Space Photo FootprintCalculator The photograph (STS093-704-50) of Limmen Bight Australia was takenfrom an altitude of 283 km using the Hasselblad camera with a 250 mm lens Mapused is 11 000 000 Operational Navigational Chart The distances shown equate toan average pixel size of 124 m for this photograph
lm and the area covered on the ground For a perfect nadir view the scalerelationship from the geometry of similar triangles is
d
D=
f
H(3)
where d=original image size D=distance of footprint on the ground f =focallength of lens and H=altitude Once D is known pixel size ( length=width) can becalculated from equation (2) These calculations represent the minimum footprintand minimum pixel size possible for a given camera system altitude and digitisingspatial resolution (table 2) Formulation 1 was used for calculating minimum pixelsizes shown in tables 2 and 4
By assuming digitizing at 2400 ppi (106 mm pixel Otilde 1 ) currently a spatial resolutioncommonly attainable from multipurpose colour transparency scanners (see sect441)this formulation was used to convert area covered to an IFOV equivalent for missionsof diVerent altitudes (table 2) Table 4 provides a comparison of IFOV of imagesfrom various satellites including the equivalent for astronaut photography Valuesin this table were derived by using formulation 1 to estimate the area covered becausea perfect nadir view represents the best possible spatial resolution and smallest eldof view that could be obtained
52 Formulation 2 Footprint for oblique views using simpli ed geometry and thegreat circle distance
A more realistic approach to determining the footprint of the photographaccounts for the fact that the camera is not usually pointing perfectly down at the
J A Robinson et al4422
Table 4 Comparison of pixel sizes (instantaneous elds of view) for satellite remote sensorswith pixel sizes for near vertical viewing photography from spacecraft Note that alldata returned from the automated satellites have the same eld of view while indicatedpixel sizes for astronaut photography are best-case for near vertical look angles See gure 5 for distributions of astronaut photographs of various look angles High-altitude aerial photography is also included for reference
Altitude PixelInstrument (km) width (mtimesm) Bands1
Landsat Multipectral Scanner2 (MSS 1974-) 880ndash940 79times56 4ndash5Landsat Thematic Mapper (TM 1982-) ~705 30times30 7Satellite pour lrsquoObservation de la Terre (SPOT 1986-) ~832
SPOT 1-3 Panchromatic 10times10 1SPOT 1-3 Multispectral (XS) 20times20 3
Astronaut Photography (1961-)Skylab3 (1973-1974 ) ~435
Multispectral Camera (6 camera stations) 17ndash19times17ndash19 3ndash8Earth Terrain Camera (hi-res colour) 875times875 3
Space Shuttle5 (1981-) 222ndash611Hasselblad 70 mm (250mm lens colour) vertical view 9ndash26times9ndash26 3Hasselblad 70 mm (100mm lens colour) vertical view 23ndash65times23ndash65 3Nikon 35 mm (400mm lens colour lm or ESC) vertical 5ndash16times5ndash16 3
High Altitude Aerial Photograph (1100 000) ~12 2times2 3
1Three bands listed for aerial photography obtainable by high-resolution colour digitizingFor high-resolution colour lm at 65 lp mmOtilde 1 (K Teitelbaum Eastman Kodak Co personalcommunication) the maximum information contained would be 3302 ppi
2Data from Simonett (1983Table 1-2)3Calculated as recommended by Jensen (19831596) the dividend of estimated ground
resolution at low contrast given in NASA (1974179-180) and 244Six lters plus colour and colour infrared lm (NASA 19747) 5Extracted from table 2
nadir point The look angle (the angle oV nadir that the camera is pointing) can becalculated trigonometrically by assuming a spherical earth and calculating the dis-tance between the coordinates of SN and PC ( gure 1) using the Great Circledistance haversine solution (Sinnott 1984 Snyder 198730ndash32 Chamberlain 1996)The diVerence between the spacecraft centre and nadir point latitudes Dlat=lat 2 shy lat 1 and the diVerence between the spacecraft centre and nadir pointlongitudes Dlon=lon 2 shy lon 1 enter the following equations
a=sin2ADlat
2 B+cos(lat 1) cos(lat 2) sin2ADlon
2 B (4)
c=2 arcsin(min[1 atilde a]) (5)
and oVset=R c with R the radius of a spherical Earth=6370 kmThe look angle (t ) is then given by
t=arctanAoffset
H B (6)
Assuming that the camera was positioned so that the imaginary line between thecentre and nadir points (the principal line) runs vertically through the centre of thephotograph the distance between the geometric centre of the photograph (principalpoint PP) and the top of the photograph is d2 ( gure 8(a)) The scale at any point
Astronaut photography as digital data for remote sensing 4423
(a) (b)
Figure 8 The geometric relationship between look angle lens focal length and the centrepoint and nadir points of a photograph (see also Estes and Simonett 1975) Note thatthe Earthrsquos surface and footprint represented in gure 1 are not shown here only arepresentation of the image area of the photograph Variables shown in the gurelook angle t focal length f PP centre of the image on the lm SN spacecraft nadird original image size (a) The special case where the camera is aligned squarely withthe isometric parallel In this case tilt occurs along a plane parallel with the centre ofthe photograph When the conditions are met calculations using Formulation 3 willgive the ground coordinates of the corner points of the photograph (b) The generalrelationship between these parameters and key photogrammetric points on the photo-graph For this more typical case coordinates of an additional point in the photographmust be input in order to adjust for twist of the camera and to nd the corner pointcoordinates
in the photograph varies as a function of the distance y along the principal linebetween the isocentre and the point according to the relationship
d
D=
f shy ysint
H(7)
(Wong 1980 equation (214) H amp the ground elevation h) Using gure 8(a) at thetop of the photo
y= f tanAt
2B+d
2(8)
and at the bottom of the photo
y= f tanAt
2Bshyd
2(9)
Thus for given PC and SN coordinates and assuming a photo orientation as in gure 8(a) and not gure 8(b) we can estimate a minimum D (using equations (7)and (8)) and a maximum D (using equations (7) and (9)) and then average the twoto determine the pixel size (P ) via equation (2)
J A Robinson et al4424
53 Formulation 3 T he L ow Oblique Space Photo Footprint CalculatorThe Low Oblique Space Photo Footprint Calculator was developed to provide
a more accurate estimation of the geographic coordinates for the footprint of a lowoblique photo of the Earthrsquos surface taken from a human-occupied spacecraft inorbit The calculator performs a series of three-dimensional coordinate transforma-tions to compute the location and orientation of the centre of the photo exposureplane relative to an earth referenced coordinate system The nominal camera focallength is then used to create a vector from the photorsquos perspective point througheach of eight points around the parameter of the image as de ned by the formatsize The geographical coordinates for the photo footprint are then computed byintersecting these photo vectors with a spherical earth model Although more sophist-icated projection algorithms are available no signi cant increase in the accuracy ofthe results would be produced by these algorithms due to inherent uncertainties inthe available input data (ie the spacecraft altitude photo centre location etc)
The calculations were initially implemented within a Microsoft Excel workbookwhich allowed one to embed the mathematical and graphical documentation nextto the actual calculations Thus interested users can inspect the mathematical pro-cessing A set of error traps was also built into the calculations to detect erroneousresults A summary of any errors generated is reported to the user with the resultsof the calculations Interested individuals are invited to download the Excel work-book from httpeoljscnasagovsseopFootprintCalculatorxls The calculations arecurrently being encoded in a high-level programming language and should soon beavailable alongside other background data provided for each photograph at OYceof Earth Sciences (2000)
For the purposes of these calculations a low oblique photograph was de ned as
one with the centre within 10 degrees of latitude and longitude of the spacecraftnadir point For the typical range of spacecraft altitudes to date this restricted thecalculations to photographs in which Earthrsquos horizon does not appear (the generalde nition of a low oblique photograph eg Campbell 199671 and gure 4)
531 Input data and resultsUpon opening the Excel workbook the user is presented with a program intro-
duction providing instructions for using the calculator The second worksheet tablsquoHow-To-Usersquo provides detailed step-by-step instructions for preparing the baselinedata for the program The third tab lsquoInput-Output rsquo contains the user input eldsand displays the results of the calculations The additional worksheets contain theactual calculations and program documentation Although users are welcome toreview these sheets an experienced user need only access the lsquoInput-Outputrsquo
spreadsheetTo begin a calculation the user enters the following information which is available
for each photo in the NASA Astronaut Photography Database (1) SN geographicalcoordinates of spacecraft nadir position at the time of photo (2) H spacecraftaltitude (3) PC the geographical coordinates of the centre of the photo (4) f nominal focal length and (5) d image format size The automatic implementationof the workbook on the web will automatically enter these values and completecalculations
For more accurate results the user may optionally enter the geographic co-ordinates and orientation for an auxiliary point on the photo which resolves thecamerarsquos rotation uncertainty about the optical axis The auxiliary point data must
Astronaut photography as digital data for remote sensing 4425
be computed by the user following the instruction contained in the lsquoHow-To-Usersquotab of the spreadsheet
After entering the input data the geographic coordinates of the photo footprint(ie four photo corner points four points at the bisector of each edge and the centreof the photo) are immediately displayed below the input elds along with any errormessages generated by the user input or by the calculations ( gure 7) Although
results are computed and displayed they should not be used when error messagesare produced by the program The program also computes the tilt angle for each ofthe photo vectors relative to the spacecraft nadir vector To the right of the photofootprint coordinates is displayed the arc distance along the surface of the sphere
between adjacent computed points
532 Calculation assumptions
The mathematical calculations implemented in the Low Oblique Space PhotoFootprint Calculator use the following assumptions
1 The SN location is used as exact even though the true value may vary by upto plusmn01 degree from the location provided with the photo
2 The spacecraft altitude is used as exact Although the determination of the
nadir point at the instant of a known spacecraft vector is relatively precise(plusmn115times10 Otilde 4 degrees) the propagator interpolates between sets of approxi-mately 10ndash40 known vectors per day and the time code recorded on the lmcan drift Thus the true value for SN may vary by up to plusmn01 degree fromthe value provided with the photo
3 The perspective centre of the camera is assumed to be at the given altitudeover the speci ed spacecraft nadir location at the time of photo exposure
4 The PC location is used as exact even though the true value may vary by upto plusmn05deg latitude and plusmn05deg longitude from the location provided with the
photo5 A spherical earth model is used with a nominal radius of 6 372 16154 m (a
common rst-order approximation for a spherical earth used in geodeticcomputations)
6 The nominal lens focal length of the camera lens is used in the computations(calibrated focal length values are not available)
7 The photo projection is based on the classic pin-hole camera model8 No correction for lens distortion or atmospheric re ection is made
9 If no auxiliary point data is provided the lsquoTop of the Imagersquo is orientedperpendicular to the vector from SN towards PC
533 T ransformation from Earth to photo coordinate systemsThe calculations begin by converting the geographic coordinates ( latitude and
longitude) of the SN and PC to a Rectangular Earth-Centred Coordinate System(R-Earth) de ned as shown in gure 9 (with the centre of the Earth at (0 0 0))
Using the vector from the Earthrsquos centre through SN and the spacecraft altitude thespacecraft location (SC) is also computed in R-Earth
For ease of computation a Rectangular Spacecraft-Centred Coordinate System(R-Spacecraft ) is de ned as shown in gure 9 The origin of R-Spacecraft is located
at SC with its +Z-axis aligned with vector from the centre of the Earth throughSN and its +X-axis aligned with the vector from SN to PC ( gure 9) The speci c
J A Robinson et al4426
Figure 9 Representation of the area included in a photograph as a geometric projectiononto the Earthrsquos surface Note that the focal length (the distance from the SpaceShuttle to the photo) is grossly overscale compared to the altitude (the distance fromthe Shuttle to the surface of the Earth
rotations and translation used to convert from the R-Earth to the R-Spacecraft arecomputed and documented in the spreadsheet
With the mathematical positions of the SC SN PC and the centre of the Earthcomputed in R-Spacecraft the program next computes the location of the camerarsquosprincipal point (PP) The principle point is the point of intersection of the opticalaxis of the lens with the image plane (ie the lm) It is nominally positioned at a
Astronaut photography as digital data for remote sensing 4427
distance equal to the focal length from the perspective centre of the camera (whichis assumed to be at SC) along the vector from PC through SC as shown in gure 9
A third coordinate system the Rectangular Photo Coordinate System (R-Photo)is created with its origin at PP its XndashY axial plane normal to the vector from PCthrough SC and its +X axis aligned with the +X axis of R-Spacecraft as shownin gure 9 The XndashY plane of this coordinate system represents the image plane ofthe photograph
534 Auxiliary point calculationsThe calculations above employ a critical assumption that all photos are taken
with the lsquotop of the imagersquo oriented perpendicular to the vector from the SN towardsPC as shown in gure 9 To avoid non-uniform solar heating of the external surfacemost orbiting spacecraft are slowly and continually rotated about one or more axesIn this condition a ight crew member taking a photo while oating in microgravitycould orient the photo with practically any orientation relative to the horizon (seealso gure 8(b)) Unfortunately since these photos are taken with conventionalhandheld cameras there is no other information available which can be used toresolve the photorsquos rotational ambiguity about the optical axis other then the photoitself This is why the above assumption is used and the footprint computed by thiscalculator is actually a lsquolocator ellipsersquo which estimates the area on the ground whatis likely to be included in a speci c photograph (see gure 6) This locator ellipse ismost accurate for square image formats and subject to additional distortion as thephotograph format becomes more rectangular
If the user wants a more precise calculation of the image footprint the photorsquosrotational ambiguity about the optical axis must be resolved This can be done inthe calculator by adding data for an auxiliary point Detailed instructions regardinghow to prepare and use auxiliary point data in the computations are included in thelsquoHow-To-Usersquo tab of the spreadsheet Basically the user determines which side ofthe photograph is top and then measures the angle between the line from PP to thetop of the photo and from PP to the auxiliary point on the photo ( gure 9)
If the user includes data for an auxiliary point a series of computations arecompleted to resolve the photo rotation ambiguity about the optical axis (ie the
+Z axis in R-Photo) A vector from the Auxiliary Point on the Earth (AE) throughthe photograph perspective centre ( located at SC) is intersected with the photo imageplane (XndashY plane of R-Photo) to compute the coordinates of the Auxiliary Point onthe photo (AP) in R-Photo A two-dimensional angle in the XndashY plane of R-Photofrom the shy X axis to a line from PP to AP is calculated as shown in gure 9 The
shy X axis is used as the origin of the angle since it represents the top of the photoonce it passes through the perspective centre The diVerence between the computedangle and the angle measured by the user on the photo resolves the ambiguity inthe rotation of the photo relative to the principal line ( gures 7 and 9) Thetransformations from R-Spacecraft and R-Photo are then modi ed to include anadditional rotation angle about the +Z axis in R-Photo
535 lsquoFootprint rsquo calculationsThe program next computes the coordinates of eight points about the perimeter
of the image format (ie located at the four photo corners plus a bisector pointalong each edge of the image) These points are identi ed in R-Photo based uponthe photograph format size and then converted to R-Spacecraft Since all computa-tions are done in orthogonal coordinate systems the R-Spacecraft to R-Photo
J A Robinson et al4428
rotation matrix is transposed to produce an R-Photo to R-Spacecraft rotation matrixOnce in R-Spacecraft a unit vector from each of the eight perimeter points throughthe photo perspective centre (the same point as SC) is computed This provides thecoordinates for points about the perimeter of the image format with their directionvectors in a common coordinate system with other key points needed to computethe photo footprint
The next step is to compute the point of intersection between the spherical earthmodel and each of the eight perimeter point vectors The scalar value for eachperimeter point unit vector is computed using two-dimensional planar trigonometryAn angle c is computed using the formula for the cosine of an angle between twoincluded vectors (the perimeter point unit vector and the vector from SC to thecentre of the Earth) Angle y is computed using the Law of Sines Angle e=180degrees shy c shy y The scalar value of the perimeter point vector is computed using eand the Law of Cosines The scalar value is then multiplied by the perimeter pointunit vector to produce the three-dimensional point of intersection of the vector withEarthrsquos surface in R-Spacecraft The process is repeated independently for each ofthe eight perimeter point vectors Aside from its mathematical simplicity the valueof arcsine (computed in step 2) will exceed the normal range for the sin of an anglewhen the perimeter point vectors fail to intersect with the surface of the earth Asimple test based on this principle allows the program to correctly handle obliquephotos which image a portion of the horizon (see results for high oblique photographsin table 5)
The nal step in the process converts the eight earth intersection points from theR-Spacecraft to R-Earth The results are then converted to the standard geographiccoordinate system and displayed on the lsquoInput-Outputrsquo page of the spreadsheet
54 ExamplesAll three formulations were applied to the photographs included in this paper
and the results are compared in table 5 Formulation 1 gives too small a value fordistance across the photograph (D) for all but the most nadir shots and thus servesas an indicator of the best theoretical case but is not a good measure for a speci cphotograph For example the photograph of Lake Eyre taken from 276 km altitude( gure 2(a)) and the photograph of Limmen Bight ( gure 7) were closest to beingnadir views (oVsetslt68 km or tlt15deg table 5) For these photographs D calculatedusing Formulation 1 was similar to the minimum D calculated using Formulations2 and 3 For almost all other more oblique photographs Formulation 1 gave asigni cant underestimate of the distance covered in the photograph For gure 5(A)(the picture of Houston taken with a 40 mm lens) Formulation 1 did not give anunderestimate for D This is because Formulation 1 does not account for curvatureof the Earth in any way With this large eld of view assuming a at Earth in atedthe value of D above the minimum from calculations that included Earth curvature
A major diVerence between Formulations 2 and 3 is the ability to estimate pixelsizes (P) in both directions (along the principal line and perpendicular to the principalline) For the more oblique photographs the vertical estimate of D and pixel sizes ismuch larger than in the horizontal direction (eg the low oblique and high obliquephotographs of Hawaii gure 4 table 5)
For the area of Limmen Bight ( gure 7) table 5 illustrates the improvement inthe estimate of distance and pixel size that can be obtained by re-estimating thelocation of the PC with greater accuracy Centre points in the catalogued data are
Astronaut photography as digital data for remote sensing 4429
plusmn05deg of latitude and longitude When the centre point was re-estimated to plusmn002degwe determined that the photograph was not taken as obliquely as rst thought(change in the estimate of the look angle t from 169 to 128deg table 5) When theauxiliary point was added to the calculations of Formula 3 the calculated look angleshrank further to 110deg indicating that this photograph was taken at very close toa nadir view Of course this improvement in accuracy could have also led to estimatesof greater obliquity and corresponding larger pixel sizes
We also tested the performance of the scale calculator with auxiliary point byestimating the corner and point locations on the photograph using a 11 000 000Operational Navigational Chart For this test we estimated our ability to readcoordinates from the map as plusmn002degand our error in nding the locations of thecorner points as plusmn015deg (this error varies among photographs depending on thedetail that can be matched between photo and map) For Limmen Bight ( gure 7)the mean diVerence between map estimates and calculator estimates for four pointswas 031deg (SD=018 n=8) For a photograph of San Francisco Bay (STS062-151-291) the mean diVerence between map estimates and calculator estimates forfour points was 0064deg (SD=018 n=8) For a photograph of San Francisco Bay(STS062-151-291) the mean diVerence between map estimates and calculator estim-ates for eight points was 0196deg (SD=0146 n=16) Thus in one case the calculatorestimates were better than our estimate of the error in locating corner points on themap It is reasonable to expect that for nadir-viewing photographs the calculatorused with an auxiliary point can estimate locations of the edges of a photograph towithin plusmn03deg
55 Empirical con rmation of spatial resolution estimatesAs stated previously a challenge to estimating system-AWAR for astronaut
photography of Earth is the lack of suitable targets Small features in an image cansometimes be used as a check on the size of objects that can be successfully resolvedgiving an approximate value for GRD Similarly the number of pixels that make upthose features in the digitized image can be used to make an independent calculationof pixel size We have successfully used features such as roads and airport runwaysto make estimates of spatial scale and resolution (eg Robinson et al 2000c) Whilerecognizing that the use of linear detail in an image is a poor approximation to abar target and that linear objects smaller than the resolving power can often bedetected (Charman 1965) few objects other than roads could be found to make anydirect estimates of GRD Thus roads and runways were used in the images ofHouston (where one can readily conduct ground veri cations and where a numberof higher-contrast concrete roadways were available) to obtain empirical estimatesof GRD and pixel size for comparison with table 5
In the all-digital ESC image of Houston ( gure 3(d )) we examined EllingtonField runway 4-22 (centre left of the image) which is 24384 mtimes457 m This runwayis approximately 6ndash7 pixels in width and 304ndash3094 pixels in length so pixelsrepresent an distance on the ground 7ndash8 m Using a lower contrast measure of astreet length between two intersections (2125 m=211 pixels) a pixel width of 101 mis estimated These results compare favourably with the minimum estimate of 81 mpixels using Formulation 1 (table 5) For an estimate of GRD that would be morecomparable to aerial photography of a line target the smallest street without treecover that could be clearly distinguished on the photograph was 792 m wide Thesmallest non-street object (a gap between stages of the Saturn rocket on display in
J A Robinson et al4430
Tab
le5
App
lica
tio
nof
met
hod
sfo
res
tim
ati
ng
spati
alre
solu
tio
no
fast
ron
aut
ph
oto
gra
ph
sin
clu
din
gF
orm
ula
tion
s12
and
3Im
pro
ved
cen
tre
po
int
esti
mat
esan
dauxilia
ryp
oin
tca
lcu
lati
ons
are
sho
wn
for
gure
7V
aria
ble
sas
use
din
the
text
fle
ns
foca
lle
ngt
hH
al
titu
de
PC
ph
oto
gra
ph
cen
tre
poin
tSN
sp
ace
craft
nadir
po
int
D
dis
tance
cover
edo
nth
egr
oun
d
Pd
ista
nce
on
the
grou
nd
repre
sen
ted
by
apix
el
OVse
td
ista
nce
bet
wee
nP
Cand
SNusi
ng
the
grea
tci
rcle
form
ula
t
look
angle
away
from
SN
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
2L
ake
Eyre
ST
S093
-717
-80
250
276
shy29
0137
0shy
28
413
71
607
11
767
413
7609
ndash64
212
013
760
9ndash64
7121
124
ST
S082
-754
-61
250
617
shy28
5137
0shy
28
113
50
135
726
1201
18
013
8ndash14
827
517
9138ndash15
3277
294
Fig
ure
3H
ou
sto
nS
TS
062
-97-
151
4029
829
5shy
95
530
5shy
95
4410
78
8112
20
5348ndash58
990
1205
351
ndash63
8951
10
36
ST
S094
-737
-28
100
294
295
shy
95
528
3shy
95
5162
31
1133
24
415
8ndash20
334
724
3158ndash20
7351
390
ST
S055
-71-
41250
302
295
shy
95
028
2shy
93
766
412
8192
32
5736
ndash84
715
232
273
9ndash85
7154
185
S73
E51
45300
2685
295
shy
95
024
78
0F
igu
re4
Haw
aii
ST
S069
-701
-65
250
370
195
shy
155
520
4shy
153
881
415
7204
28
9876
ndash98
917
928
688
0ndash10
9181
210
ST
S069
-701
-70
250
370
210
shy
157
519
2shy
150
781
415
7738
63
414
9ndash23
336
760
7148ndash55
6394
10
63
ST
S069
-729
-27
100
398
200
shy
156
620
5shy
148
2219
42
1867
65
432
8ndash13
10
158
62
4323+
311
N
A
Astronaut photography as digital data for remote sensing 4431
Tab
le5
(Cont
inued
)
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
7L
imm
enB
igh
tS
TS
093
-704
-50
250
283
shy15
0135
0shy
15
013
58
623
120
859
16
963
0ndash67
3125
16
863
0ndash67
612
613
2P
CR
e-es
tim
ated
shy14
733
1352
67
64
5128
623
ndash65
5123
128
623
ndash65
712
3127
Auxilia
ryP
oin
t611
062
3ndash65
112
312
5F
igu
re10
N
ear
Au
stin
T
XS
TS
095
-716
-46
250
543
30
5shy
975
28
6shy
1007
120
230
375
34
613
5ndash157
281
34
1136ndash16
128
636
0
1 All
pho
togr
aphs
exce
pt
on
eta
ken
wit
hH
ass
elbla
dca
mer
aso
dis
tan
ceacr
oss
the
image
d=
55m
m
for
S73
E51
45
taken
wit
hel
ectr
onic
still
cam
erad=
276
5m
mho
rizo
nta
l2 P
Cand
SN
are
giv
enin
the
form
(lati
tud
elo
ngit
ude)
deg
rees
w
ith
neg
ativ
en
um
ber
sre
pre
senti
ng
south
and
wes
tA
ccura
cyo
fP
Cis
plusmn0
5and
SN
isplusmn
01
as
reco
rded
inth
eA
stro
naut
Ph
oto
gra
phy
Data
base
ex
cep
tw
her
en
ote
dth
atce
ntr
ep
oin
tw
asre
-est
imate
dw
ith
grea
ter
accu
racy
3 C
alc
ula
ted
usi
ng
the
mea
nofth
em
inim
um
an
dm
axim
um
esti
mate
sof
DF
orm
ula
tion
2co
nsi
der
sD
inth
ed
irec
tion
per
pen
dic
ula
rto
the
Pri
nci
pal
Lin
eo
nly
4 C
alc
ula
ted
inho
rizo
nta
lan
dve
rtic
aldir
ecti
on
sb
ut
still
ass
um
ing
geom
etry
of
gure
6(B
)F
irst
nu
mb
eris
the
mea
no
fth
em
inim
um
Dat
the
bott
om
of
the
ph
oto
and
the
maxi
mum
Dat
the
top
of
the
ph
oto
(range
show
nin
the
tab
le)
and
isco
mpara
ble
toF
orm
ula
tio
n2
Sec
ond
num
ber
isth
em
ean
of
Des
tim
ated
alon
gth
eP
rin
cip
alline
and
alo
ng
the
ou
tsid
eed
ge
(range
not
show
n)
5 Alt
itu
de
isap
pro
xim
ate
for
mis
sion
as
exac
tti
me
pho
togr
ap
hw
asta
ken
and
corr
esp
on
din
gn
adir
data
are
no
tava
ilab
le
Lack
of
nad
irdata
pre
ven
tsap
plica
tion
of
Form
ula
tion
s2
an
d3
6 Au
xilliar
ypo
int
add
edw
as
shy14
80
lati
tud
e13
51
2lo
ngit
ude
shy46
5degro
tati
on
angle
in
stru
ctio
ns
for
esti
mati
ng
the
rota
tio
nan
gle
para
met
erare
conta
ined
as
par
tof
the
scal
eca
lcu
lato
rsp
read
shee
tdo
cum
enta
tio
n
J A Robinson et al4432
a park at Johnson Space Center) that could clearly be distinguished on the photo-graph was 853 m wide
For the photograph of Houston taken with a 250 mm lens ( gure 3(C)) anddigitized from second generation lm at 2400 ppi (106 mm pixel Otilde 1 ) Ellington Fieldrunway 4-22 is 3 pixels in width and 1613 pixels in length so pixels represent151ndash152 m on the ground These results compare favourably with the minimumestimate of 152 m using Formulation 2 and 154ndash185 m pixels using Formulation 3(table 5) For an estimate of GRD using an 8times8 inch print (1369 enlargement) and4times magni cation the smallest street that could clearly be distinguished was 822 mwide the same feature could barely be distinguished on the digitized image
We also made an empirical estimate of spatial resolution for lower contrastvegetation boundaries By clearing forest so that a pattern would be visible to landingaircraft a landowner outside Austin Texas (see also aerial photo in Lisheron 2000)created a target that is also useful for evaluating spatial resolution of astronautphotographs The forest was selectively cleared in order to spell the landownerrsquosname lsquoLUECKErsquo with the remaining trees ( gure 10) According to local surveyorswho planned the clearing the plan was to create letters that were 3100 fttimes1700 ft(9449 mtimes5182 m) Photographed at a high altitude relative to most Shuttle missions(543 km) with a 250 mm lens Formula 3 predicts that each pixel would represent anarea 286 mtimes360 m on the ground (table 5) When original lm was digitized at2400 ppi (106 mm pixel Otilde 1 ) letters correspond to 294times188 pixels for a comparablepixel size of 27ndash32 m
6 Summary and conclusionsAstronaut photographs can be an excellent source of data for remote sensing
applications Best-case resolutions are similar to that for Landsat or SPOT withpixels as small as lt10 m It was estimated that of images taken to date 504 hadlens and obliquity characteristics that make them potential candidates for remotesensing information Digitized astronaut photographs can be overlaid with othersatellite data using GIS (Eckardt et al 2000) or used to ll in gaps in time serieswhen other imagery is not available As a source of public-domain information theycan be very useful for scientists who do not have access to satellite imagery eitherbecause of the expense of image acquisition (to the end user) or the computersystems needed for processing satellite images These diVerences in image acquisitioncosts (to the user) are summarized in table 6
Searching of the complete database of NASA astronaut photography includinglow-spatial-resolution browse images is available via the Web (OYce of EarthSciences 2000) This provides nearly global access for identifying images that willcontribute to a speci c scienti c project For the most detailed studies digitalproducts posted to the web will not be of suYcient quality To date we have madeit a practice of digitizing small numbers of images when requested by scientists atno charge Such requests (including a description of the project involved) can bemade through the authors or using the contact information listed on the websiteInvestigators with needs for larger numbers of digital images have also been servedthrough collaborative agreements
In our experience scientists that nd the data most valuable are those in develop-ing countries those studying areas that have not usually been targets for the majorsatellites those having diYculty in nding low-cloud images those interested inconstructing time series or those interested in using a large number of images
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
Astronaut photography as digital data for remote sensing 4417
413 Film duplicationThe original lm is archived permanently after producing about 20 duplicate
printing masters (second generation) Duplicate printing masters are disseminatedto regional archives and used to produce products (thirdndashfourth generation) for themedia the public and for scienti c use When prints slides transparencies anddigital products are produced for public distribution they are often colour adjustedto correspond to a more realistic look Digital products from second generationcopies can be requested by scienti c users (contact the authors for information onobtaining such products for a particular project) and these are recommended forremote sensing applications Care should be taken that the digital product acquiredfor remote sensing analysis has not been colour adjusted for presentation purposesBased on qualitative observations there is little visible increase in GRD in the thirdor fourth generation products There is a signi cant degradation in delity of colourin third and fourth generation duplicates because an increase in contrast occurswith every copy
42 Shutter speedsThe impact of camera motion (both due to the photographer and to the motion
of the spacecraft ) on image quality is determined by shutter speedmdash1250 to 1500second have been used because slower speeds record obvious blurring caused bythe rapid motion of the spacecraft relative to the surface of the Earth Generally1250 was used for ISO 64 lms (and slower) and after the switch to ISO 100 lmsa 1500 setting became standard New Hasselblad cameras that have been ownbeginning with STS-92 in October 2000 vary shutter speed using a focal plane shutterfor exposure bracketing (rather than varying aperture) and 1250 1500 and 11000second are used
Across an altitude range of 120ndash300 nautical miles (222ndash555 km) and orbitalinclinations of 285deg and 570deg the median relative ground velocity of orbitingspacecraft is 73 km s Otilde 1 At a nominal shutter speed of 1500 the expected blur dueto motion relative to the ground would be 146 m When cameras are handheld (asopposed to mounted in a bracket) blur can be reduced by physical compensationfor the motion (tracking) by the photographer many photographs show a level ofdetail indicating that blur at this level did not occur (eg gure 3(d )) Thus motionrelative to the ground is not an absolute barrier to ground resolution for handheldphotographs
43 Spacecraft windowsMost of the photographs in the NASA Astronaut Photography Database were
taken through window ports on the spacecraft used The transmittance of the windowport may be aVected by material inhomogeniety of the glass the number of layersof panes coatings used the quality of the surface polish environmentally inducedchanges in window materials (pressure loads thermal gradients) or deposited con-tamination Such degradation of the window cannot be corrected is diVerent foreach window and changes over time See Eppler et al (1996) and Scott (2000) fordiscussion of the spectral transmittance of the fused quartz window that is part ofthe US Laboratory Module (Destiny) of the International Space Station
44 Digitized images from lmWhen lm is scanned digitally the amount of information retained depends on
the spectral information extracted from the lm at the spatial limits of the scanner
J A Robinson et al4418
To date standard digitizing methodologies for astronaut photographs have not beenestablished and lm is digitized on a case-by-case basis using the equipment available
441 Digitizing and spatial resolutionLight (1993 1996) provided equations for determining the digitizing spatial
resolution needed to preserve spatial resolution of aerial photography based on thestatic resolving power (AWAR) for the system For lms where the manufacturer hasprovided data (Eastman Kodak 1998 K Teitelbaum personal communication) the resolving power of lms used for astronaut photography ranges from 32 lp mm Otilde 1to 100 lp mm Otilde 1 at low contrast (objectbackground ratio 161 table 3) The AWARfor the static case of the Hasselblad camera has been measured at high and lowcontrast (using lenses shown in table 1) with a maximum of approximately55 lp mm Otilde 1 (Fred Pearce unpublished data)
Based on the method of Light (1993 1996) the dimension of one spatial resolutionelement for a photograph with maximum static AWAR of 55 lp mm Otilde 1 would be18 mm lp Otilde 1 The acceptable range of spot size to preserve spatial information wouldthen be
18 mm
2 atilde 2aring scan spot size aring
18 mm
2(1)
and 6 mm aring scan spot size aring 9 mm Similarly for a more typical static AWAR of30 lp mm Otilde 1 (33 mm lpOtilde 1 low contrast Fred Pearce unpublished data) 11 mm aring scanspot sizearing 17 mm These scan spot sizes correspond to digitizing resolutions rangingfrom 4233ndash2822 ppi (pixels inch Otilde 1 ) for AWAR of 55 lp mm Otilde 1 and 2309ndash1494 ppi forAWAR of 30 lp mm Otilde 1 Scan spot sizes calculated for astronaut photography arecomparable to those calculated for the National Aerial Photography Program(9ndash13 mm Light 1996)
Widely available scanners that can digitize colour transparency lm currentlyhave a maximum spatial resolution of approximately 2400 ppi (106 mm pixel Otilde 1) Forexample we routinely use an Agfa Arcus II desktop scanner (see also Baltsavias1996) with 2400 ppi digitizing spatial resolution (2400 ppi optical resolution in onedirection and 1200 ppi interpolated to 2400 ppi in the other direction) and 2400 ppiwas used to calculate IFOV equivalents (tables 2 and 4) For some combinations oflens camera and contrast 2400 ppi will capture nearly all of the spatial informationcontained in the lm However for lm with higher resolving power thanEktachrome-64 for better lenses and for higher contrast targets digitizing at 2400 ppiwill not capture all of the spatial information in the lm
Improvements in digitizing technology will only produce real increases in IFOVto the limit of the AWAR of the photography system The incremental increase inspatial information above 3000 ppi (85 mm pixel Otilde 1 ) is not likely to outweigh thecosts of storing large images (Luman et al 1997) At 2400 ppi a Hasselblad frameis approximately 5200times5200 or 27 million pixels (table 2) while the same imagedigitized at 4000 ppi would contain 75 million pixels
Initial studies using astronaut photographs digitized at 2400 ppi (106 mm pixel Otilde 1 Webb et al 2000 Robinson et al 2000c) indicate that some GRD is lost comparedto photographic products Nevertheless the spatial resolution is still comparablewith other widely used data sources (Webb et al 2000) Robinson et al (2000c)found that digitizing at 21 mm pixel Otilde 1 provided information equivalent to 106 mmpixel Otilde 1 for identifying general urban area boundaries for six cities except for a single
Astronaut photography as digital data for remote sensing 4419
photo that required higher spatial resolution digitizing (it had been taken with ashorter lens and thus had less spatial resolution) Part of the equivalence observedin the urban areas study may be attributable to the fact that the atbed scannerused interpolates from 1200 ppi to 2400 ppi in one direction The appropriate digitiz-ing spatial resolution will in part depend on the scale of features of interest to theresearchers with a maximum upper limit set of a scan spot size of approximately6 mm (4233 ppi)
442 Digitizing and spectral resolutionWhen colour lm is digitized there will be loss of spectral resolution The three
lm emulsion layers (red green blue) have relatively distinct spectral responses butare fused together so that digitizers detect one colour signal and must convert itback into three (red green blue) digital channels Digital scanning is also subject tospectral calibration and reproduction errors Studies using digital astronaut photo-graphs to date have used 8 bits per channel However this is another parameterthat can be controlled when the lm is digitized We do not further address spectralresolution of digitally scanned images in this paper
45 External factors that in uence GRDAlthough not discussed in detail in this paper factors external to the spacecraft
and camera system (as listed by Moik (1980) for remote sensing in general) alsoimpact GRD These include atmospheric interference (due to scattering attenuationhaze) variable surface illumination (diVerences in terrain slope and orientation) andchange of terrain re ectance with viewing angle (bidirectional re ectance) For astro-naut photographs variable illumination is particularly important because orbits arenot sun-synchronous Photographs are illuminated by diVerent sun angles and imagesof a given location will have colour intensities that vary widely In addition theviewing angle has an eVect on the degree of object-to-backgroun d contrast andatmospheric interference
5 Estimating spatial resolution of astronaut photographsIn order to use astronaut photographs for digital remote sensing it is important
to be able to calculate the equivalent to an IFOVmdashthe ground area represented bya single pixel in a digitized orbital photograph The obliquity of most photographsmeans that pixel lsquosizesrsquo vary at diVerent places in an image Given a ground distanceD represented by a photograph in each direction (horizontal vertical ) an approxi-mate average pixel width (P the equivalent of IFOV) for the entire image can becalculated as follows
P=D
dS(2)
where D is the projected distance on the ground covered by the image in the samedirection as the pixel is measured d is the actual width of the image on the original lm (table 1) and S is the digitizing spatial resolution
Here we present three mathematical formulations for estimating the size of thefootprint or area on the ground covered by the image Example results from theapplication of all three formulations are given in table 5 The rst and simplestcalculation (formulation 1) gives an idea of the maximum spatial resolution attainableat a given altitude of orbit with a given lm format and a perfectly vertical (nadir)
J A Robinson et al4420
view downward Formulation 2 takes into account obliquity by calculating lookangle from the diVerence between the location on the ground represented at thecentre of the photograph and the nadir location of the spacecraft at the time thephotograph was taken ( gure 1) Formulation 3 describes an alternate solution tothe oblique look angle problem using coordinate-system transformations Thisformulation has been implemented in a documented spreadsheet and is availablefor download (httpeoljscnasagovsseopFootprintCalculatorxls) and is in theprocess of being implemented on a Web-based user interface to the AstronautPhotography Database (OYce of Earth Sciences 2000)
Although formulations 2 and 3 account for obliquity for the purposes of calcula-tion they treat the position of the spacecraft and position of the camera as one Inactuality astronauts are generally holding the cameras by hand (although camerasbracketed in the window are also used) and the selection of window position of theastronaut in the window and rotation of the camera relative to the movement ofthe spacecraft are not known Thus calculations using only the photo centre point(PC) and spacecraft nadir point (SN) give a locator ellipse and not the locations ofthe corners of the photograph A locator ellipse describes an estimated area on theground that is likely to be included in a speci c photograph regardless of the rotationof the lm plane about the camerarsquos optical axis ( gure 6)
Estimating the corner positions of the photo requires additional user input of asingle auxiliary pointmdasha location on the image that has a known location on theground Addition of this auxiliary point is an option available to users of thespreadsheet An example of the results of adding an auxiliary point is shown in gure 7 with comparisons of the various calculations in table 5
51 Formulation 1 Footprint for a nadir viewThe simplest way to estimate footprint size is to use the geometry of camera lens
and spacecraft altitude to calculate the scaling relationship between the image in the
Figure 6 Sketch representation of the locator ellipse an estimated area on the ground thatis likely to be included in a speci c photograph regardless of the rotation of the lmplane about the camerarsquos optical axis
Astronaut photography as digital data for remote sensing 4421
Figure 7 An example of the application of the Low Oblique Space Photo FootprintCalculator The photograph (STS093-704-50) of Limmen Bight Australia was takenfrom an altitude of 283 km using the Hasselblad camera with a 250 mm lens Mapused is 11 000 000 Operational Navigational Chart The distances shown equate toan average pixel size of 124 m for this photograph
lm and the area covered on the ground For a perfect nadir view the scalerelationship from the geometry of similar triangles is
d
D=
f
H(3)
where d=original image size D=distance of footprint on the ground f =focallength of lens and H=altitude Once D is known pixel size ( length=width) can becalculated from equation (2) These calculations represent the minimum footprintand minimum pixel size possible for a given camera system altitude and digitisingspatial resolution (table 2) Formulation 1 was used for calculating minimum pixelsizes shown in tables 2 and 4
By assuming digitizing at 2400 ppi (106 mm pixel Otilde 1 ) currently a spatial resolutioncommonly attainable from multipurpose colour transparency scanners (see sect441)this formulation was used to convert area covered to an IFOV equivalent for missionsof diVerent altitudes (table 2) Table 4 provides a comparison of IFOV of imagesfrom various satellites including the equivalent for astronaut photography Valuesin this table were derived by using formulation 1 to estimate the area covered becausea perfect nadir view represents the best possible spatial resolution and smallest eldof view that could be obtained
52 Formulation 2 Footprint for oblique views using simpli ed geometry and thegreat circle distance
A more realistic approach to determining the footprint of the photographaccounts for the fact that the camera is not usually pointing perfectly down at the
J A Robinson et al4422
Table 4 Comparison of pixel sizes (instantaneous elds of view) for satellite remote sensorswith pixel sizes for near vertical viewing photography from spacecraft Note that alldata returned from the automated satellites have the same eld of view while indicatedpixel sizes for astronaut photography are best-case for near vertical look angles See gure 5 for distributions of astronaut photographs of various look angles High-altitude aerial photography is also included for reference
Altitude PixelInstrument (km) width (mtimesm) Bands1
Landsat Multipectral Scanner2 (MSS 1974-) 880ndash940 79times56 4ndash5Landsat Thematic Mapper (TM 1982-) ~705 30times30 7Satellite pour lrsquoObservation de la Terre (SPOT 1986-) ~832
SPOT 1-3 Panchromatic 10times10 1SPOT 1-3 Multispectral (XS) 20times20 3
Astronaut Photography (1961-)Skylab3 (1973-1974 ) ~435
Multispectral Camera (6 camera stations) 17ndash19times17ndash19 3ndash8Earth Terrain Camera (hi-res colour) 875times875 3
Space Shuttle5 (1981-) 222ndash611Hasselblad 70 mm (250mm lens colour) vertical view 9ndash26times9ndash26 3Hasselblad 70 mm (100mm lens colour) vertical view 23ndash65times23ndash65 3Nikon 35 mm (400mm lens colour lm or ESC) vertical 5ndash16times5ndash16 3
High Altitude Aerial Photograph (1100 000) ~12 2times2 3
1Three bands listed for aerial photography obtainable by high-resolution colour digitizingFor high-resolution colour lm at 65 lp mmOtilde 1 (K Teitelbaum Eastman Kodak Co personalcommunication) the maximum information contained would be 3302 ppi
2Data from Simonett (1983Table 1-2)3Calculated as recommended by Jensen (19831596) the dividend of estimated ground
resolution at low contrast given in NASA (1974179-180) and 244Six lters plus colour and colour infrared lm (NASA 19747) 5Extracted from table 2
nadir point The look angle (the angle oV nadir that the camera is pointing) can becalculated trigonometrically by assuming a spherical earth and calculating the dis-tance between the coordinates of SN and PC ( gure 1) using the Great Circledistance haversine solution (Sinnott 1984 Snyder 198730ndash32 Chamberlain 1996)The diVerence between the spacecraft centre and nadir point latitudes Dlat=lat 2 shy lat 1 and the diVerence between the spacecraft centre and nadir pointlongitudes Dlon=lon 2 shy lon 1 enter the following equations
a=sin2ADlat
2 B+cos(lat 1) cos(lat 2) sin2ADlon
2 B (4)
c=2 arcsin(min[1 atilde a]) (5)
and oVset=R c with R the radius of a spherical Earth=6370 kmThe look angle (t ) is then given by
t=arctanAoffset
H B (6)
Assuming that the camera was positioned so that the imaginary line between thecentre and nadir points (the principal line) runs vertically through the centre of thephotograph the distance between the geometric centre of the photograph (principalpoint PP) and the top of the photograph is d2 ( gure 8(a)) The scale at any point
Astronaut photography as digital data for remote sensing 4423
(a) (b)
Figure 8 The geometric relationship between look angle lens focal length and the centrepoint and nadir points of a photograph (see also Estes and Simonett 1975) Note thatthe Earthrsquos surface and footprint represented in gure 1 are not shown here only arepresentation of the image area of the photograph Variables shown in the gurelook angle t focal length f PP centre of the image on the lm SN spacecraft nadird original image size (a) The special case where the camera is aligned squarely withthe isometric parallel In this case tilt occurs along a plane parallel with the centre ofthe photograph When the conditions are met calculations using Formulation 3 willgive the ground coordinates of the corner points of the photograph (b) The generalrelationship between these parameters and key photogrammetric points on the photo-graph For this more typical case coordinates of an additional point in the photographmust be input in order to adjust for twist of the camera and to nd the corner pointcoordinates
in the photograph varies as a function of the distance y along the principal linebetween the isocentre and the point according to the relationship
d
D=
f shy ysint
H(7)
(Wong 1980 equation (214) H amp the ground elevation h) Using gure 8(a) at thetop of the photo
y= f tanAt
2B+d
2(8)
and at the bottom of the photo
y= f tanAt
2Bshyd
2(9)
Thus for given PC and SN coordinates and assuming a photo orientation as in gure 8(a) and not gure 8(b) we can estimate a minimum D (using equations (7)and (8)) and a maximum D (using equations (7) and (9)) and then average the twoto determine the pixel size (P ) via equation (2)
J A Robinson et al4424
53 Formulation 3 T he L ow Oblique Space Photo Footprint CalculatorThe Low Oblique Space Photo Footprint Calculator was developed to provide
a more accurate estimation of the geographic coordinates for the footprint of a lowoblique photo of the Earthrsquos surface taken from a human-occupied spacecraft inorbit The calculator performs a series of three-dimensional coordinate transforma-tions to compute the location and orientation of the centre of the photo exposureplane relative to an earth referenced coordinate system The nominal camera focallength is then used to create a vector from the photorsquos perspective point througheach of eight points around the parameter of the image as de ned by the formatsize The geographical coordinates for the photo footprint are then computed byintersecting these photo vectors with a spherical earth model Although more sophist-icated projection algorithms are available no signi cant increase in the accuracy ofthe results would be produced by these algorithms due to inherent uncertainties inthe available input data (ie the spacecraft altitude photo centre location etc)
The calculations were initially implemented within a Microsoft Excel workbookwhich allowed one to embed the mathematical and graphical documentation nextto the actual calculations Thus interested users can inspect the mathematical pro-cessing A set of error traps was also built into the calculations to detect erroneousresults A summary of any errors generated is reported to the user with the resultsof the calculations Interested individuals are invited to download the Excel work-book from httpeoljscnasagovsseopFootprintCalculatorxls The calculations arecurrently being encoded in a high-level programming language and should soon beavailable alongside other background data provided for each photograph at OYceof Earth Sciences (2000)
For the purposes of these calculations a low oblique photograph was de ned as
one with the centre within 10 degrees of latitude and longitude of the spacecraftnadir point For the typical range of spacecraft altitudes to date this restricted thecalculations to photographs in which Earthrsquos horizon does not appear (the generalde nition of a low oblique photograph eg Campbell 199671 and gure 4)
531 Input data and resultsUpon opening the Excel workbook the user is presented with a program intro-
duction providing instructions for using the calculator The second worksheet tablsquoHow-To-Usersquo provides detailed step-by-step instructions for preparing the baselinedata for the program The third tab lsquoInput-Output rsquo contains the user input eldsand displays the results of the calculations The additional worksheets contain theactual calculations and program documentation Although users are welcome toreview these sheets an experienced user need only access the lsquoInput-Outputrsquo
spreadsheetTo begin a calculation the user enters the following information which is available
for each photo in the NASA Astronaut Photography Database (1) SN geographicalcoordinates of spacecraft nadir position at the time of photo (2) H spacecraftaltitude (3) PC the geographical coordinates of the centre of the photo (4) f nominal focal length and (5) d image format size The automatic implementationof the workbook on the web will automatically enter these values and completecalculations
For more accurate results the user may optionally enter the geographic co-ordinates and orientation for an auxiliary point on the photo which resolves thecamerarsquos rotation uncertainty about the optical axis The auxiliary point data must
Astronaut photography as digital data for remote sensing 4425
be computed by the user following the instruction contained in the lsquoHow-To-Usersquotab of the spreadsheet
After entering the input data the geographic coordinates of the photo footprint(ie four photo corner points four points at the bisector of each edge and the centreof the photo) are immediately displayed below the input elds along with any errormessages generated by the user input or by the calculations ( gure 7) Although
results are computed and displayed they should not be used when error messagesare produced by the program The program also computes the tilt angle for each ofthe photo vectors relative to the spacecraft nadir vector To the right of the photofootprint coordinates is displayed the arc distance along the surface of the sphere
between adjacent computed points
532 Calculation assumptions
The mathematical calculations implemented in the Low Oblique Space PhotoFootprint Calculator use the following assumptions
1 The SN location is used as exact even though the true value may vary by upto plusmn01 degree from the location provided with the photo
2 The spacecraft altitude is used as exact Although the determination of the
nadir point at the instant of a known spacecraft vector is relatively precise(plusmn115times10 Otilde 4 degrees) the propagator interpolates between sets of approxi-mately 10ndash40 known vectors per day and the time code recorded on the lmcan drift Thus the true value for SN may vary by up to plusmn01 degree fromthe value provided with the photo
3 The perspective centre of the camera is assumed to be at the given altitudeover the speci ed spacecraft nadir location at the time of photo exposure
4 The PC location is used as exact even though the true value may vary by upto plusmn05deg latitude and plusmn05deg longitude from the location provided with the
photo5 A spherical earth model is used with a nominal radius of 6 372 16154 m (a
common rst-order approximation for a spherical earth used in geodeticcomputations)
6 The nominal lens focal length of the camera lens is used in the computations(calibrated focal length values are not available)
7 The photo projection is based on the classic pin-hole camera model8 No correction for lens distortion or atmospheric re ection is made
9 If no auxiliary point data is provided the lsquoTop of the Imagersquo is orientedperpendicular to the vector from SN towards PC
533 T ransformation from Earth to photo coordinate systemsThe calculations begin by converting the geographic coordinates ( latitude and
longitude) of the SN and PC to a Rectangular Earth-Centred Coordinate System(R-Earth) de ned as shown in gure 9 (with the centre of the Earth at (0 0 0))
Using the vector from the Earthrsquos centre through SN and the spacecraft altitude thespacecraft location (SC) is also computed in R-Earth
For ease of computation a Rectangular Spacecraft-Centred Coordinate System(R-Spacecraft ) is de ned as shown in gure 9 The origin of R-Spacecraft is located
at SC with its +Z-axis aligned with vector from the centre of the Earth throughSN and its +X-axis aligned with the vector from SN to PC ( gure 9) The speci c
J A Robinson et al4426
Figure 9 Representation of the area included in a photograph as a geometric projectiononto the Earthrsquos surface Note that the focal length (the distance from the SpaceShuttle to the photo) is grossly overscale compared to the altitude (the distance fromthe Shuttle to the surface of the Earth
rotations and translation used to convert from the R-Earth to the R-Spacecraft arecomputed and documented in the spreadsheet
With the mathematical positions of the SC SN PC and the centre of the Earthcomputed in R-Spacecraft the program next computes the location of the camerarsquosprincipal point (PP) The principle point is the point of intersection of the opticalaxis of the lens with the image plane (ie the lm) It is nominally positioned at a
Astronaut photography as digital data for remote sensing 4427
distance equal to the focal length from the perspective centre of the camera (whichis assumed to be at SC) along the vector from PC through SC as shown in gure 9
A third coordinate system the Rectangular Photo Coordinate System (R-Photo)is created with its origin at PP its XndashY axial plane normal to the vector from PCthrough SC and its +X axis aligned with the +X axis of R-Spacecraft as shownin gure 9 The XndashY plane of this coordinate system represents the image plane ofthe photograph
534 Auxiliary point calculationsThe calculations above employ a critical assumption that all photos are taken
with the lsquotop of the imagersquo oriented perpendicular to the vector from the SN towardsPC as shown in gure 9 To avoid non-uniform solar heating of the external surfacemost orbiting spacecraft are slowly and continually rotated about one or more axesIn this condition a ight crew member taking a photo while oating in microgravitycould orient the photo with practically any orientation relative to the horizon (seealso gure 8(b)) Unfortunately since these photos are taken with conventionalhandheld cameras there is no other information available which can be used toresolve the photorsquos rotational ambiguity about the optical axis other then the photoitself This is why the above assumption is used and the footprint computed by thiscalculator is actually a lsquolocator ellipsersquo which estimates the area on the ground whatis likely to be included in a speci c photograph (see gure 6) This locator ellipse ismost accurate for square image formats and subject to additional distortion as thephotograph format becomes more rectangular
If the user wants a more precise calculation of the image footprint the photorsquosrotational ambiguity about the optical axis must be resolved This can be done inthe calculator by adding data for an auxiliary point Detailed instructions regardinghow to prepare and use auxiliary point data in the computations are included in thelsquoHow-To-Usersquo tab of the spreadsheet Basically the user determines which side ofthe photograph is top and then measures the angle between the line from PP to thetop of the photo and from PP to the auxiliary point on the photo ( gure 9)
If the user includes data for an auxiliary point a series of computations arecompleted to resolve the photo rotation ambiguity about the optical axis (ie the
+Z axis in R-Photo) A vector from the Auxiliary Point on the Earth (AE) throughthe photograph perspective centre ( located at SC) is intersected with the photo imageplane (XndashY plane of R-Photo) to compute the coordinates of the Auxiliary Point onthe photo (AP) in R-Photo A two-dimensional angle in the XndashY plane of R-Photofrom the shy X axis to a line from PP to AP is calculated as shown in gure 9 The
shy X axis is used as the origin of the angle since it represents the top of the photoonce it passes through the perspective centre The diVerence between the computedangle and the angle measured by the user on the photo resolves the ambiguity inthe rotation of the photo relative to the principal line ( gures 7 and 9) Thetransformations from R-Spacecraft and R-Photo are then modi ed to include anadditional rotation angle about the +Z axis in R-Photo
535 lsquoFootprint rsquo calculationsThe program next computes the coordinates of eight points about the perimeter
of the image format (ie located at the four photo corners plus a bisector pointalong each edge of the image) These points are identi ed in R-Photo based uponthe photograph format size and then converted to R-Spacecraft Since all computa-tions are done in orthogonal coordinate systems the R-Spacecraft to R-Photo
J A Robinson et al4428
rotation matrix is transposed to produce an R-Photo to R-Spacecraft rotation matrixOnce in R-Spacecraft a unit vector from each of the eight perimeter points throughthe photo perspective centre (the same point as SC) is computed This provides thecoordinates for points about the perimeter of the image format with their directionvectors in a common coordinate system with other key points needed to computethe photo footprint
The next step is to compute the point of intersection between the spherical earthmodel and each of the eight perimeter point vectors The scalar value for eachperimeter point unit vector is computed using two-dimensional planar trigonometryAn angle c is computed using the formula for the cosine of an angle between twoincluded vectors (the perimeter point unit vector and the vector from SC to thecentre of the Earth) Angle y is computed using the Law of Sines Angle e=180degrees shy c shy y The scalar value of the perimeter point vector is computed using eand the Law of Cosines The scalar value is then multiplied by the perimeter pointunit vector to produce the three-dimensional point of intersection of the vector withEarthrsquos surface in R-Spacecraft The process is repeated independently for each ofthe eight perimeter point vectors Aside from its mathematical simplicity the valueof arcsine (computed in step 2) will exceed the normal range for the sin of an anglewhen the perimeter point vectors fail to intersect with the surface of the earth Asimple test based on this principle allows the program to correctly handle obliquephotos which image a portion of the horizon (see results for high oblique photographsin table 5)
The nal step in the process converts the eight earth intersection points from theR-Spacecraft to R-Earth The results are then converted to the standard geographiccoordinate system and displayed on the lsquoInput-Outputrsquo page of the spreadsheet
54 ExamplesAll three formulations were applied to the photographs included in this paper
and the results are compared in table 5 Formulation 1 gives too small a value fordistance across the photograph (D) for all but the most nadir shots and thus servesas an indicator of the best theoretical case but is not a good measure for a speci cphotograph For example the photograph of Lake Eyre taken from 276 km altitude( gure 2(a)) and the photograph of Limmen Bight ( gure 7) were closest to beingnadir views (oVsetslt68 km or tlt15deg table 5) For these photographs D calculatedusing Formulation 1 was similar to the minimum D calculated using Formulations2 and 3 For almost all other more oblique photographs Formulation 1 gave asigni cant underestimate of the distance covered in the photograph For gure 5(A)(the picture of Houston taken with a 40 mm lens) Formulation 1 did not give anunderestimate for D This is because Formulation 1 does not account for curvatureof the Earth in any way With this large eld of view assuming a at Earth in atedthe value of D above the minimum from calculations that included Earth curvature
A major diVerence between Formulations 2 and 3 is the ability to estimate pixelsizes (P) in both directions (along the principal line and perpendicular to the principalline) For the more oblique photographs the vertical estimate of D and pixel sizes ismuch larger than in the horizontal direction (eg the low oblique and high obliquephotographs of Hawaii gure 4 table 5)
For the area of Limmen Bight ( gure 7) table 5 illustrates the improvement inthe estimate of distance and pixel size that can be obtained by re-estimating thelocation of the PC with greater accuracy Centre points in the catalogued data are
Astronaut photography as digital data for remote sensing 4429
plusmn05deg of latitude and longitude When the centre point was re-estimated to plusmn002degwe determined that the photograph was not taken as obliquely as rst thought(change in the estimate of the look angle t from 169 to 128deg table 5) When theauxiliary point was added to the calculations of Formula 3 the calculated look angleshrank further to 110deg indicating that this photograph was taken at very close toa nadir view Of course this improvement in accuracy could have also led to estimatesof greater obliquity and corresponding larger pixel sizes
We also tested the performance of the scale calculator with auxiliary point byestimating the corner and point locations on the photograph using a 11 000 000Operational Navigational Chart For this test we estimated our ability to readcoordinates from the map as plusmn002degand our error in nding the locations of thecorner points as plusmn015deg (this error varies among photographs depending on thedetail that can be matched between photo and map) For Limmen Bight ( gure 7)the mean diVerence between map estimates and calculator estimates for four pointswas 031deg (SD=018 n=8) For a photograph of San Francisco Bay (STS062-151-291) the mean diVerence between map estimates and calculator estimates forfour points was 0064deg (SD=018 n=8) For a photograph of San Francisco Bay(STS062-151-291) the mean diVerence between map estimates and calculator estim-ates for eight points was 0196deg (SD=0146 n=16) Thus in one case the calculatorestimates were better than our estimate of the error in locating corner points on themap It is reasonable to expect that for nadir-viewing photographs the calculatorused with an auxiliary point can estimate locations of the edges of a photograph towithin plusmn03deg
55 Empirical con rmation of spatial resolution estimatesAs stated previously a challenge to estimating system-AWAR for astronaut
photography of Earth is the lack of suitable targets Small features in an image cansometimes be used as a check on the size of objects that can be successfully resolvedgiving an approximate value for GRD Similarly the number of pixels that make upthose features in the digitized image can be used to make an independent calculationof pixel size We have successfully used features such as roads and airport runwaysto make estimates of spatial scale and resolution (eg Robinson et al 2000c) Whilerecognizing that the use of linear detail in an image is a poor approximation to abar target and that linear objects smaller than the resolving power can often bedetected (Charman 1965) few objects other than roads could be found to make anydirect estimates of GRD Thus roads and runways were used in the images ofHouston (where one can readily conduct ground veri cations and where a numberof higher-contrast concrete roadways were available) to obtain empirical estimatesof GRD and pixel size for comparison with table 5
In the all-digital ESC image of Houston ( gure 3(d )) we examined EllingtonField runway 4-22 (centre left of the image) which is 24384 mtimes457 m This runwayis approximately 6ndash7 pixels in width and 304ndash3094 pixels in length so pixelsrepresent an distance on the ground 7ndash8 m Using a lower contrast measure of astreet length between two intersections (2125 m=211 pixels) a pixel width of 101 mis estimated These results compare favourably with the minimum estimate of 81 mpixels using Formulation 1 (table 5) For an estimate of GRD that would be morecomparable to aerial photography of a line target the smallest street without treecover that could be clearly distinguished on the photograph was 792 m wide Thesmallest non-street object (a gap between stages of the Saturn rocket on display in
J A Robinson et al4430
Tab
le5
App
lica
tio
nof
met
hod
sfo
res
tim
ati
ng
spati
alre
solu
tio
no
fast
ron
aut
ph
oto
gra
ph
sin
clu
din
gF
orm
ula
tion
s12
and
3Im
pro
ved
cen
tre
po
int
esti
mat
esan
dauxilia
ryp
oin
tca
lcu
lati
ons
are
sho
wn
for
gure
7V
aria
ble
sas
use
din
the
text
fle
ns
foca
lle
ngt
hH
al
titu
de
PC
ph
oto
gra
ph
cen
tre
poin
tSN
sp
ace
craft
nadir
po
int
D
dis
tance
cover
edo
nth
egr
oun
d
Pd
ista
nce
on
the
grou
nd
repre
sen
ted
by
apix
el
OVse
td
ista
nce
bet
wee
nP
Cand
SNusi
ng
the
grea
tci
rcle
form
ula
t
look
angle
away
from
SN
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
2L
ake
Eyre
ST
S093
-717
-80
250
276
shy29
0137
0shy
28
413
71
607
11
767
413
7609
ndash64
212
013
760
9ndash64
7121
124
ST
S082
-754
-61
250
617
shy28
5137
0shy
28
113
50
135
726
1201
18
013
8ndash14
827
517
9138ndash15
3277
294
Fig
ure
3H
ou
sto
nS
TS
062
-97-
151
4029
829
5shy
95
530
5shy
95
4410
78
8112
20
5348ndash58
990
1205
351
ndash63
8951
10
36
ST
S094
-737
-28
100
294
295
shy
95
528
3shy
95
5162
31
1133
24
415
8ndash20
334
724
3158ndash20
7351
390
ST
S055
-71-
41250
302
295
shy
95
028
2shy
93
766
412
8192
32
5736
ndash84
715
232
273
9ndash85
7154
185
S73
E51
45300
2685
295
shy
95
024
78
0F
igu
re4
Haw
aii
ST
S069
-701
-65
250
370
195
shy
155
520
4shy
153
881
415
7204
28
9876
ndash98
917
928
688
0ndash10
9181
210
ST
S069
-701
-70
250
370
210
shy
157
519
2shy
150
781
415
7738
63
414
9ndash23
336
760
7148ndash55
6394
10
63
ST
S069
-729
-27
100
398
200
shy
156
620
5shy
148
2219
42
1867
65
432
8ndash13
10
158
62
4323+
311
N
A
Astronaut photography as digital data for remote sensing 4431
Tab
le5
(Cont
inued
)
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
7L
imm
enB
igh
tS
TS
093
-704
-50
250
283
shy15
0135
0shy
15
013
58
623
120
859
16
963
0ndash67
3125
16
863
0ndash67
612
613
2P
CR
e-es
tim
ated
shy14
733
1352
67
64
5128
623
ndash65
5123
128
623
ndash65
712
3127
Auxilia
ryP
oin
t611
062
3ndash65
112
312
5F
igu
re10
N
ear
Au
stin
T
XS
TS
095
-716
-46
250
543
30
5shy
975
28
6shy
1007
120
230
375
34
613
5ndash157
281
34
1136ndash16
128
636
0
1 All
pho
togr
aphs
exce
pt
on
eta
ken
wit
hH
ass
elbla
dca
mer
aso
dis
tan
ceacr
oss
the
image
d=
55m
m
for
S73
E51
45
taken
wit
hel
ectr
onic
still
cam
erad=
276
5m
mho
rizo
nta
l2 P
Cand
SN
are
giv
enin
the
form
(lati
tud
elo
ngit
ude)
deg
rees
w
ith
neg
ativ
en
um
ber
sre
pre
senti
ng
south
and
wes
tA
ccura
cyo
fP
Cis
plusmn0
5and
SN
isplusmn
01
as
reco
rded
inth
eA
stro
naut
Ph
oto
gra
phy
Data
base
ex
cep
tw
her
en
ote
dth
atce
ntr
ep
oin
tw
asre
-est
imate
dw
ith
grea
ter
accu
racy
3 C
alc
ula
ted
usi
ng
the
mea
nofth
em
inim
um
an
dm
axim
um
esti
mate
sof
DF
orm
ula
tion
2co
nsi
der
sD
inth
ed
irec
tion
per
pen
dic
ula
rto
the
Pri
nci
pal
Lin
eo
nly
4 C
alc
ula
ted
inho
rizo
nta
lan
dve
rtic
aldir
ecti
on
sb
ut
still
ass
um
ing
geom
etry
of
gure
6(B
)F
irst
nu
mb
eris
the
mea
no
fth
em
inim
um
Dat
the
bott
om
of
the
ph
oto
and
the
maxi
mum
Dat
the
top
of
the
ph
oto
(range
show
nin
the
tab
le)
and
isco
mpara
ble
toF
orm
ula
tio
n2
Sec
ond
num
ber
isth
em
ean
of
Des
tim
ated
alon
gth
eP
rin
cip
alline
and
alo
ng
the
ou
tsid
eed
ge
(range
not
show
n)
5 Alt
itu
de
isap
pro
xim
ate
for
mis
sion
as
exac
tti
me
pho
togr
ap
hw
asta
ken
and
corr
esp
on
din
gn
adir
data
are
no
tava
ilab
le
Lack
of
nad
irdata
pre
ven
tsap
plica
tion
of
Form
ula
tion
s2
an
d3
6 Au
xilliar
ypo
int
add
edw
as
shy14
80
lati
tud
e13
51
2lo
ngit
ude
shy46
5degro
tati
on
angle
in
stru
ctio
ns
for
esti
mati
ng
the
rota
tio
nan
gle
para
met
erare
conta
ined
as
par
tof
the
scal
eca
lcu
lato
rsp
read
shee
tdo
cum
enta
tio
n
J A Robinson et al4432
a park at Johnson Space Center) that could clearly be distinguished on the photo-graph was 853 m wide
For the photograph of Houston taken with a 250 mm lens ( gure 3(C)) anddigitized from second generation lm at 2400 ppi (106 mm pixel Otilde 1 ) Ellington Fieldrunway 4-22 is 3 pixels in width and 1613 pixels in length so pixels represent151ndash152 m on the ground These results compare favourably with the minimumestimate of 152 m using Formulation 2 and 154ndash185 m pixels using Formulation 3(table 5) For an estimate of GRD using an 8times8 inch print (1369 enlargement) and4times magni cation the smallest street that could clearly be distinguished was 822 mwide the same feature could barely be distinguished on the digitized image
We also made an empirical estimate of spatial resolution for lower contrastvegetation boundaries By clearing forest so that a pattern would be visible to landingaircraft a landowner outside Austin Texas (see also aerial photo in Lisheron 2000)created a target that is also useful for evaluating spatial resolution of astronautphotographs The forest was selectively cleared in order to spell the landownerrsquosname lsquoLUECKErsquo with the remaining trees ( gure 10) According to local surveyorswho planned the clearing the plan was to create letters that were 3100 fttimes1700 ft(9449 mtimes5182 m) Photographed at a high altitude relative to most Shuttle missions(543 km) with a 250 mm lens Formula 3 predicts that each pixel would represent anarea 286 mtimes360 m on the ground (table 5) When original lm was digitized at2400 ppi (106 mm pixel Otilde 1 ) letters correspond to 294times188 pixels for a comparablepixel size of 27ndash32 m
6 Summary and conclusionsAstronaut photographs can be an excellent source of data for remote sensing
applications Best-case resolutions are similar to that for Landsat or SPOT withpixels as small as lt10 m It was estimated that of images taken to date 504 hadlens and obliquity characteristics that make them potential candidates for remotesensing information Digitized astronaut photographs can be overlaid with othersatellite data using GIS (Eckardt et al 2000) or used to ll in gaps in time serieswhen other imagery is not available As a source of public-domain information theycan be very useful for scientists who do not have access to satellite imagery eitherbecause of the expense of image acquisition (to the end user) or the computersystems needed for processing satellite images These diVerences in image acquisitioncosts (to the user) are summarized in table 6
Searching of the complete database of NASA astronaut photography includinglow-spatial-resolution browse images is available via the Web (OYce of EarthSciences 2000) This provides nearly global access for identifying images that willcontribute to a speci c scienti c project For the most detailed studies digitalproducts posted to the web will not be of suYcient quality To date we have madeit a practice of digitizing small numbers of images when requested by scientists atno charge Such requests (including a description of the project involved) can bemade through the authors or using the contact information listed on the websiteInvestigators with needs for larger numbers of digital images have also been servedthrough collaborative agreements
In our experience scientists that nd the data most valuable are those in develop-ing countries those studying areas that have not usually been targets for the majorsatellites those having diYculty in nding low-cloud images those interested inconstructing time series or those interested in using a large number of images
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
J A Robinson et al4418
To date standard digitizing methodologies for astronaut photographs have not beenestablished and lm is digitized on a case-by-case basis using the equipment available
441 Digitizing and spatial resolutionLight (1993 1996) provided equations for determining the digitizing spatial
resolution needed to preserve spatial resolution of aerial photography based on thestatic resolving power (AWAR) for the system For lms where the manufacturer hasprovided data (Eastman Kodak 1998 K Teitelbaum personal communication) the resolving power of lms used for astronaut photography ranges from 32 lp mm Otilde 1to 100 lp mm Otilde 1 at low contrast (objectbackground ratio 161 table 3) The AWARfor the static case of the Hasselblad camera has been measured at high and lowcontrast (using lenses shown in table 1) with a maximum of approximately55 lp mm Otilde 1 (Fred Pearce unpublished data)
Based on the method of Light (1993 1996) the dimension of one spatial resolutionelement for a photograph with maximum static AWAR of 55 lp mm Otilde 1 would be18 mm lp Otilde 1 The acceptable range of spot size to preserve spatial information wouldthen be
18 mm
2 atilde 2aring scan spot size aring
18 mm
2(1)
and 6 mm aring scan spot size aring 9 mm Similarly for a more typical static AWAR of30 lp mm Otilde 1 (33 mm lpOtilde 1 low contrast Fred Pearce unpublished data) 11 mm aring scanspot sizearing 17 mm These scan spot sizes correspond to digitizing resolutions rangingfrom 4233ndash2822 ppi (pixels inch Otilde 1 ) for AWAR of 55 lp mm Otilde 1 and 2309ndash1494 ppi forAWAR of 30 lp mm Otilde 1 Scan spot sizes calculated for astronaut photography arecomparable to those calculated for the National Aerial Photography Program(9ndash13 mm Light 1996)
Widely available scanners that can digitize colour transparency lm currentlyhave a maximum spatial resolution of approximately 2400 ppi (106 mm pixel Otilde 1) Forexample we routinely use an Agfa Arcus II desktop scanner (see also Baltsavias1996) with 2400 ppi digitizing spatial resolution (2400 ppi optical resolution in onedirection and 1200 ppi interpolated to 2400 ppi in the other direction) and 2400 ppiwas used to calculate IFOV equivalents (tables 2 and 4) For some combinations oflens camera and contrast 2400 ppi will capture nearly all of the spatial informationcontained in the lm However for lm with higher resolving power thanEktachrome-64 for better lenses and for higher contrast targets digitizing at 2400 ppiwill not capture all of the spatial information in the lm
Improvements in digitizing technology will only produce real increases in IFOVto the limit of the AWAR of the photography system The incremental increase inspatial information above 3000 ppi (85 mm pixel Otilde 1 ) is not likely to outweigh thecosts of storing large images (Luman et al 1997) At 2400 ppi a Hasselblad frameis approximately 5200times5200 or 27 million pixels (table 2) while the same imagedigitized at 4000 ppi would contain 75 million pixels
Initial studies using astronaut photographs digitized at 2400 ppi (106 mm pixel Otilde 1 Webb et al 2000 Robinson et al 2000c) indicate that some GRD is lost comparedto photographic products Nevertheless the spatial resolution is still comparablewith other widely used data sources (Webb et al 2000) Robinson et al (2000c)found that digitizing at 21 mm pixel Otilde 1 provided information equivalent to 106 mmpixel Otilde 1 for identifying general urban area boundaries for six cities except for a single
Astronaut photography as digital data for remote sensing 4419
photo that required higher spatial resolution digitizing (it had been taken with ashorter lens and thus had less spatial resolution) Part of the equivalence observedin the urban areas study may be attributable to the fact that the atbed scannerused interpolates from 1200 ppi to 2400 ppi in one direction The appropriate digitiz-ing spatial resolution will in part depend on the scale of features of interest to theresearchers with a maximum upper limit set of a scan spot size of approximately6 mm (4233 ppi)
442 Digitizing and spectral resolutionWhen colour lm is digitized there will be loss of spectral resolution The three
lm emulsion layers (red green blue) have relatively distinct spectral responses butare fused together so that digitizers detect one colour signal and must convert itback into three (red green blue) digital channels Digital scanning is also subject tospectral calibration and reproduction errors Studies using digital astronaut photo-graphs to date have used 8 bits per channel However this is another parameterthat can be controlled when the lm is digitized We do not further address spectralresolution of digitally scanned images in this paper
45 External factors that in uence GRDAlthough not discussed in detail in this paper factors external to the spacecraft
and camera system (as listed by Moik (1980) for remote sensing in general) alsoimpact GRD These include atmospheric interference (due to scattering attenuationhaze) variable surface illumination (diVerences in terrain slope and orientation) andchange of terrain re ectance with viewing angle (bidirectional re ectance) For astro-naut photographs variable illumination is particularly important because orbits arenot sun-synchronous Photographs are illuminated by diVerent sun angles and imagesof a given location will have colour intensities that vary widely In addition theviewing angle has an eVect on the degree of object-to-backgroun d contrast andatmospheric interference
5 Estimating spatial resolution of astronaut photographsIn order to use astronaut photographs for digital remote sensing it is important
to be able to calculate the equivalent to an IFOVmdashthe ground area represented bya single pixel in a digitized orbital photograph The obliquity of most photographsmeans that pixel lsquosizesrsquo vary at diVerent places in an image Given a ground distanceD represented by a photograph in each direction (horizontal vertical ) an approxi-mate average pixel width (P the equivalent of IFOV) for the entire image can becalculated as follows
P=D
dS(2)
where D is the projected distance on the ground covered by the image in the samedirection as the pixel is measured d is the actual width of the image on the original lm (table 1) and S is the digitizing spatial resolution
Here we present three mathematical formulations for estimating the size of thefootprint or area on the ground covered by the image Example results from theapplication of all three formulations are given in table 5 The rst and simplestcalculation (formulation 1) gives an idea of the maximum spatial resolution attainableat a given altitude of orbit with a given lm format and a perfectly vertical (nadir)
J A Robinson et al4420
view downward Formulation 2 takes into account obliquity by calculating lookangle from the diVerence between the location on the ground represented at thecentre of the photograph and the nadir location of the spacecraft at the time thephotograph was taken ( gure 1) Formulation 3 describes an alternate solution tothe oblique look angle problem using coordinate-system transformations Thisformulation has been implemented in a documented spreadsheet and is availablefor download (httpeoljscnasagovsseopFootprintCalculatorxls) and is in theprocess of being implemented on a Web-based user interface to the AstronautPhotography Database (OYce of Earth Sciences 2000)
Although formulations 2 and 3 account for obliquity for the purposes of calcula-tion they treat the position of the spacecraft and position of the camera as one Inactuality astronauts are generally holding the cameras by hand (although camerasbracketed in the window are also used) and the selection of window position of theastronaut in the window and rotation of the camera relative to the movement ofthe spacecraft are not known Thus calculations using only the photo centre point(PC) and spacecraft nadir point (SN) give a locator ellipse and not the locations ofthe corners of the photograph A locator ellipse describes an estimated area on theground that is likely to be included in a speci c photograph regardless of the rotationof the lm plane about the camerarsquos optical axis ( gure 6)
Estimating the corner positions of the photo requires additional user input of asingle auxiliary pointmdasha location on the image that has a known location on theground Addition of this auxiliary point is an option available to users of thespreadsheet An example of the results of adding an auxiliary point is shown in gure 7 with comparisons of the various calculations in table 5
51 Formulation 1 Footprint for a nadir viewThe simplest way to estimate footprint size is to use the geometry of camera lens
and spacecraft altitude to calculate the scaling relationship between the image in the
Figure 6 Sketch representation of the locator ellipse an estimated area on the ground thatis likely to be included in a speci c photograph regardless of the rotation of the lmplane about the camerarsquos optical axis
Astronaut photography as digital data for remote sensing 4421
Figure 7 An example of the application of the Low Oblique Space Photo FootprintCalculator The photograph (STS093-704-50) of Limmen Bight Australia was takenfrom an altitude of 283 km using the Hasselblad camera with a 250 mm lens Mapused is 11 000 000 Operational Navigational Chart The distances shown equate toan average pixel size of 124 m for this photograph
lm and the area covered on the ground For a perfect nadir view the scalerelationship from the geometry of similar triangles is
d
D=
f
H(3)
where d=original image size D=distance of footprint on the ground f =focallength of lens and H=altitude Once D is known pixel size ( length=width) can becalculated from equation (2) These calculations represent the minimum footprintand minimum pixel size possible for a given camera system altitude and digitisingspatial resolution (table 2) Formulation 1 was used for calculating minimum pixelsizes shown in tables 2 and 4
By assuming digitizing at 2400 ppi (106 mm pixel Otilde 1 ) currently a spatial resolutioncommonly attainable from multipurpose colour transparency scanners (see sect441)this formulation was used to convert area covered to an IFOV equivalent for missionsof diVerent altitudes (table 2) Table 4 provides a comparison of IFOV of imagesfrom various satellites including the equivalent for astronaut photography Valuesin this table were derived by using formulation 1 to estimate the area covered becausea perfect nadir view represents the best possible spatial resolution and smallest eldof view that could be obtained
52 Formulation 2 Footprint for oblique views using simpli ed geometry and thegreat circle distance
A more realistic approach to determining the footprint of the photographaccounts for the fact that the camera is not usually pointing perfectly down at the
J A Robinson et al4422
Table 4 Comparison of pixel sizes (instantaneous elds of view) for satellite remote sensorswith pixel sizes for near vertical viewing photography from spacecraft Note that alldata returned from the automated satellites have the same eld of view while indicatedpixel sizes for astronaut photography are best-case for near vertical look angles See gure 5 for distributions of astronaut photographs of various look angles High-altitude aerial photography is also included for reference
Altitude PixelInstrument (km) width (mtimesm) Bands1
Landsat Multipectral Scanner2 (MSS 1974-) 880ndash940 79times56 4ndash5Landsat Thematic Mapper (TM 1982-) ~705 30times30 7Satellite pour lrsquoObservation de la Terre (SPOT 1986-) ~832
SPOT 1-3 Panchromatic 10times10 1SPOT 1-3 Multispectral (XS) 20times20 3
Astronaut Photography (1961-)Skylab3 (1973-1974 ) ~435
Multispectral Camera (6 camera stations) 17ndash19times17ndash19 3ndash8Earth Terrain Camera (hi-res colour) 875times875 3
Space Shuttle5 (1981-) 222ndash611Hasselblad 70 mm (250mm lens colour) vertical view 9ndash26times9ndash26 3Hasselblad 70 mm (100mm lens colour) vertical view 23ndash65times23ndash65 3Nikon 35 mm (400mm lens colour lm or ESC) vertical 5ndash16times5ndash16 3
High Altitude Aerial Photograph (1100 000) ~12 2times2 3
1Three bands listed for aerial photography obtainable by high-resolution colour digitizingFor high-resolution colour lm at 65 lp mmOtilde 1 (K Teitelbaum Eastman Kodak Co personalcommunication) the maximum information contained would be 3302 ppi
2Data from Simonett (1983Table 1-2)3Calculated as recommended by Jensen (19831596) the dividend of estimated ground
resolution at low contrast given in NASA (1974179-180) and 244Six lters plus colour and colour infrared lm (NASA 19747) 5Extracted from table 2
nadir point The look angle (the angle oV nadir that the camera is pointing) can becalculated trigonometrically by assuming a spherical earth and calculating the dis-tance between the coordinates of SN and PC ( gure 1) using the Great Circledistance haversine solution (Sinnott 1984 Snyder 198730ndash32 Chamberlain 1996)The diVerence between the spacecraft centre and nadir point latitudes Dlat=lat 2 shy lat 1 and the diVerence between the spacecraft centre and nadir pointlongitudes Dlon=lon 2 shy lon 1 enter the following equations
a=sin2ADlat
2 B+cos(lat 1) cos(lat 2) sin2ADlon
2 B (4)
c=2 arcsin(min[1 atilde a]) (5)
and oVset=R c with R the radius of a spherical Earth=6370 kmThe look angle (t ) is then given by
t=arctanAoffset
H B (6)
Assuming that the camera was positioned so that the imaginary line between thecentre and nadir points (the principal line) runs vertically through the centre of thephotograph the distance between the geometric centre of the photograph (principalpoint PP) and the top of the photograph is d2 ( gure 8(a)) The scale at any point
Astronaut photography as digital data for remote sensing 4423
(a) (b)
Figure 8 The geometric relationship between look angle lens focal length and the centrepoint and nadir points of a photograph (see also Estes and Simonett 1975) Note thatthe Earthrsquos surface and footprint represented in gure 1 are not shown here only arepresentation of the image area of the photograph Variables shown in the gurelook angle t focal length f PP centre of the image on the lm SN spacecraft nadird original image size (a) The special case where the camera is aligned squarely withthe isometric parallel In this case tilt occurs along a plane parallel with the centre ofthe photograph When the conditions are met calculations using Formulation 3 willgive the ground coordinates of the corner points of the photograph (b) The generalrelationship between these parameters and key photogrammetric points on the photo-graph For this more typical case coordinates of an additional point in the photographmust be input in order to adjust for twist of the camera and to nd the corner pointcoordinates
in the photograph varies as a function of the distance y along the principal linebetween the isocentre and the point according to the relationship
d
D=
f shy ysint
H(7)
(Wong 1980 equation (214) H amp the ground elevation h) Using gure 8(a) at thetop of the photo
y= f tanAt
2B+d
2(8)
and at the bottom of the photo
y= f tanAt
2Bshyd
2(9)
Thus for given PC and SN coordinates and assuming a photo orientation as in gure 8(a) and not gure 8(b) we can estimate a minimum D (using equations (7)and (8)) and a maximum D (using equations (7) and (9)) and then average the twoto determine the pixel size (P ) via equation (2)
J A Robinson et al4424
53 Formulation 3 T he L ow Oblique Space Photo Footprint CalculatorThe Low Oblique Space Photo Footprint Calculator was developed to provide
a more accurate estimation of the geographic coordinates for the footprint of a lowoblique photo of the Earthrsquos surface taken from a human-occupied spacecraft inorbit The calculator performs a series of three-dimensional coordinate transforma-tions to compute the location and orientation of the centre of the photo exposureplane relative to an earth referenced coordinate system The nominal camera focallength is then used to create a vector from the photorsquos perspective point througheach of eight points around the parameter of the image as de ned by the formatsize The geographical coordinates for the photo footprint are then computed byintersecting these photo vectors with a spherical earth model Although more sophist-icated projection algorithms are available no signi cant increase in the accuracy ofthe results would be produced by these algorithms due to inherent uncertainties inthe available input data (ie the spacecraft altitude photo centre location etc)
The calculations were initially implemented within a Microsoft Excel workbookwhich allowed one to embed the mathematical and graphical documentation nextto the actual calculations Thus interested users can inspect the mathematical pro-cessing A set of error traps was also built into the calculations to detect erroneousresults A summary of any errors generated is reported to the user with the resultsof the calculations Interested individuals are invited to download the Excel work-book from httpeoljscnasagovsseopFootprintCalculatorxls The calculations arecurrently being encoded in a high-level programming language and should soon beavailable alongside other background data provided for each photograph at OYceof Earth Sciences (2000)
For the purposes of these calculations a low oblique photograph was de ned as
one with the centre within 10 degrees of latitude and longitude of the spacecraftnadir point For the typical range of spacecraft altitudes to date this restricted thecalculations to photographs in which Earthrsquos horizon does not appear (the generalde nition of a low oblique photograph eg Campbell 199671 and gure 4)
531 Input data and resultsUpon opening the Excel workbook the user is presented with a program intro-
duction providing instructions for using the calculator The second worksheet tablsquoHow-To-Usersquo provides detailed step-by-step instructions for preparing the baselinedata for the program The third tab lsquoInput-Output rsquo contains the user input eldsand displays the results of the calculations The additional worksheets contain theactual calculations and program documentation Although users are welcome toreview these sheets an experienced user need only access the lsquoInput-Outputrsquo
spreadsheetTo begin a calculation the user enters the following information which is available
for each photo in the NASA Astronaut Photography Database (1) SN geographicalcoordinates of spacecraft nadir position at the time of photo (2) H spacecraftaltitude (3) PC the geographical coordinates of the centre of the photo (4) f nominal focal length and (5) d image format size The automatic implementationof the workbook on the web will automatically enter these values and completecalculations
For more accurate results the user may optionally enter the geographic co-ordinates and orientation for an auxiliary point on the photo which resolves thecamerarsquos rotation uncertainty about the optical axis The auxiliary point data must
Astronaut photography as digital data for remote sensing 4425
be computed by the user following the instruction contained in the lsquoHow-To-Usersquotab of the spreadsheet
After entering the input data the geographic coordinates of the photo footprint(ie four photo corner points four points at the bisector of each edge and the centreof the photo) are immediately displayed below the input elds along with any errormessages generated by the user input or by the calculations ( gure 7) Although
results are computed and displayed they should not be used when error messagesare produced by the program The program also computes the tilt angle for each ofthe photo vectors relative to the spacecraft nadir vector To the right of the photofootprint coordinates is displayed the arc distance along the surface of the sphere
between adjacent computed points
532 Calculation assumptions
The mathematical calculations implemented in the Low Oblique Space PhotoFootprint Calculator use the following assumptions
1 The SN location is used as exact even though the true value may vary by upto plusmn01 degree from the location provided with the photo
2 The spacecraft altitude is used as exact Although the determination of the
nadir point at the instant of a known spacecraft vector is relatively precise(plusmn115times10 Otilde 4 degrees) the propagator interpolates between sets of approxi-mately 10ndash40 known vectors per day and the time code recorded on the lmcan drift Thus the true value for SN may vary by up to plusmn01 degree fromthe value provided with the photo
3 The perspective centre of the camera is assumed to be at the given altitudeover the speci ed spacecraft nadir location at the time of photo exposure
4 The PC location is used as exact even though the true value may vary by upto plusmn05deg latitude and plusmn05deg longitude from the location provided with the
photo5 A spherical earth model is used with a nominal radius of 6 372 16154 m (a
common rst-order approximation for a spherical earth used in geodeticcomputations)
6 The nominal lens focal length of the camera lens is used in the computations(calibrated focal length values are not available)
7 The photo projection is based on the classic pin-hole camera model8 No correction for lens distortion or atmospheric re ection is made
9 If no auxiliary point data is provided the lsquoTop of the Imagersquo is orientedperpendicular to the vector from SN towards PC
533 T ransformation from Earth to photo coordinate systemsThe calculations begin by converting the geographic coordinates ( latitude and
longitude) of the SN and PC to a Rectangular Earth-Centred Coordinate System(R-Earth) de ned as shown in gure 9 (with the centre of the Earth at (0 0 0))
Using the vector from the Earthrsquos centre through SN and the spacecraft altitude thespacecraft location (SC) is also computed in R-Earth
For ease of computation a Rectangular Spacecraft-Centred Coordinate System(R-Spacecraft ) is de ned as shown in gure 9 The origin of R-Spacecraft is located
at SC with its +Z-axis aligned with vector from the centre of the Earth throughSN and its +X-axis aligned with the vector from SN to PC ( gure 9) The speci c
J A Robinson et al4426
Figure 9 Representation of the area included in a photograph as a geometric projectiononto the Earthrsquos surface Note that the focal length (the distance from the SpaceShuttle to the photo) is grossly overscale compared to the altitude (the distance fromthe Shuttle to the surface of the Earth
rotations and translation used to convert from the R-Earth to the R-Spacecraft arecomputed and documented in the spreadsheet
With the mathematical positions of the SC SN PC and the centre of the Earthcomputed in R-Spacecraft the program next computes the location of the camerarsquosprincipal point (PP) The principle point is the point of intersection of the opticalaxis of the lens with the image plane (ie the lm) It is nominally positioned at a
Astronaut photography as digital data for remote sensing 4427
distance equal to the focal length from the perspective centre of the camera (whichis assumed to be at SC) along the vector from PC through SC as shown in gure 9
A third coordinate system the Rectangular Photo Coordinate System (R-Photo)is created with its origin at PP its XndashY axial plane normal to the vector from PCthrough SC and its +X axis aligned with the +X axis of R-Spacecraft as shownin gure 9 The XndashY plane of this coordinate system represents the image plane ofthe photograph
534 Auxiliary point calculationsThe calculations above employ a critical assumption that all photos are taken
with the lsquotop of the imagersquo oriented perpendicular to the vector from the SN towardsPC as shown in gure 9 To avoid non-uniform solar heating of the external surfacemost orbiting spacecraft are slowly and continually rotated about one or more axesIn this condition a ight crew member taking a photo while oating in microgravitycould orient the photo with practically any orientation relative to the horizon (seealso gure 8(b)) Unfortunately since these photos are taken with conventionalhandheld cameras there is no other information available which can be used toresolve the photorsquos rotational ambiguity about the optical axis other then the photoitself This is why the above assumption is used and the footprint computed by thiscalculator is actually a lsquolocator ellipsersquo which estimates the area on the ground whatis likely to be included in a speci c photograph (see gure 6) This locator ellipse ismost accurate for square image formats and subject to additional distortion as thephotograph format becomes more rectangular
If the user wants a more precise calculation of the image footprint the photorsquosrotational ambiguity about the optical axis must be resolved This can be done inthe calculator by adding data for an auxiliary point Detailed instructions regardinghow to prepare and use auxiliary point data in the computations are included in thelsquoHow-To-Usersquo tab of the spreadsheet Basically the user determines which side ofthe photograph is top and then measures the angle between the line from PP to thetop of the photo and from PP to the auxiliary point on the photo ( gure 9)
If the user includes data for an auxiliary point a series of computations arecompleted to resolve the photo rotation ambiguity about the optical axis (ie the
+Z axis in R-Photo) A vector from the Auxiliary Point on the Earth (AE) throughthe photograph perspective centre ( located at SC) is intersected with the photo imageplane (XndashY plane of R-Photo) to compute the coordinates of the Auxiliary Point onthe photo (AP) in R-Photo A two-dimensional angle in the XndashY plane of R-Photofrom the shy X axis to a line from PP to AP is calculated as shown in gure 9 The
shy X axis is used as the origin of the angle since it represents the top of the photoonce it passes through the perspective centre The diVerence between the computedangle and the angle measured by the user on the photo resolves the ambiguity inthe rotation of the photo relative to the principal line ( gures 7 and 9) Thetransformations from R-Spacecraft and R-Photo are then modi ed to include anadditional rotation angle about the +Z axis in R-Photo
535 lsquoFootprint rsquo calculationsThe program next computes the coordinates of eight points about the perimeter
of the image format (ie located at the four photo corners plus a bisector pointalong each edge of the image) These points are identi ed in R-Photo based uponthe photograph format size and then converted to R-Spacecraft Since all computa-tions are done in orthogonal coordinate systems the R-Spacecraft to R-Photo
J A Robinson et al4428
rotation matrix is transposed to produce an R-Photo to R-Spacecraft rotation matrixOnce in R-Spacecraft a unit vector from each of the eight perimeter points throughthe photo perspective centre (the same point as SC) is computed This provides thecoordinates for points about the perimeter of the image format with their directionvectors in a common coordinate system with other key points needed to computethe photo footprint
The next step is to compute the point of intersection between the spherical earthmodel and each of the eight perimeter point vectors The scalar value for eachperimeter point unit vector is computed using two-dimensional planar trigonometryAn angle c is computed using the formula for the cosine of an angle between twoincluded vectors (the perimeter point unit vector and the vector from SC to thecentre of the Earth) Angle y is computed using the Law of Sines Angle e=180degrees shy c shy y The scalar value of the perimeter point vector is computed using eand the Law of Cosines The scalar value is then multiplied by the perimeter pointunit vector to produce the three-dimensional point of intersection of the vector withEarthrsquos surface in R-Spacecraft The process is repeated independently for each ofthe eight perimeter point vectors Aside from its mathematical simplicity the valueof arcsine (computed in step 2) will exceed the normal range for the sin of an anglewhen the perimeter point vectors fail to intersect with the surface of the earth Asimple test based on this principle allows the program to correctly handle obliquephotos which image a portion of the horizon (see results for high oblique photographsin table 5)
The nal step in the process converts the eight earth intersection points from theR-Spacecraft to R-Earth The results are then converted to the standard geographiccoordinate system and displayed on the lsquoInput-Outputrsquo page of the spreadsheet
54 ExamplesAll three formulations were applied to the photographs included in this paper
and the results are compared in table 5 Formulation 1 gives too small a value fordistance across the photograph (D) for all but the most nadir shots and thus servesas an indicator of the best theoretical case but is not a good measure for a speci cphotograph For example the photograph of Lake Eyre taken from 276 km altitude( gure 2(a)) and the photograph of Limmen Bight ( gure 7) were closest to beingnadir views (oVsetslt68 km or tlt15deg table 5) For these photographs D calculatedusing Formulation 1 was similar to the minimum D calculated using Formulations2 and 3 For almost all other more oblique photographs Formulation 1 gave asigni cant underestimate of the distance covered in the photograph For gure 5(A)(the picture of Houston taken with a 40 mm lens) Formulation 1 did not give anunderestimate for D This is because Formulation 1 does not account for curvatureof the Earth in any way With this large eld of view assuming a at Earth in atedthe value of D above the minimum from calculations that included Earth curvature
A major diVerence between Formulations 2 and 3 is the ability to estimate pixelsizes (P) in both directions (along the principal line and perpendicular to the principalline) For the more oblique photographs the vertical estimate of D and pixel sizes ismuch larger than in the horizontal direction (eg the low oblique and high obliquephotographs of Hawaii gure 4 table 5)
For the area of Limmen Bight ( gure 7) table 5 illustrates the improvement inthe estimate of distance and pixel size that can be obtained by re-estimating thelocation of the PC with greater accuracy Centre points in the catalogued data are
Astronaut photography as digital data for remote sensing 4429
plusmn05deg of latitude and longitude When the centre point was re-estimated to plusmn002degwe determined that the photograph was not taken as obliquely as rst thought(change in the estimate of the look angle t from 169 to 128deg table 5) When theauxiliary point was added to the calculations of Formula 3 the calculated look angleshrank further to 110deg indicating that this photograph was taken at very close toa nadir view Of course this improvement in accuracy could have also led to estimatesof greater obliquity and corresponding larger pixel sizes
We also tested the performance of the scale calculator with auxiliary point byestimating the corner and point locations on the photograph using a 11 000 000Operational Navigational Chart For this test we estimated our ability to readcoordinates from the map as plusmn002degand our error in nding the locations of thecorner points as plusmn015deg (this error varies among photographs depending on thedetail that can be matched between photo and map) For Limmen Bight ( gure 7)the mean diVerence between map estimates and calculator estimates for four pointswas 031deg (SD=018 n=8) For a photograph of San Francisco Bay (STS062-151-291) the mean diVerence between map estimates and calculator estimates forfour points was 0064deg (SD=018 n=8) For a photograph of San Francisco Bay(STS062-151-291) the mean diVerence between map estimates and calculator estim-ates for eight points was 0196deg (SD=0146 n=16) Thus in one case the calculatorestimates were better than our estimate of the error in locating corner points on themap It is reasonable to expect that for nadir-viewing photographs the calculatorused with an auxiliary point can estimate locations of the edges of a photograph towithin plusmn03deg
55 Empirical con rmation of spatial resolution estimatesAs stated previously a challenge to estimating system-AWAR for astronaut
photography of Earth is the lack of suitable targets Small features in an image cansometimes be used as a check on the size of objects that can be successfully resolvedgiving an approximate value for GRD Similarly the number of pixels that make upthose features in the digitized image can be used to make an independent calculationof pixel size We have successfully used features such as roads and airport runwaysto make estimates of spatial scale and resolution (eg Robinson et al 2000c) Whilerecognizing that the use of linear detail in an image is a poor approximation to abar target and that linear objects smaller than the resolving power can often bedetected (Charman 1965) few objects other than roads could be found to make anydirect estimates of GRD Thus roads and runways were used in the images ofHouston (where one can readily conduct ground veri cations and where a numberof higher-contrast concrete roadways were available) to obtain empirical estimatesof GRD and pixel size for comparison with table 5
In the all-digital ESC image of Houston ( gure 3(d )) we examined EllingtonField runway 4-22 (centre left of the image) which is 24384 mtimes457 m This runwayis approximately 6ndash7 pixels in width and 304ndash3094 pixels in length so pixelsrepresent an distance on the ground 7ndash8 m Using a lower contrast measure of astreet length between two intersections (2125 m=211 pixels) a pixel width of 101 mis estimated These results compare favourably with the minimum estimate of 81 mpixels using Formulation 1 (table 5) For an estimate of GRD that would be morecomparable to aerial photography of a line target the smallest street without treecover that could be clearly distinguished on the photograph was 792 m wide Thesmallest non-street object (a gap between stages of the Saturn rocket on display in
J A Robinson et al4430
Tab
le5
App
lica
tio
nof
met
hod
sfo
res
tim
ati
ng
spati
alre
solu
tio
no
fast
ron
aut
ph
oto
gra
ph
sin
clu
din
gF
orm
ula
tion
s12
and
3Im
pro
ved
cen
tre
po
int
esti
mat
esan
dauxilia
ryp
oin
tca
lcu
lati
ons
are
sho
wn
for
gure
7V
aria
ble
sas
use
din
the
text
fle
ns
foca
lle
ngt
hH
al
titu
de
PC
ph
oto
gra
ph
cen
tre
poin
tSN
sp
ace
craft
nadir
po
int
D
dis
tance
cover
edo
nth
egr
oun
d
Pd
ista
nce
on
the
grou
nd
repre
sen
ted
by
apix
el
OVse
td
ista
nce
bet
wee
nP
Cand
SNusi
ng
the
grea
tci
rcle
form
ula
t
look
angle
away
from
SN
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
2L
ake
Eyre
ST
S093
-717
-80
250
276
shy29
0137
0shy
28
413
71
607
11
767
413
7609
ndash64
212
013
760
9ndash64
7121
124
ST
S082
-754
-61
250
617
shy28
5137
0shy
28
113
50
135
726
1201
18
013
8ndash14
827
517
9138ndash15
3277
294
Fig
ure
3H
ou
sto
nS
TS
062
-97-
151
4029
829
5shy
95
530
5shy
95
4410
78
8112
20
5348ndash58
990
1205
351
ndash63
8951
10
36
ST
S094
-737
-28
100
294
295
shy
95
528
3shy
95
5162
31
1133
24
415
8ndash20
334
724
3158ndash20
7351
390
ST
S055
-71-
41250
302
295
shy
95
028
2shy
93
766
412
8192
32
5736
ndash84
715
232
273
9ndash85
7154
185
S73
E51
45300
2685
295
shy
95
024
78
0F
igu
re4
Haw
aii
ST
S069
-701
-65
250
370
195
shy
155
520
4shy
153
881
415
7204
28
9876
ndash98
917
928
688
0ndash10
9181
210
ST
S069
-701
-70
250
370
210
shy
157
519
2shy
150
781
415
7738
63
414
9ndash23
336
760
7148ndash55
6394
10
63
ST
S069
-729
-27
100
398
200
shy
156
620
5shy
148
2219
42
1867
65
432
8ndash13
10
158
62
4323+
311
N
A
Astronaut photography as digital data for remote sensing 4431
Tab
le5
(Cont
inued
)
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
7L
imm
enB
igh
tS
TS
093
-704
-50
250
283
shy15
0135
0shy
15
013
58
623
120
859
16
963
0ndash67
3125
16
863
0ndash67
612
613
2P
CR
e-es
tim
ated
shy14
733
1352
67
64
5128
623
ndash65
5123
128
623
ndash65
712
3127
Auxilia
ryP
oin
t611
062
3ndash65
112
312
5F
igu
re10
N
ear
Au
stin
T
XS
TS
095
-716
-46
250
543
30
5shy
975
28
6shy
1007
120
230
375
34
613
5ndash157
281
34
1136ndash16
128
636
0
1 All
pho
togr
aphs
exce
pt
on
eta
ken
wit
hH
ass
elbla
dca
mer
aso
dis
tan
ceacr
oss
the
image
d=
55m
m
for
S73
E51
45
taken
wit
hel
ectr
onic
still
cam
erad=
276
5m
mho
rizo
nta
l2 P
Cand
SN
are
giv
enin
the
form
(lati
tud
elo
ngit
ude)
deg
rees
w
ith
neg
ativ
en
um
ber
sre
pre
senti
ng
south
and
wes
tA
ccura
cyo
fP
Cis
plusmn0
5and
SN
isplusmn
01
as
reco
rded
inth
eA
stro
naut
Ph
oto
gra
phy
Data
base
ex
cep
tw
her
en
ote
dth
atce
ntr
ep
oin
tw
asre
-est
imate
dw
ith
grea
ter
accu
racy
3 C
alc
ula
ted
usi
ng
the
mea
nofth
em
inim
um
an
dm
axim
um
esti
mate
sof
DF
orm
ula
tion
2co
nsi
der
sD
inth
ed
irec
tion
per
pen
dic
ula
rto
the
Pri
nci
pal
Lin
eo
nly
4 C
alc
ula
ted
inho
rizo
nta
lan
dve
rtic
aldir
ecti
on
sb
ut
still
ass
um
ing
geom
etry
of
gure
6(B
)F
irst
nu
mb
eris
the
mea
no
fth
em
inim
um
Dat
the
bott
om
of
the
ph
oto
and
the
maxi
mum
Dat
the
top
of
the
ph
oto
(range
show
nin
the
tab
le)
and
isco
mpara
ble
toF
orm
ula
tio
n2
Sec
ond
num
ber
isth
em
ean
of
Des
tim
ated
alon
gth
eP
rin
cip
alline
and
alo
ng
the
ou
tsid
eed
ge
(range
not
show
n)
5 Alt
itu
de
isap
pro
xim
ate
for
mis
sion
as
exac
tti
me
pho
togr
ap
hw
asta
ken
and
corr
esp
on
din
gn
adir
data
are
no
tava
ilab
le
Lack
of
nad
irdata
pre
ven
tsap
plica
tion
of
Form
ula
tion
s2
an
d3
6 Au
xilliar
ypo
int
add
edw
as
shy14
80
lati
tud
e13
51
2lo
ngit
ude
shy46
5degro
tati
on
angle
in
stru
ctio
ns
for
esti
mati
ng
the
rota
tio
nan
gle
para
met
erare
conta
ined
as
par
tof
the
scal
eca
lcu
lato
rsp
read
shee
tdo
cum
enta
tio
n
J A Robinson et al4432
a park at Johnson Space Center) that could clearly be distinguished on the photo-graph was 853 m wide
For the photograph of Houston taken with a 250 mm lens ( gure 3(C)) anddigitized from second generation lm at 2400 ppi (106 mm pixel Otilde 1 ) Ellington Fieldrunway 4-22 is 3 pixels in width and 1613 pixels in length so pixels represent151ndash152 m on the ground These results compare favourably with the minimumestimate of 152 m using Formulation 2 and 154ndash185 m pixels using Formulation 3(table 5) For an estimate of GRD using an 8times8 inch print (1369 enlargement) and4times magni cation the smallest street that could clearly be distinguished was 822 mwide the same feature could barely be distinguished on the digitized image
We also made an empirical estimate of spatial resolution for lower contrastvegetation boundaries By clearing forest so that a pattern would be visible to landingaircraft a landowner outside Austin Texas (see also aerial photo in Lisheron 2000)created a target that is also useful for evaluating spatial resolution of astronautphotographs The forest was selectively cleared in order to spell the landownerrsquosname lsquoLUECKErsquo with the remaining trees ( gure 10) According to local surveyorswho planned the clearing the plan was to create letters that were 3100 fttimes1700 ft(9449 mtimes5182 m) Photographed at a high altitude relative to most Shuttle missions(543 km) with a 250 mm lens Formula 3 predicts that each pixel would represent anarea 286 mtimes360 m on the ground (table 5) When original lm was digitized at2400 ppi (106 mm pixel Otilde 1 ) letters correspond to 294times188 pixels for a comparablepixel size of 27ndash32 m
6 Summary and conclusionsAstronaut photographs can be an excellent source of data for remote sensing
applications Best-case resolutions are similar to that for Landsat or SPOT withpixels as small as lt10 m It was estimated that of images taken to date 504 hadlens and obliquity characteristics that make them potential candidates for remotesensing information Digitized astronaut photographs can be overlaid with othersatellite data using GIS (Eckardt et al 2000) or used to ll in gaps in time serieswhen other imagery is not available As a source of public-domain information theycan be very useful for scientists who do not have access to satellite imagery eitherbecause of the expense of image acquisition (to the end user) or the computersystems needed for processing satellite images These diVerences in image acquisitioncosts (to the user) are summarized in table 6
Searching of the complete database of NASA astronaut photography includinglow-spatial-resolution browse images is available via the Web (OYce of EarthSciences 2000) This provides nearly global access for identifying images that willcontribute to a speci c scienti c project For the most detailed studies digitalproducts posted to the web will not be of suYcient quality To date we have madeit a practice of digitizing small numbers of images when requested by scientists atno charge Such requests (including a description of the project involved) can bemade through the authors or using the contact information listed on the websiteInvestigators with needs for larger numbers of digital images have also been servedthrough collaborative agreements
In our experience scientists that nd the data most valuable are those in develop-ing countries those studying areas that have not usually been targets for the majorsatellites those having diYculty in nding low-cloud images those interested inconstructing time series or those interested in using a large number of images
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
Astronaut photography as digital data for remote sensing 4419
photo that required higher spatial resolution digitizing (it had been taken with ashorter lens and thus had less spatial resolution) Part of the equivalence observedin the urban areas study may be attributable to the fact that the atbed scannerused interpolates from 1200 ppi to 2400 ppi in one direction The appropriate digitiz-ing spatial resolution will in part depend on the scale of features of interest to theresearchers with a maximum upper limit set of a scan spot size of approximately6 mm (4233 ppi)
442 Digitizing and spectral resolutionWhen colour lm is digitized there will be loss of spectral resolution The three
lm emulsion layers (red green blue) have relatively distinct spectral responses butare fused together so that digitizers detect one colour signal and must convert itback into three (red green blue) digital channels Digital scanning is also subject tospectral calibration and reproduction errors Studies using digital astronaut photo-graphs to date have used 8 bits per channel However this is another parameterthat can be controlled when the lm is digitized We do not further address spectralresolution of digitally scanned images in this paper
45 External factors that in uence GRDAlthough not discussed in detail in this paper factors external to the spacecraft
and camera system (as listed by Moik (1980) for remote sensing in general) alsoimpact GRD These include atmospheric interference (due to scattering attenuationhaze) variable surface illumination (diVerences in terrain slope and orientation) andchange of terrain re ectance with viewing angle (bidirectional re ectance) For astro-naut photographs variable illumination is particularly important because orbits arenot sun-synchronous Photographs are illuminated by diVerent sun angles and imagesof a given location will have colour intensities that vary widely In addition theviewing angle has an eVect on the degree of object-to-backgroun d contrast andatmospheric interference
5 Estimating spatial resolution of astronaut photographsIn order to use astronaut photographs for digital remote sensing it is important
to be able to calculate the equivalent to an IFOVmdashthe ground area represented bya single pixel in a digitized orbital photograph The obliquity of most photographsmeans that pixel lsquosizesrsquo vary at diVerent places in an image Given a ground distanceD represented by a photograph in each direction (horizontal vertical ) an approxi-mate average pixel width (P the equivalent of IFOV) for the entire image can becalculated as follows
P=D
dS(2)
where D is the projected distance on the ground covered by the image in the samedirection as the pixel is measured d is the actual width of the image on the original lm (table 1) and S is the digitizing spatial resolution
Here we present three mathematical formulations for estimating the size of thefootprint or area on the ground covered by the image Example results from theapplication of all three formulations are given in table 5 The rst and simplestcalculation (formulation 1) gives an idea of the maximum spatial resolution attainableat a given altitude of orbit with a given lm format and a perfectly vertical (nadir)
J A Robinson et al4420
view downward Formulation 2 takes into account obliquity by calculating lookangle from the diVerence between the location on the ground represented at thecentre of the photograph and the nadir location of the spacecraft at the time thephotograph was taken ( gure 1) Formulation 3 describes an alternate solution tothe oblique look angle problem using coordinate-system transformations Thisformulation has been implemented in a documented spreadsheet and is availablefor download (httpeoljscnasagovsseopFootprintCalculatorxls) and is in theprocess of being implemented on a Web-based user interface to the AstronautPhotography Database (OYce of Earth Sciences 2000)
Although formulations 2 and 3 account for obliquity for the purposes of calcula-tion they treat the position of the spacecraft and position of the camera as one Inactuality astronauts are generally holding the cameras by hand (although camerasbracketed in the window are also used) and the selection of window position of theastronaut in the window and rotation of the camera relative to the movement ofthe spacecraft are not known Thus calculations using only the photo centre point(PC) and spacecraft nadir point (SN) give a locator ellipse and not the locations ofthe corners of the photograph A locator ellipse describes an estimated area on theground that is likely to be included in a speci c photograph regardless of the rotationof the lm plane about the camerarsquos optical axis ( gure 6)
Estimating the corner positions of the photo requires additional user input of asingle auxiliary pointmdasha location on the image that has a known location on theground Addition of this auxiliary point is an option available to users of thespreadsheet An example of the results of adding an auxiliary point is shown in gure 7 with comparisons of the various calculations in table 5
51 Formulation 1 Footprint for a nadir viewThe simplest way to estimate footprint size is to use the geometry of camera lens
and spacecraft altitude to calculate the scaling relationship between the image in the
Figure 6 Sketch representation of the locator ellipse an estimated area on the ground thatis likely to be included in a speci c photograph regardless of the rotation of the lmplane about the camerarsquos optical axis
Astronaut photography as digital data for remote sensing 4421
Figure 7 An example of the application of the Low Oblique Space Photo FootprintCalculator The photograph (STS093-704-50) of Limmen Bight Australia was takenfrom an altitude of 283 km using the Hasselblad camera with a 250 mm lens Mapused is 11 000 000 Operational Navigational Chart The distances shown equate toan average pixel size of 124 m for this photograph
lm and the area covered on the ground For a perfect nadir view the scalerelationship from the geometry of similar triangles is
d
D=
f
H(3)
where d=original image size D=distance of footprint on the ground f =focallength of lens and H=altitude Once D is known pixel size ( length=width) can becalculated from equation (2) These calculations represent the minimum footprintand minimum pixel size possible for a given camera system altitude and digitisingspatial resolution (table 2) Formulation 1 was used for calculating minimum pixelsizes shown in tables 2 and 4
By assuming digitizing at 2400 ppi (106 mm pixel Otilde 1 ) currently a spatial resolutioncommonly attainable from multipurpose colour transparency scanners (see sect441)this formulation was used to convert area covered to an IFOV equivalent for missionsof diVerent altitudes (table 2) Table 4 provides a comparison of IFOV of imagesfrom various satellites including the equivalent for astronaut photography Valuesin this table were derived by using formulation 1 to estimate the area covered becausea perfect nadir view represents the best possible spatial resolution and smallest eldof view that could be obtained
52 Formulation 2 Footprint for oblique views using simpli ed geometry and thegreat circle distance
A more realistic approach to determining the footprint of the photographaccounts for the fact that the camera is not usually pointing perfectly down at the
J A Robinson et al4422
Table 4 Comparison of pixel sizes (instantaneous elds of view) for satellite remote sensorswith pixel sizes for near vertical viewing photography from spacecraft Note that alldata returned from the automated satellites have the same eld of view while indicatedpixel sizes for astronaut photography are best-case for near vertical look angles See gure 5 for distributions of astronaut photographs of various look angles High-altitude aerial photography is also included for reference
Altitude PixelInstrument (km) width (mtimesm) Bands1
Landsat Multipectral Scanner2 (MSS 1974-) 880ndash940 79times56 4ndash5Landsat Thematic Mapper (TM 1982-) ~705 30times30 7Satellite pour lrsquoObservation de la Terre (SPOT 1986-) ~832
SPOT 1-3 Panchromatic 10times10 1SPOT 1-3 Multispectral (XS) 20times20 3
Astronaut Photography (1961-)Skylab3 (1973-1974 ) ~435
Multispectral Camera (6 camera stations) 17ndash19times17ndash19 3ndash8Earth Terrain Camera (hi-res colour) 875times875 3
Space Shuttle5 (1981-) 222ndash611Hasselblad 70 mm (250mm lens colour) vertical view 9ndash26times9ndash26 3Hasselblad 70 mm (100mm lens colour) vertical view 23ndash65times23ndash65 3Nikon 35 mm (400mm lens colour lm or ESC) vertical 5ndash16times5ndash16 3
High Altitude Aerial Photograph (1100 000) ~12 2times2 3
1Three bands listed for aerial photography obtainable by high-resolution colour digitizingFor high-resolution colour lm at 65 lp mmOtilde 1 (K Teitelbaum Eastman Kodak Co personalcommunication) the maximum information contained would be 3302 ppi
2Data from Simonett (1983Table 1-2)3Calculated as recommended by Jensen (19831596) the dividend of estimated ground
resolution at low contrast given in NASA (1974179-180) and 244Six lters plus colour and colour infrared lm (NASA 19747) 5Extracted from table 2
nadir point The look angle (the angle oV nadir that the camera is pointing) can becalculated trigonometrically by assuming a spherical earth and calculating the dis-tance between the coordinates of SN and PC ( gure 1) using the Great Circledistance haversine solution (Sinnott 1984 Snyder 198730ndash32 Chamberlain 1996)The diVerence between the spacecraft centre and nadir point latitudes Dlat=lat 2 shy lat 1 and the diVerence between the spacecraft centre and nadir pointlongitudes Dlon=lon 2 shy lon 1 enter the following equations
a=sin2ADlat
2 B+cos(lat 1) cos(lat 2) sin2ADlon
2 B (4)
c=2 arcsin(min[1 atilde a]) (5)
and oVset=R c with R the radius of a spherical Earth=6370 kmThe look angle (t ) is then given by
t=arctanAoffset
H B (6)
Assuming that the camera was positioned so that the imaginary line between thecentre and nadir points (the principal line) runs vertically through the centre of thephotograph the distance between the geometric centre of the photograph (principalpoint PP) and the top of the photograph is d2 ( gure 8(a)) The scale at any point
Astronaut photography as digital data for remote sensing 4423
(a) (b)
Figure 8 The geometric relationship between look angle lens focal length and the centrepoint and nadir points of a photograph (see also Estes and Simonett 1975) Note thatthe Earthrsquos surface and footprint represented in gure 1 are not shown here only arepresentation of the image area of the photograph Variables shown in the gurelook angle t focal length f PP centre of the image on the lm SN spacecraft nadird original image size (a) The special case where the camera is aligned squarely withthe isometric parallel In this case tilt occurs along a plane parallel with the centre ofthe photograph When the conditions are met calculations using Formulation 3 willgive the ground coordinates of the corner points of the photograph (b) The generalrelationship between these parameters and key photogrammetric points on the photo-graph For this more typical case coordinates of an additional point in the photographmust be input in order to adjust for twist of the camera and to nd the corner pointcoordinates
in the photograph varies as a function of the distance y along the principal linebetween the isocentre and the point according to the relationship
d
D=
f shy ysint
H(7)
(Wong 1980 equation (214) H amp the ground elevation h) Using gure 8(a) at thetop of the photo
y= f tanAt
2B+d
2(8)
and at the bottom of the photo
y= f tanAt
2Bshyd
2(9)
Thus for given PC and SN coordinates and assuming a photo orientation as in gure 8(a) and not gure 8(b) we can estimate a minimum D (using equations (7)and (8)) and a maximum D (using equations (7) and (9)) and then average the twoto determine the pixel size (P ) via equation (2)
J A Robinson et al4424
53 Formulation 3 T he L ow Oblique Space Photo Footprint CalculatorThe Low Oblique Space Photo Footprint Calculator was developed to provide
a more accurate estimation of the geographic coordinates for the footprint of a lowoblique photo of the Earthrsquos surface taken from a human-occupied spacecraft inorbit The calculator performs a series of three-dimensional coordinate transforma-tions to compute the location and orientation of the centre of the photo exposureplane relative to an earth referenced coordinate system The nominal camera focallength is then used to create a vector from the photorsquos perspective point througheach of eight points around the parameter of the image as de ned by the formatsize The geographical coordinates for the photo footprint are then computed byintersecting these photo vectors with a spherical earth model Although more sophist-icated projection algorithms are available no signi cant increase in the accuracy ofthe results would be produced by these algorithms due to inherent uncertainties inthe available input data (ie the spacecraft altitude photo centre location etc)
The calculations were initially implemented within a Microsoft Excel workbookwhich allowed one to embed the mathematical and graphical documentation nextto the actual calculations Thus interested users can inspect the mathematical pro-cessing A set of error traps was also built into the calculations to detect erroneousresults A summary of any errors generated is reported to the user with the resultsof the calculations Interested individuals are invited to download the Excel work-book from httpeoljscnasagovsseopFootprintCalculatorxls The calculations arecurrently being encoded in a high-level programming language and should soon beavailable alongside other background data provided for each photograph at OYceof Earth Sciences (2000)
For the purposes of these calculations a low oblique photograph was de ned as
one with the centre within 10 degrees of latitude and longitude of the spacecraftnadir point For the typical range of spacecraft altitudes to date this restricted thecalculations to photographs in which Earthrsquos horizon does not appear (the generalde nition of a low oblique photograph eg Campbell 199671 and gure 4)
531 Input data and resultsUpon opening the Excel workbook the user is presented with a program intro-
duction providing instructions for using the calculator The second worksheet tablsquoHow-To-Usersquo provides detailed step-by-step instructions for preparing the baselinedata for the program The third tab lsquoInput-Output rsquo contains the user input eldsand displays the results of the calculations The additional worksheets contain theactual calculations and program documentation Although users are welcome toreview these sheets an experienced user need only access the lsquoInput-Outputrsquo
spreadsheetTo begin a calculation the user enters the following information which is available
for each photo in the NASA Astronaut Photography Database (1) SN geographicalcoordinates of spacecraft nadir position at the time of photo (2) H spacecraftaltitude (3) PC the geographical coordinates of the centre of the photo (4) f nominal focal length and (5) d image format size The automatic implementationof the workbook on the web will automatically enter these values and completecalculations
For more accurate results the user may optionally enter the geographic co-ordinates and orientation for an auxiliary point on the photo which resolves thecamerarsquos rotation uncertainty about the optical axis The auxiliary point data must
Astronaut photography as digital data for remote sensing 4425
be computed by the user following the instruction contained in the lsquoHow-To-Usersquotab of the spreadsheet
After entering the input data the geographic coordinates of the photo footprint(ie four photo corner points four points at the bisector of each edge and the centreof the photo) are immediately displayed below the input elds along with any errormessages generated by the user input or by the calculations ( gure 7) Although
results are computed and displayed they should not be used when error messagesare produced by the program The program also computes the tilt angle for each ofthe photo vectors relative to the spacecraft nadir vector To the right of the photofootprint coordinates is displayed the arc distance along the surface of the sphere
between adjacent computed points
532 Calculation assumptions
The mathematical calculations implemented in the Low Oblique Space PhotoFootprint Calculator use the following assumptions
1 The SN location is used as exact even though the true value may vary by upto plusmn01 degree from the location provided with the photo
2 The spacecraft altitude is used as exact Although the determination of the
nadir point at the instant of a known spacecraft vector is relatively precise(plusmn115times10 Otilde 4 degrees) the propagator interpolates between sets of approxi-mately 10ndash40 known vectors per day and the time code recorded on the lmcan drift Thus the true value for SN may vary by up to plusmn01 degree fromthe value provided with the photo
3 The perspective centre of the camera is assumed to be at the given altitudeover the speci ed spacecraft nadir location at the time of photo exposure
4 The PC location is used as exact even though the true value may vary by upto plusmn05deg latitude and plusmn05deg longitude from the location provided with the
photo5 A spherical earth model is used with a nominal radius of 6 372 16154 m (a
common rst-order approximation for a spherical earth used in geodeticcomputations)
6 The nominal lens focal length of the camera lens is used in the computations(calibrated focal length values are not available)
7 The photo projection is based on the classic pin-hole camera model8 No correction for lens distortion or atmospheric re ection is made
9 If no auxiliary point data is provided the lsquoTop of the Imagersquo is orientedperpendicular to the vector from SN towards PC
533 T ransformation from Earth to photo coordinate systemsThe calculations begin by converting the geographic coordinates ( latitude and
longitude) of the SN and PC to a Rectangular Earth-Centred Coordinate System(R-Earth) de ned as shown in gure 9 (with the centre of the Earth at (0 0 0))
Using the vector from the Earthrsquos centre through SN and the spacecraft altitude thespacecraft location (SC) is also computed in R-Earth
For ease of computation a Rectangular Spacecraft-Centred Coordinate System(R-Spacecraft ) is de ned as shown in gure 9 The origin of R-Spacecraft is located
at SC with its +Z-axis aligned with vector from the centre of the Earth throughSN and its +X-axis aligned with the vector from SN to PC ( gure 9) The speci c
J A Robinson et al4426
Figure 9 Representation of the area included in a photograph as a geometric projectiononto the Earthrsquos surface Note that the focal length (the distance from the SpaceShuttle to the photo) is grossly overscale compared to the altitude (the distance fromthe Shuttle to the surface of the Earth
rotations and translation used to convert from the R-Earth to the R-Spacecraft arecomputed and documented in the spreadsheet
With the mathematical positions of the SC SN PC and the centre of the Earthcomputed in R-Spacecraft the program next computes the location of the camerarsquosprincipal point (PP) The principle point is the point of intersection of the opticalaxis of the lens with the image plane (ie the lm) It is nominally positioned at a
Astronaut photography as digital data for remote sensing 4427
distance equal to the focal length from the perspective centre of the camera (whichis assumed to be at SC) along the vector from PC through SC as shown in gure 9
A third coordinate system the Rectangular Photo Coordinate System (R-Photo)is created with its origin at PP its XndashY axial plane normal to the vector from PCthrough SC and its +X axis aligned with the +X axis of R-Spacecraft as shownin gure 9 The XndashY plane of this coordinate system represents the image plane ofthe photograph
534 Auxiliary point calculationsThe calculations above employ a critical assumption that all photos are taken
with the lsquotop of the imagersquo oriented perpendicular to the vector from the SN towardsPC as shown in gure 9 To avoid non-uniform solar heating of the external surfacemost orbiting spacecraft are slowly and continually rotated about one or more axesIn this condition a ight crew member taking a photo while oating in microgravitycould orient the photo with practically any orientation relative to the horizon (seealso gure 8(b)) Unfortunately since these photos are taken with conventionalhandheld cameras there is no other information available which can be used toresolve the photorsquos rotational ambiguity about the optical axis other then the photoitself This is why the above assumption is used and the footprint computed by thiscalculator is actually a lsquolocator ellipsersquo which estimates the area on the ground whatis likely to be included in a speci c photograph (see gure 6) This locator ellipse ismost accurate for square image formats and subject to additional distortion as thephotograph format becomes more rectangular
If the user wants a more precise calculation of the image footprint the photorsquosrotational ambiguity about the optical axis must be resolved This can be done inthe calculator by adding data for an auxiliary point Detailed instructions regardinghow to prepare and use auxiliary point data in the computations are included in thelsquoHow-To-Usersquo tab of the spreadsheet Basically the user determines which side ofthe photograph is top and then measures the angle between the line from PP to thetop of the photo and from PP to the auxiliary point on the photo ( gure 9)
If the user includes data for an auxiliary point a series of computations arecompleted to resolve the photo rotation ambiguity about the optical axis (ie the
+Z axis in R-Photo) A vector from the Auxiliary Point on the Earth (AE) throughthe photograph perspective centre ( located at SC) is intersected with the photo imageplane (XndashY plane of R-Photo) to compute the coordinates of the Auxiliary Point onthe photo (AP) in R-Photo A two-dimensional angle in the XndashY plane of R-Photofrom the shy X axis to a line from PP to AP is calculated as shown in gure 9 The
shy X axis is used as the origin of the angle since it represents the top of the photoonce it passes through the perspective centre The diVerence between the computedangle and the angle measured by the user on the photo resolves the ambiguity inthe rotation of the photo relative to the principal line ( gures 7 and 9) Thetransformations from R-Spacecraft and R-Photo are then modi ed to include anadditional rotation angle about the +Z axis in R-Photo
535 lsquoFootprint rsquo calculationsThe program next computes the coordinates of eight points about the perimeter
of the image format (ie located at the four photo corners plus a bisector pointalong each edge of the image) These points are identi ed in R-Photo based uponthe photograph format size and then converted to R-Spacecraft Since all computa-tions are done in orthogonal coordinate systems the R-Spacecraft to R-Photo
J A Robinson et al4428
rotation matrix is transposed to produce an R-Photo to R-Spacecraft rotation matrixOnce in R-Spacecraft a unit vector from each of the eight perimeter points throughthe photo perspective centre (the same point as SC) is computed This provides thecoordinates for points about the perimeter of the image format with their directionvectors in a common coordinate system with other key points needed to computethe photo footprint
The next step is to compute the point of intersection between the spherical earthmodel and each of the eight perimeter point vectors The scalar value for eachperimeter point unit vector is computed using two-dimensional planar trigonometryAn angle c is computed using the formula for the cosine of an angle between twoincluded vectors (the perimeter point unit vector and the vector from SC to thecentre of the Earth) Angle y is computed using the Law of Sines Angle e=180degrees shy c shy y The scalar value of the perimeter point vector is computed using eand the Law of Cosines The scalar value is then multiplied by the perimeter pointunit vector to produce the three-dimensional point of intersection of the vector withEarthrsquos surface in R-Spacecraft The process is repeated independently for each ofthe eight perimeter point vectors Aside from its mathematical simplicity the valueof arcsine (computed in step 2) will exceed the normal range for the sin of an anglewhen the perimeter point vectors fail to intersect with the surface of the earth Asimple test based on this principle allows the program to correctly handle obliquephotos which image a portion of the horizon (see results for high oblique photographsin table 5)
The nal step in the process converts the eight earth intersection points from theR-Spacecraft to R-Earth The results are then converted to the standard geographiccoordinate system and displayed on the lsquoInput-Outputrsquo page of the spreadsheet
54 ExamplesAll three formulations were applied to the photographs included in this paper
and the results are compared in table 5 Formulation 1 gives too small a value fordistance across the photograph (D) for all but the most nadir shots and thus servesas an indicator of the best theoretical case but is not a good measure for a speci cphotograph For example the photograph of Lake Eyre taken from 276 km altitude( gure 2(a)) and the photograph of Limmen Bight ( gure 7) were closest to beingnadir views (oVsetslt68 km or tlt15deg table 5) For these photographs D calculatedusing Formulation 1 was similar to the minimum D calculated using Formulations2 and 3 For almost all other more oblique photographs Formulation 1 gave asigni cant underestimate of the distance covered in the photograph For gure 5(A)(the picture of Houston taken with a 40 mm lens) Formulation 1 did not give anunderestimate for D This is because Formulation 1 does not account for curvatureof the Earth in any way With this large eld of view assuming a at Earth in atedthe value of D above the minimum from calculations that included Earth curvature
A major diVerence between Formulations 2 and 3 is the ability to estimate pixelsizes (P) in both directions (along the principal line and perpendicular to the principalline) For the more oblique photographs the vertical estimate of D and pixel sizes ismuch larger than in the horizontal direction (eg the low oblique and high obliquephotographs of Hawaii gure 4 table 5)
For the area of Limmen Bight ( gure 7) table 5 illustrates the improvement inthe estimate of distance and pixel size that can be obtained by re-estimating thelocation of the PC with greater accuracy Centre points in the catalogued data are
Astronaut photography as digital data for remote sensing 4429
plusmn05deg of latitude and longitude When the centre point was re-estimated to plusmn002degwe determined that the photograph was not taken as obliquely as rst thought(change in the estimate of the look angle t from 169 to 128deg table 5) When theauxiliary point was added to the calculations of Formula 3 the calculated look angleshrank further to 110deg indicating that this photograph was taken at very close toa nadir view Of course this improvement in accuracy could have also led to estimatesof greater obliquity and corresponding larger pixel sizes
We also tested the performance of the scale calculator with auxiliary point byestimating the corner and point locations on the photograph using a 11 000 000Operational Navigational Chart For this test we estimated our ability to readcoordinates from the map as plusmn002degand our error in nding the locations of thecorner points as plusmn015deg (this error varies among photographs depending on thedetail that can be matched between photo and map) For Limmen Bight ( gure 7)the mean diVerence between map estimates and calculator estimates for four pointswas 031deg (SD=018 n=8) For a photograph of San Francisco Bay (STS062-151-291) the mean diVerence between map estimates and calculator estimates forfour points was 0064deg (SD=018 n=8) For a photograph of San Francisco Bay(STS062-151-291) the mean diVerence between map estimates and calculator estim-ates for eight points was 0196deg (SD=0146 n=16) Thus in one case the calculatorestimates were better than our estimate of the error in locating corner points on themap It is reasonable to expect that for nadir-viewing photographs the calculatorused with an auxiliary point can estimate locations of the edges of a photograph towithin plusmn03deg
55 Empirical con rmation of spatial resolution estimatesAs stated previously a challenge to estimating system-AWAR for astronaut
photography of Earth is the lack of suitable targets Small features in an image cansometimes be used as a check on the size of objects that can be successfully resolvedgiving an approximate value for GRD Similarly the number of pixels that make upthose features in the digitized image can be used to make an independent calculationof pixel size We have successfully used features such as roads and airport runwaysto make estimates of spatial scale and resolution (eg Robinson et al 2000c) Whilerecognizing that the use of linear detail in an image is a poor approximation to abar target and that linear objects smaller than the resolving power can often bedetected (Charman 1965) few objects other than roads could be found to make anydirect estimates of GRD Thus roads and runways were used in the images ofHouston (where one can readily conduct ground veri cations and where a numberof higher-contrast concrete roadways were available) to obtain empirical estimatesof GRD and pixel size for comparison with table 5
In the all-digital ESC image of Houston ( gure 3(d )) we examined EllingtonField runway 4-22 (centre left of the image) which is 24384 mtimes457 m This runwayis approximately 6ndash7 pixels in width and 304ndash3094 pixels in length so pixelsrepresent an distance on the ground 7ndash8 m Using a lower contrast measure of astreet length between two intersections (2125 m=211 pixels) a pixel width of 101 mis estimated These results compare favourably with the minimum estimate of 81 mpixels using Formulation 1 (table 5) For an estimate of GRD that would be morecomparable to aerial photography of a line target the smallest street without treecover that could be clearly distinguished on the photograph was 792 m wide Thesmallest non-street object (a gap between stages of the Saturn rocket on display in
J A Robinson et al4430
Tab
le5
App
lica
tio
nof
met
hod
sfo
res
tim
ati
ng
spati
alre
solu
tio
no
fast
ron
aut
ph
oto
gra
ph
sin
clu
din
gF
orm
ula
tion
s12
and
3Im
pro
ved
cen
tre
po
int
esti
mat
esan
dauxilia
ryp
oin
tca
lcu
lati
ons
are
sho
wn
for
gure
7V
aria
ble
sas
use
din
the
text
fle
ns
foca
lle
ngt
hH
al
titu
de
PC
ph
oto
gra
ph
cen
tre
poin
tSN
sp
ace
craft
nadir
po
int
D
dis
tance
cover
edo
nth
egr
oun
d
Pd
ista
nce
on
the
grou
nd
repre
sen
ted
by
apix
el
OVse
td
ista
nce
bet
wee
nP
Cand
SNusi
ng
the
grea
tci
rcle
form
ula
t
look
angle
away
from
SN
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
2L
ake
Eyre
ST
S093
-717
-80
250
276
shy29
0137
0shy
28
413
71
607
11
767
413
7609
ndash64
212
013
760
9ndash64
7121
124
ST
S082
-754
-61
250
617
shy28
5137
0shy
28
113
50
135
726
1201
18
013
8ndash14
827
517
9138ndash15
3277
294
Fig
ure
3H
ou
sto
nS
TS
062
-97-
151
4029
829
5shy
95
530
5shy
95
4410
78
8112
20
5348ndash58
990
1205
351
ndash63
8951
10
36
ST
S094
-737
-28
100
294
295
shy
95
528
3shy
95
5162
31
1133
24
415
8ndash20
334
724
3158ndash20
7351
390
ST
S055
-71-
41250
302
295
shy
95
028
2shy
93
766
412
8192
32
5736
ndash84
715
232
273
9ndash85
7154
185
S73
E51
45300
2685
295
shy
95
024
78
0F
igu
re4
Haw
aii
ST
S069
-701
-65
250
370
195
shy
155
520
4shy
153
881
415
7204
28
9876
ndash98
917
928
688
0ndash10
9181
210
ST
S069
-701
-70
250
370
210
shy
157
519
2shy
150
781
415
7738
63
414
9ndash23
336
760
7148ndash55
6394
10
63
ST
S069
-729
-27
100
398
200
shy
156
620
5shy
148
2219
42
1867
65
432
8ndash13
10
158
62
4323+
311
N
A
Astronaut photography as digital data for remote sensing 4431
Tab
le5
(Cont
inued
)
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
7L
imm
enB
igh
tS
TS
093
-704
-50
250
283
shy15
0135
0shy
15
013
58
623
120
859
16
963
0ndash67
3125
16
863
0ndash67
612
613
2P
CR
e-es
tim
ated
shy14
733
1352
67
64
5128
623
ndash65
5123
128
623
ndash65
712
3127
Auxilia
ryP
oin
t611
062
3ndash65
112
312
5F
igu
re10
N
ear
Au
stin
T
XS
TS
095
-716
-46
250
543
30
5shy
975
28
6shy
1007
120
230
375
34
613
5ndash157
281
34
1136ndash16
128
636
0
1 All
pho
togr
aphs
exce
pt
on
eta
ken
wit
hH
ass
elbla
dca
mer
aso
dis
tan
ceacr
oss
the
image
d=
55m
m
for
S73
E51
45
taken
wit
hel
ectr
onic
still
cam
erad=
276
5m
mho
rizo
nta
l2 P
Cand
SN
are
giv
enin
the
form
(lati
tud
elo
ngit
ude)
deg
rees
w
ith
neg
ativ
en
um
ber
sre
pre
senti
ng
south
and
wes
tA
ccura
cyo
fP
Cis
plusmn0
5and
SN
isplusmn
01
as
reco
rded
inth
eA
stro
naut
Ph
oto
gra
phy
Data
base
ex
cep
tw
her
en
ote
dth
atce
ntr
ep
oin
tw
asre
-est
imate
dw
ith
grea
ter
accu
racy
3 C
alc
ula
ted
usi
ng
the
mea
nofth
em
inim
um
an
dm
axim
um
esti
mate
sof
DF
orm
ula
tion
2co
nsi
der
sD
inth
ed
irec
tion
per
pen
dic
ula
rto
the
Pri
nci
pal
Lin
eo
nly
4 C
alc
ula
ted
inho
rizo
nta
lan
dve
rtic
aldir
ecti
on
sb
ut
still
ass
um
ing
geom
etry
of
gure
6(B
)F
irst
nu
mb
eris
the
mea
no
fth
em
inim
um
Dat
the
bott
om
of
the
ph
oto
and
the
maxi
mum
Dat
the
top
of
the
ph
oto
(range
show
nin
the
tab
le)
and
isco
mpara
ble
toF
orm
ula
tio
n2
Sec
ond
num
ber
isth
em
ean
of
Des
tim
ated
alon
gth
eP
rin
cip
alline
and
alo
ng
the
ou
tsid
eed
ge
(range
not
show
n)
5 Alt
itu
de
isap
pro
xim
ate
for
mis
sion
as
exac
tti
me
pho
togr
ap
hw
asta
ken
and
corr
esp
on
din
gn
adir
data
are
no
tava
ilab
le
Lack
of
nad
irdata
pre
ven
tsap
plica
tion
of
Form
ula
tion
s2
an
d3
6 Au
xilliar
ypo
int
add
edw
as
shy14
80
lati
tud
e13
51
2lo
ngit
ude
shy46
5degro
tati
on
angle
in
stru
ctio
ns
for
esti
mati
ng
the
rota
tio
nan
gle
para
met
erare
conta
ined
as
par
tof
the
scal
eca
lcu
lato
rsp
read
shee
tdo
cum
enta
tio
n
J A Robinson et al4432
a park at Johnson Space Center) that could clearly be distinguished on the photo-graph was 853 m wide
For the photograph of Houston taken with a 250 mm lens ( gure 3(C)) anddigitized from second generation lm at 2400 ppi (106 mm pixel Otilde 1 ) Ellington Fieldrunway 4-22 is 3 pixels in width and 1613 pixels in length so pixels represent151ndash152 m on the ground These results compare favourably with the minimumestimate of 152 m using Formulation 2 and 154ndash185 m pixels using Formulation 3(table 5) For an estimate of GRD using an 8times8 inch print (1369 enlargement) and4times magni cation the smallest street that could clearly be distinguished was 822 mwide the same feature could barely be distinguished on the digitized image
We also made an empirical estimate of spatial resolution for lower contrastvegetation boundaries By clearing forest so that a pattern would be visible to landingaircraft a landowner outside Austin Texas (see also aerial photo in Lisheron 2000)created a target that is also useful for evaluating spatial resolution of astronautphotographs The forest was selectively cleared in order to spell the landownerrsquosname lsquoLUECKErsquo with the remaining trees ( gure 10) According to local surveyorswho planned the clearing the plan was to create letters that were 3100 fttimes1700 ft(9449 mtimes5182 m) Photographed at a high altitude relative to most Shuttle missions(543 km) with a 250 mm lens Formula 3 predicts that each pixel would represent anarea 286 mtimes360 m on the ground (table 5) When original lm was digitized at2400 ppi (106 mm pixel Otilde 1 ) letters correspond to 294times188 pixels for a comparablepixel size of 27ndash32 m
6 Summary and conclusionsAstronaut photographs can be an excellent source of data for remote sensing
applications Best-case resolutions are similar to that for Landsat or SPOT withpixels as small as lt10 m It was estimated that of images taken to date 504 hadlens and obliquity characteristics that make them potential candidates for remotesensing information Digitized astronaut photographs can be overlaid with othersatellite data using GIS (Eckardt et al 2000) or used to ll in gaps in time serieswhen other imagery is not available As a source of public-domain information theycan be very useful for scientists who do not have access to satellite imagery eitherbecause of the expense of image acquisition (to the end user) or the computersystems needed for processing satellite images These diVerences in image acquisitioncosts (to the user) are summarized in table 6
Searching of the complete database of NASA astronaut photography includinglow-spatial-resolution browse images is available via the Web (OYce of EarthSciences 2000) This provides nearly global access for identifying images that willcontribute to a speci c scienti c project For the most detailed studies digitalproducts posted to the web will not be of suYcient quality To date we have madeit a practice of digitizing small numbers of images when requested by scientists atno charge Such requests (including a description of the project involved) can bemade through the authors or using the contact information listed on the websiteInvestigators with needs for larger numbers of digital images have also been servedthrough collaborative agreements
In our experience scientists that nd the data most valuable are those in develop-ing countries those studying areas that have not usually been targets for the majorsatellites those having diYculty in nding low-cloud images those interested inconstructing time series or those interested in using a large number of images
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
J A Robinson et al4420
view downward Formulation 2 takes into account obliquity by calculating lookangle from the diVerence between the location on the ground represented at thecentre of the photograph and the nadir location of the spacecraft at the time thephotograph was taken ( gure 1) Formulation 3 describes an alternate solution tothe oblique look angle problem using coordinate-system transformations Thisformulation has been implemented in a documented spreadsheet and is availablefor download (httpeoljscnasagovsseopFootprintCalculatorxls) and is in theprocess of being implemented on a Web-based user interface to the AstronautPhotography Database (OYce of Earth Sciences 2000)
Although formulations 2 and 3 account for obliquity for the purposes of calcula-tion they treat the position of the spacecraft and position of the camera as one Inactuality astronauts are generally holding the cameras by hand (although camerasbracketed in the window are also used) and the selection of window position of theastronaut in the window and rotation of the camera relative to the movement ofthe spacecraft are not known Thus calculations using only the photo centre point(PC) and spacecraft nadir point (SN) give a locator ellipse and not the locations ofthe corners of the photograph A locator ellipse describes an estimated area on theground that is likely to be included in a speci c photograph regardless of the rotationof the lm plane about the camerarsquos optical axis ( gure 6)
Estimating the corner positions of the photo requires additional user input of asingle auxiliary pointmdasha location on the image that has a known location on theground Addition of this auxiliary point is an option available to users of thespreadsheet An example of the results of adding an auxiliary point is shown in gure 7 with comparisons of the various calculations in table 5
51 Formulation 1 Footprint for a nadir viewThe simplest way to estimate footprint size is to use the geometry of camera lens
and spacecraft altitude to calculate the scaling relationship between the image in the
Figure 6 Sketch representation of the locator ellipse an estimated area on the ground thatis likely to be included in a speci c photograph regardless of the rotation of the lmplane about the camerarsquos optical axis
Astronaut photography as digital data for remote sensing 4421
Figure 7 An example of the application of the Low Oblique Space Photo FootprintCalculator The photograph (STS093-704-50) of Limmen Bight Australia was takenfrom an altitude of 283 km using the Hasselblad camera with a 250 mm lens Mapused is 11 000 000 Operational Navigational Chart The distances shown equate toan average pixel size of 124 m for this photograph
lm and the area covered on the ground For a perfect nadir view the scalerelationship from the geometry of similar triangles is
d
D=
f
H(3)
where d=original image size D=distance of footprint on the ground f =focallength of lens and H=altitude Once D is known pixel size ( length=width) can becalculated from equation (2) These calculations represent the minimum footprintand minimum pixel size possible for a given camera system altitude and digitisingspatial resolution (table 2) Formulation 1 was used for calculating minimum pixelsizes shown in tables 2 and 4
By assuming digitizing at 2400 ppi (106 mm pixel Otilde 1 ) currently a spatial resolutioncommonly attainable from multipurpose colour transparency scanners (see sect441)this formulation was used to convert area covered to an IFOV equivalent for missionsof diVerent altitudes (table 2) Table 4 provides a comparison of IFOV of imagesfrom various satellites including the equivalent for astronaut photography Valuesin this table were derived by using formulation 1 to estimate the area covered becausea perfect nadir view represents the best possible spatial resolution and smallest eldof view that could be obtained
52 Formulation 2 Footprint for oblique views using simpli ed geometry and thegreat circle distance
A more realistic approach to determining the footprint of the photographaccounts for the fact that the camera is not usually pointing perfectly down at the
J A Robinson et al4422
Table 4 Comparison of pixel sizes (instantaneous elds of view) for satellite remote sensorswith pixel sizes for near vertical viewing photography from spacecraft Note that alldata returned from the automated satellites have the same eld of view while indicatedpixel sizes for astronaut photography are best-case for near vertical look angles See gure 5 for distributions of astronaut photographs of various look angles High-altitude aerial photography is also included for reference
Altitude PixelInstrument (km) width (mtimesm) Bands1
Landsat Multipectral Scanner2 (MSS 1974-) 880ndash940 79times56 4ndash5Landsat Thematic Mapper (TM 1982-) ~705 30times30 7Satellite pour lrsquoObservation de la Terre (SPOT 1986-) ~832
SPOT 1-3 Panchromatic 10times10 1SPOT 1-3 Multispectral (XS) 20times20 3
Astronaut Photography (1961-)Skylab3 (1973-1974 ) ~435
Multispectral Camera (6 camera stations) 17ndash19times17ndash19 3ndash8Earth Terrain Camera (hi-res colour) 875times875 3
Space Shuttle5 (1981-) 222ndash611Hasselblad 70 mm (250mm lens colour) vertical view 9ndash26times9ndash26 3Hasselblad 70 mm (100mm lens colour) vertical view 23ndash65times23ndash65 3Nikon 35 mm (400mm lens colour lm or ESC) vertical 5ndash16times5ndash16 3
High Altitude Aerial Photograph (1100 000) ~12 2times2 3
1Three bands listed for aerial photography obtainable by high-resolution colour digitizingFor high-resolution colour lm at 65 lp mmOtilde 1 (K Teitelbaum Eastman Kodak Co personalcommunication) the maximum information contained would be 3302 ppi
2Data from Simonett (1983Table 1-2)3Calculated as recommended by Jensen (19831596) the dividend of estimated ground
resolution at low contrast given in NASA (1974179-180) and 244Six lters plus colour and colour infrared lm (NASA 19747) 5Extracted from table 2
nadir point The look angle (the angle oV nadir that the camera is pointing) can becalculated trigonometrically by assuming a spherical earth and calculating the dis-tance between the coordinates of SN and PC ( gure 1) using the Great Circledistance haversine solution (Sinnott 1984 Snyder 198730ndash32 Chamberlain 1996)The diVerence between the spacecraft centre and nadir point latitudes Dlat=lat 2 shy lat 1 and the diVerence between the spacecraft centre and nadir pointlongitudes Dlon=lon 2 shy lon 1 enter the following equations
a=sin2ADlat
2 B+cos(lat 1) cos(lat 2) sin2ADlon
2 B (4)
c=2 arcsin(min[1 atilde a]) (5)
and oVset=R c with R the radius of a spherical Earth=6370 kmThe look angle (t ) is then given by
t=arctanAoffset
H B (6)
Assuming that the camera was positioned so that the imaginary line between thecentre and nadir points (the principal line) runs vertically through the centre of thephotograph the distance between the geometric centre of the photograph (principalpoint PP) and the top of the photograph is d2 ( gure 8(a)) The scale at any point
Astronaut photography as digital data for remote sensing 4423
(a) (b)
Figure 8 The geometric relationship between look angle lens focal length and the centrepoint and nadir points of a photograph (see also Estes and Simonett 1975) Note thatthe Earthrsquos surface and footprint represented in gure 1 are not shown here only arepresentation of the image area of the photograph Variables shown in the gurelook angle t focal length f PP centre of the image on the lm SN spacecraft nadird original image size (a) The special case where the camera is aligned squarely withthe isometric parallel In this case tilt occurs along a plane parallel with the centre ofthe photograph When the conditions are met calculations using Formulation 3 willgive the ground coordinates of the corner points of the photograph (b) The generalrelationship between these parameters and key photogrammetric points on the photo-graph For this more typical case coordinates of an additional point in the photographmust be input in order to adjust for twist of the camera and to nd the corner pointcoordinates
in the photograph varies as a function of the distance y along the principal linebetween the isocentre and the point according to the relationship
d
D=
f shy ysint
H(7)
(Wong 1980 equation (214) H amp the ground elevation h) Using gure 8(a) at thetop of the photo
y= f tanAt
2B+d
2(8)
and at the bottom of the photo
y= f tanAt
2Bshyd
2(9)
Thus for given PC and SN coordinates and assuming a photo orientation as in gure 8(a) and not gure 8(b) we can estimate a minimum D (using equations (7)and (8)) and a maximum D (using equations (7) and (9)) and then average the twoto determine the pixel size (P ) via equation (2)
J A Robinson et al4424
53 Formulation 3 T he L ow Oblique Space Photo Footprint CalculatorThe Low Oblique Space Photo Footprint Calculator was developed to provide
a more accurate estimation of the geographic coordinates for the footprint of a lowoblique photo of the Earthrsquos surface taken from a human-occupied spacecraft inorbit The calculator performs a series of three-dimensional coordinate transforma-tions to compute the location and orientation of the centre of the photo exposureplane relative to an earth referenced coordinate system The nominal camera focallength is then used to create a vector from the photorsquos perspective point througheach of eight points around the parameter of the image as de ned by the formatsize The geographical coordinates for the photo footprint are then computed byintersecting these photo vectors with a spherical earth model Although more sophist-icated projection algorithms are available no signi cant increase in the accuracy ofthe results would be produced by these algorithms due to inherent uncertainties inthe available input data (ie the spacecraft altitude photo centre location etc)
The calculations were initially implemented within a Microsoft Excel workbookwhich allowed one to embed the mathematical and graphical documentation nextto the actual calculations Thus interested users can inspect the mathematical pro-cessing A set of error traps was also built into the calculations to detect erroneousresults A summary of any errors generated is reported to the user with the resultsof the calculations Interested individuals are invited to download the Excel work-book from httpeoljscnasagovsseopFootprintCalculatorxls The calculations arecurrently being encoded in a high-level programming language and should soon beavailable alongside other background data provided for each photograph at OYceof Earth Sciences (2000)
For the purposes of these calculations a low oblique photograph was de ned as
one with the centre within 10 degrees of latitude and longitude of the spacecraftnadir point For the typical range of spacecraft altitudes to date this restricted thecalculations to photographs in which Earthrsquos horizon does not appear (the generalde nition of a low oblique photograph eg Campbell 199671 and gure 4)
531 Input data and resultsUpon opening the Excel workbook the user is presented with a program intro-
duction providing instructions for using the calculator The second worksheet tablsquoHow-To-Usersquo provides detailed step-by-step instructions for preparing the baselinedata for the program The third tab lsquoInput-Output rsquo contains the user input eldsand displays the results of the calculations The additional worksheets contain theactual calculations and program documentation Although users are welcome toreview these sheets an experienced user need only access the lsquoInput-Outputrsquo
spreadsheetTo begin a calculation the user enters the following information which is available
for each photo in the NASA Astronaut Photography Database (1) SN geographicalcoordinates of spacecraft nadir position at the time of photo (2) H spacecraftaltitude (3) PC the geographical coordinates of the centre of the photo (4) f nominal focal length and (5) d image format size The automatic implementationof the workbook on the web will automatically enter these values and completecalculations
For more accurate results the user may optionally enter the geographic co-ordinates and orientation for an auxiliary point on the photo which resolves thecamerarsquos rotation uncertainty about the optical axis The auxiliary point data must
Astronaut photography as digital data for remote sensing 4425
be computed by the user following the instruction contained in the lsquoHow-To-Usersquotab of the spreadsheet
After entering the input data the geographic coordinates of the photo footprint(ie four photo corner points four points at the bisector of each edge and the centreof the photo) are immediately displayed below the input elds along with any errormessages generated by the user input or by the calculations ( gure 7) Although
results are computed and displayed they should not be used when error messagesare produced by the program The program also computes the tilt angle for each ofthe photo vectors relative to the spacecraft nadir vector To the right of the photofootprint coordinates is displayed the arc distance along the surface of the sphere
between adjacent computed points
532 Calculation assumptions
The mathematical calculations implemented in the Low Oblique Space PhotoFootprint Calculator use the following assumptions
1 The SN location is used as exact even though the true value may vary by upto plusmn01 degree from the location provided with the photo
2 The spacecraft altitude is used as exact Although the determination of the
nadir point at the instant of a known spacecraft vector is relatively precise(plusmn115times10 Otilde 4 degrees) the propagator interpolates between sets of approxi-mately 10ndash40 known vectors per day and the time code recorded on the lmcan drift Thus the true value for SN may vary by up to plusmn01 degree fromthe value provided with the photo
3 The perspective centre of the camera is assumed to be at the given altitudeover the speci ed spacecraft nadir location at the time of photo exposure
4 The PC location is used as exact even though the true value may vary by upto plusmn05deg latitude and plusmn05deg longitude from the location provided with the
photo5 A spherical earth model is used with a nominal radius of 6 372 16154 m (a
common rst-order approximation for a spherical earth used in geodeticcomputations)
6 The nominal lens focal length of the camera lens is used in the computations(calibrated focal length values are not available)
7 The photo projection is based on the classic pin-hole camera model8 No correction for lens distortion or atmospheric re ection is made
9 If no auxiliary point data is provided the lsquoTop of the Imagersquo is orientedperpendicular to the vector from SN towards PC
533 T ransformation from Earth to photo coordinate systemsThe calculations begin by converting the geographic coordinates ( latitude and
longitude) of the SN and PC to a Rectangular Earth-Centred Coordinate System(R-Earth) de ned as shown in gure 9 (with the centre of the Earth at (0 0 0))
Using the vector from the Earthrsquos centre through SN and the spacecraft altitude thespacecraft location (SC) is also computed in R-Earth
For ease of computation a Rectangular Spacecraft-Centred Coordinate System(R-Spacecraft ) is de ned as shown in gure 9 The origin of R-Spacecraft is located
at SC with its +Z-axis aligned with vector from the centre of the Earth throughSN and its +X-axis aligned with the vector from SN to PC ( gure 9) The speci c
J A Robinson et al4426
Figure 9 Representation of the area included in a photograph as a geometric projectiononto the Earthrsquos surface Note that the focal length (the distance from the SpaceShuttle to the photo) is grossly overscale compared to the altitude (the distance fromthe Shuttle to the surface of the Earth
rotations and translation used to convert from the R-Earth to the R-Spacecraft arecomputed and documented in the spreadsheet
With the mathematical positions of the SC SN PC and the centre of the Earthcomputed in R-Spacecraft the program next computes the location of the camerarsquosprincipal point (PP) The principle point is the point of intersection of the opticalaxis of the lens with the image plane (ie the lm) It is nominally positioned at a
Astronaut photography as digital data for remote sensing 4427
distance equal to the focal length from the perspective centre of the camera (whichis assumed to be at SC) along the vector from PC through SC as shown in gure 9
A third coordinate system the Rectangular Photo Coordinate System (R-Photo)is created with its origin at PP its XndashY axial plane normal to the vector from PCthrough SC and its +X axis aligned with the +X axis of R-Spacecraft as shownin gure 9 The XndashY plane of this coordinate system represents the image plane ofthe photograph
534 Auxiliary point calculationsThe calculations above employ a critical assumption that all photos are taken
with the lsquotop of the imagersquo oriented perpendicular to the vector from the SN towardsPC as shown in gure 9 To avoid non-uniform solar heating of the external surfacemost orbiting spacecraft are slowly and continually rotated about one or more axesIn this condition a ight crew member taking a photo while oating in microgravitycould orient the photo with practically any orientation relative to the horizon (seealso gure 8(b)) Unfortunately since these photos are taken with conventionalhandheld cameras there is no other information available which can be used toresolve the photorsquos rotational ambiguity about the optical axis other then the photoitself This is why the above assumption is used and the footprint computed by thiscalculator is actually a lsquolocator ellipsersquo which estimates the area on the ground whatis likely to be included in a speci c photograph (see gure 6) This locator ellipse ismost accurate for square image formats and subject to additional distortion as thephotograph format becomes more rectangular
If the user wants a more precise calculation of the image footprint the photorsquosrotational ambiguity about the optical axis must be resolved This can be done inthe calculator by adding data for an auxiliary point Detailed instructions regardinghow to prepare and use auxiliary point data in the computations are included in thelsquoHow-To-Usersquo tab of the spreadsheet Basically the user determines which side ofthe photograph is top and then measures the angle between the line from PP to thetop of the photo and from PP to the auxiliary point on the photo ( gure 9)
If the user includes data for an auxiliary point a series of computations arecompleted to resolve the photo rotation ambiguity about the optical axis (ie the
+Z axis in R-Photo) A vector from the Auxiliary Point on the Earth (AE) throughthe photograph perspective centre ( located at SC) is intersected with the photo imageplane (XndashY plane of R-Photo) to compute the coordinates of the Auxiliary Point onthe photo (AP) in R-Photo A two-dimensional angle in the XndashY plane of R-Photofrom the shy X axis to a line from PP to AP is calculated as shown in gure 9 The
shy X axis is used as the origin of the angle since it represents the top of the photoonce it passes through the perspective centre The diVerence between the computedangle and the angle measured by the user on the photo resolves the ambiguity inthe rotation of the photo relative to the principal line ( gures 7 and 9) Thetransformations from R-Spacecraft and R-Photo are then modi ed to include anadditional rotation angle about the +Z axis in R-Photo
535 lsquoFootprint rsquo calculationsThe program next computes the coordinates of eight points about the perimeter
of the image format (ie located at the four photo corners plus a bisector pointalong each edge of the image) These points are identi ed in R-Photo based uponthe photograph format size and then converted to R-Spacecraft Since all computa-tions are done in orthogonal coordinate systems the R-Spacecraft to R-Photo
J A Robinson et al4428
rotation matrix is transposed to produce an R-Photo to R-Spacecraft rotation matrixOnce in R-Spacecraft a unit vector from each of the eight perimeter points throughthe photo perspective centre (the same point as SC) is computed This provides thecoordinates for points about the perimeter of the image format with their directionvectors in a common coordinate system with other key points needed to computethe photo footprint
The next step is to compute the point of intersection between the spherical earthmodel and each of the eight perimeter point vectors The scalar value for eachperimeter point unit vector is computed using two-dimensional planar trigonometryAn angle c is computed using the formula for the cosine of an angle between twoincluded vectors (the perimeter point unit vector and the vector from SC to thecentre of the Earth) Angle y is computed using the Law of Sines Angle e=180degrees shy c shy y The scalar value of the perimeter point vector is computed using eand the Law of Cosines The scalar value is then multiplied by the perimeter pointunit vector to produce the three-dimensional point of intersection of the vector withEarthrsquos surface in R-Spacecraft The process is repeated independently for each ofthe eight perimeter point vectors Aside from its mathematical simplicity the valueof arcsine (computed in step 2) will exceed the normal range for the sin of an anglewhen the perimeter point vectors fail to intersect with the surface of the earth Asimple test based on this principle allows the program to correctly handle obliquephotos which image a portion of the horizon (see results for high oblique photographsin table 5)
The nal step in the process converts the eight earth intersection points from theR-Spacecraft to R-Earth The results are then converted to the standard geographiccoordinate system and displayed on the lsquoInput-Outputrsquo page of the spreadsheet
54 ExamplesAll three formulations were applied to the photographs included in this paper
and the results are compared in table 5 Formulation 1 gives too small a value fordistance across the photograph (D) for all but the most nadir shots and thus servesas an indicator of the best theoretical case but is not a good measure for a speci cphotograph For example the photograph of Lake Eyre taken from 276 km altitude( gure 2(a)) and the photograph of Limmen Bight ( gure 7) were closest to beingnadir views (oVsetslt68 km or tlt15deg table 5) For these photographs D calculatedusing Formulation 1 was similar to the minimum D calculated using Formulations2 and 3 For almost all other more oblique photographs Formulation 1 gave asigni cant underestimate of the distance covered in the photograph For gure 5(A)(the picture of Houston taken with a 40 mm lens) Formulation 1 did not give anunderestimate for D This is because Formulation 1 does not account for curvatureof the Earth in any way With this large eld of view assuming a at Earth in atedthe value of D above the minimum from calculations that included Earth curvature
A major diVerence between Formulations 2 and 3 is the ability to estimate pixelsizes (P) in both directions (along the principal line and perpendicular to the principalline) For the more oblique photographs the vertical estimate of D and pixel sizes ismuch larger than in the horizontal direction (eg the low oblique and high obliquephotographs of Hawaii gure 4 table 5)
For the area of Limmen Bight ( gure 7) table 5 illustrates the improvement inthe estimate of distance and pixel size that can be obtained by re-estimating thelocation of the PC with greater accuracy Centre points in the catalogued data are
Astronaut photography as digital data for remote sensing 4429
plusmn05deg of latitude and longitude When the centre point was re-estimated to plusmn002degwe determined that the photograph was not taken as obliquely as rst thought(change in the estimate of the look angle t from 169 to 128deg table 5) When theauxiliary point was added to the calculations of Formula 3 the calculated look angleshrank further to 110deg indicating that this photograph was taken at very close toa nadir view Of course this improvement in accuracy could have also led to estimatesof greater obliquity and corresponding larger pixel sizes
We also tested the performance of the scale calculator with auxiliary point byestimating the corner and point locations on the photograph using a 11 000 000Operational Navigational Chart For this test we estimated our ability to readcoordinates from the map as plusmn002degand our error in nding the locations of thecorner points as plusmn015deg (this error varies among photographs depending on thedetail that can be matched between photo and map) For Limmen Bight ( gure 7)the mean diVerence between map estimates and calculator estimates for four pointswas 031deg (SD=018 n=8) For a photograph of San Francisco Bay (STS062-151-291) the mean diVerence between map estimates and calculator estimates forfour points was 0064deg (SD=018 n=8) For a photograph of San Francisco Bay(STS062-151-291) the mean diVerence between map estimates and calculator estim-ates for eight points was 0196deg (SD=0146 n=16) Thus in one case the calculatorestimates were better than our estimate of the error in locating corner points on themap It is reasonable to expect that for nadir-viewing photographs the calculatorused with an auxiliary point can estimate locations of the edges of a photograph towithin plusmn03deg
55 Empirical con rmation of spatial resolution estimatesAs stated previously a challenge to estimating system-AWAR for astronaut
photography of Earth is the lack of suitable targets Small features in an image cansometimes be used as a check on the size of objects that can be successfully resolvedgiving an approximate value for GRD Similarly the number of pixels that make upthose features in the digitized image can be used to make an independent calculationof pixel size We have successfully used features such as roads and airport runwaysto make estimates of spatial scale and resolution (eg Robinson et al 2000c) Whilerecognizing that the use of linear detail in an image is a poor approximation to abar target and that linear objects smaller than the resolving power can often bedetected (Charman 1965) few objects other than roads could be found to make anydirect estimates of GRD Thus roads and runways were used in the images ofHouston (where one can readily conduct ground veri cations and where a numberof higher-contrast concrete roadways were available) to obtain empirical estimatesof GRD and pixel size for comparison with table 5
In the all-digital ESC image of Houston ( gure 3(d )) we examined EllingtonField runway 4-22 (centre left of the image) which is 24384 mtimes457 m This runwayis approximately 6ndash7 pixels in width and 304ndash3094 pixels in length so pixelsrepresent an distance on the ground 7ndash8 m Using a lower contrast measure of astreet length between two intersections (2125 m=211 pixels) a pixel width of 101 mis estimated These results compare favourably with the minimum estimate of 81 mpixels using Formulation 1 (table 5) For an estimate of GRD that would be morecomparable to aerial photography of a line target the smallest street without treecover that could be clearly distinguished on the photograph was 792 m wide Thesmallest non-street object (a gap between stages of the Saturn rocket on display in
J A Robinson et al4430
Tab
le5
App
lica
tio
nof
met
hod
sfo
res
tim
ati
ng
spati
alre
solu
tio
no
fast
ron
aut
ph
oto
gra
ph
sin
clu
din
gF
orm
ula
tion
s12
and
3Im
pro
ved
cen
tre
po
int
esti
mat
esan
dauxilia
ryp
oin
tca
lcu
lati
ons
are
sho
wn
for
gure
7V
aria
ble
sas
use
din
the
text
fle
ns
foca
lle
ngt
hH
al
titu
de
PC
ph
oto
gra
ph
cen
tre
poin
tSN
sp
ace
craft
nadir
po
int
D
dis
tance
cover
edo
nth
egr
oun
d
Pd
ista
nce
on
the
grou
nd
repre
sen
ted
by
apix
el
OVse
td
ista
nce
bet
wee
nP
Cand
SNusi
ng
the
grea
tci
rcle
form
ula
t
look
angle
away
from
SN
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
2L
ake
Eyre
ST
S093
-717
-80
250
276
shy29
0137
0shy
28
413
71
607
11
767
413
7609
ndash64
212
013
760
9ndash64
7121
124
ST
S082
-754
-61
250
617
shy28
5137
0shy
28
113
50
135
726
1201
18
013
8ndash14
827
517
9138ndash15
3277
294
Fig
ure
3H
ou
sto
nS
TS
062
-97-
151
4029
829
5shy
95
530
5shy
95
4410
78
8112
20
5348ndash58
990
1205
351
ndash63
8951
10
36
ST
S094
-737
-28
100
294
295
shy
95
528
3shy
95
5162
31
1133
24
415
8ndash20
334
724
3158ndash20
7351
390
ST
S055
-71-
41250
302
295
shy
95
028
2shy
93
766
412
8192
32
5736
ndash84
715
232
273
9ndash85
7154
185
S73
E51
45300
2685
295
shy
95
024
78
0F
igu
re4
Haw
aii
ST
S069
-701
-65
250
370
195
shy
155
520
4shy
153
881
415
7204
28
9876
ndash98
917
928
688
0ndash10
9181
210
ST
S069
-701
-70
250
370
210
shy
157
519
2shy
150
781
415
7738
63
414
9ndash23
336
760
7148ndash55
6394
10
63
ST
S069
-729
-27
100
398
200
shy
156
620
5shy
148
2219
42
1867
65
432
8ndash13
10
158
62
4323+
311
N
A
Astronaut photography as digital data for remote sensing 4431
Tab
le5
(Cont
inued
)
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
7L
imm
enB
igh
tS
TS
093
-704
-50
250
283
shy15
0135
0shy
15
013
58
623
120
859
16
963
0ndash67
3125
16
863
0ndash67
612
613
2P
CR
e-es
tim
ated
shy14
733
1352
67
64
5128
623
ndash65
5123
128
623
ndash65
712
3127
Auxilia
ryP
oin
t611
062
3ndash65
112
312
5F
igu
re10
N
ear
Au
stin
T
XS
TS
095
-716
-46
250
543
30
5shy
975
28
6shy
1007
120
230
375
34
613
5ndash157
281
34
1136ndash16
128
636
0
1 All
pho
togr
aphs
exce
pt
on
eta
ken
wit
hH
ass
elbla
dca
mer
aso
dis
tan
ceacr
oss
the
image
d=
55m
m
for
S73
E51
45
taken
wit
hel
ectr
onic
still
cam
erad=
276
5m
mho
rizo
nta
l2 P
Cand
SN
are
giv
enin
the
form
(lati
tud
elo
ngit
ude)
deg
rees
w
ith
neg
ativ
en
um
ber
sre
pre
senti
ng
south
and
wes
tA
ccura
cyo
fP
Cis
plusmn0
5and
SN
isplusmn
01
as
reco
rded
inth
eA
stro
naut
Ph
oto
gra
phy
Data
base
ex
cep
tw
her
en
ote
dth
atce
ntr
ep
oin
tw
asre
-est
imate
dw
ith
grea
ter
accu
racy
3 C
alc
ula
ted
usi
ng
the
mea
nofth
em
inim
um
an
dm
axim
um
esti
mate
sof
DF
orm
ula
tion
2co
nsi
der
sD
inth
ed
irec
tion
per
pen
dic
ula
rto
the
Pri
nci
pal
Lin
eo
nly
4 C
alc
ula
ted
inho
rizo
nta
lan
dve
rtic
aldir
ecti
on
sb
ut
still
ass
um
ing
geom
etry
of
gure
6(B
)F
irst
nu
mb
eris
the
mea
no
fth
em
inim
um
Dat
the
bott
om
of
the
ph
oto
and
the
maxi
mum
Dat
the
top
of
the
ph
oto
(range
show
nin
the
tab
le)
and
isco
mpara
ble
toF
orm
ula
tio
n2
Sec
ond
num
ber
isth
em
ean
of
Des
tim
ated
alon
gth
eP
rin
cip
alline
and
alo
ng
the
ou
tsid
eed
ge
(range
not
show
n)
5 Alt
itu
de
isap
pro
xim
ate
for
mis
sion
as
exac
tti
me
pho
togr
ap
hw
asta
ken
and
corr
esp
on
din
gn
adir
data
are
no
tava
ilab
le
Lack
of
nad
irdata
pre
ven
tsap
plica
tion
of
Form
ula
tion
s2
an
d3
6 Au
xilliar
ypo
int
add
edw
as
shy14
80
lati
tud
e13
51
2lo
ngit
ude
shy46
5degro
tati
on
angle
in
stru
ctio
ns
for
esti
mati
ng
the
rota
tio
nan
gle
para
met
erare
conta
ined
as
par
tof
the
scal
eca
lcu
lato
rsp
read
shee
tdo
cum
enta
tio
n
J A Robinson et al4432
a park at Johnson Space Center) that could clearly be distinguished on the photo-graph was 853 m wide
For the photograph of Houston taken with a 250 mm lens ( gure 3(C)) anddigitized from second generation lm at 2400 ppi (106 mm pixel Otilde 1 ) Ellington Fieldrunway 4-22 is 3 pixels in width and 1613 pixels in length so pixels represent151ndash152 m on the ground These results compare favourably with the minimumestimate of 152 m using Formulation 2 and 154ndash185 m pixels using Formulation 3(table 5) For an estimate of GRD using an 8times8 inch print (1369 enlargement) and4times magni cation the smallest street that could clearly be distinguished was 822 mwide the same feature could barely be distinguished on the digitized image
We also made an empirical estimate of spatial resolution for lower contrastvegetation boundaries By clearing forest so that a pattern would be visible to landingaircraft a landowner outside Austin Texas (see also aerial photo in Lisheron 2000)created a target that is also useful for evaluating spatial resolution of astronautphotographs The forest was selectively cleared in order to spell the landownerrsquosname lsquoLUECKErsquo with the remaining trees ( gure 10) According to local surveyorswho planned the clearing the plan was to create letters that were 3100 fttimes1700 ft(9449 mtimes5182 m) Photographed at a high altitude relative to most Shuttle missions(543 km) with a 250 mm lens Formula 3 predicts that each pixel would represent anarea 286 mtimes360 m on the ground (table 5) When original lm was digitized at2400 ppi (106 mm pixel Otilde 1 ) letters correspond to 294times188 pixels for a comparablepixel size of 27ndash32 m
6 Summary and conclusionsAstronaut photographs can be an excellent source of data for remote sensing
applications Best-case resolutions are similar to that for Landsat or SPOT withpixels as small as lt10 m It was estimated that of images taken to date 504 hadlens and obliquity characteristics that make them potential candidates for remotesensing information Digitized astronaut photographs can be overlaid with othersatellite data using GIS (Eckardt et al 2000) or used to ll in gaps in time serieswhen other imagery is not available As a source of public-domain information theycan be very useful for scientists who do not have access to satellite imagery eitherbecause of the expense of image acquisition (to the end user) or the computersystems needed for processing satellite images These diVerences in image acquisitioncosts (to the user) are summarized in table 6
Searching of the complete database of NASA astronaut photography includinglow-spatial-resolution browse images is available via the Web (OYce of EarthSciences 2000) This provides nearly global access for identifying images that willcontribute to a speci c scienti c project For the most detailed studies digitalproducts posted to the web will not be of suYcient quality To date we have madeit a practice of digitizing small numbers of images when requested by scientists atno charge Such requests (including a description of the project involved) can bemade through the authors or using the contact information listed on the websiteInvestigators with needs for larger numbers of digital images have also been servedthrough collaborative agreements
In our experience scientists that nd the data most valuable are those in develop-ing countries those studying areas that have not usually been targets for the majorsatellites those having diYculty in nding low-cloud images those interested inconstructing time series or those interested in using a large number of images
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
Astronaut photography as digital data for remote sensing 4421
Figure 7 An example of the application of the Low Oblique Space Photo FootprintCalculator The photograph (STS093-704-50) of Limmen Bight Australia was takenfrom an altitude of 283 km using the Hasselblad camera with a 250 mm lens Mapused is 11 000 000 Operational Navigational Chart The distances shown equate toan average pixel size of 124 m for this photograph
lm and the area covered on the ground For a perfect nadir view the scalerelationship from the geometry of similar triangles is
d
D=
f
H(3)
where d=original image size D=distance of footprint on the ground f =focallength of lens and H=altitude Once D is known pixel size ( length=width) can becalculated from equation (2) These calculations represent the minimum footprintand minimum pixel size possible for a given camera system altitude and digitisingspatial resolution (table 2) Formulation 1 was used for calculating minimum pixelsizes shown in tables 2 and 4
By assuming digitizing at 2400 ppi (106 mm pixel Otilde 1 ) currently a spatial resolutioncommonly attainable from multipurpose colour transparency scanners (see sect441)this formulation was used to convert area covered to an IFOV equivalent for missionsof diVerent altitudes (table 2) Table 4 provides a comparison of IFOV of imagesfrom various satellites including the equivalent for astronaut photography Valuesin this table were derived by using formulation 1 to estimate the area covered becausea perfect nadir view represents the best possible spatial resolution and smallest eldof view that could be obtained
52 Formulation 2 Footprint for oblique views using simpli ed geometry and thegreat circle distance
A more realistic approach to determining the footprint of the photographaccounts for the fact that the camera is not usually pointing perfectly down at the
J A Robinson et al4422
Table 4 Comparison of pixel sizes (instantaneous elds of view) for satellite remote sensorswith pixel sizes for near vertical viewing photography from spacecraft Note that alldata returned from the automated satellites have the same eld of view while indicatedpixel sizes for astronaut photography are best-case for near vertical look angles See gure 5 for distributions of astronaut photographs of various look angles High-altitude aerial photography is also included for reference
Altitude PixelInstrument (km) width (mtimesm) Bands1
Landsat Multipectral Scanner2 (MSS 1974-) 880ndash940 79times56 4ndash5Landsat Thematic Mapper (TM 1982-) ~705 30times30 7Satellite pour lrsquoObservation de la Terre (SPOT 1986-) ~832
SPOT 1-3 Panchromatic 10times10 1SPOT 1-3 Multispectral (XS) 20times20 3
Astronaut Photography (1961-)Skylab3 (1973-1974 ) ~435
Multispectral Camera (6 camera stations) 17ndash19times17ndash19 3ndash8Earth Terrain Camera (hi-res colour) 875times875 3
Space Shuttle5 (1981-) 222ndash611Hasselblad 70 mm (250mm lens colour) vertical view 9ndash26times9ndash26 3Hasselblad 70 mm (100mm lens colour) vertical view 23ndash65times23ndash65 3Nikon 35 mm (400mm lens colour lm or ESC) vertical 5ndash16times5ndash16 3
High Altitude Aerial Photograph (1100 000) ~12 2times2 3
1Three bands listed for aerial photography obtainable by high-resolution colour digitizingFor high-resolution colour lm at 65 lp mmOtilde 1 (K Teitelbaum Eastman Kodak Co personalcommunication) the maximum information contained would be 3302 ppi
2Data from Simonett (1983Table 1-2)3Calculated as recommended by Jensen (19831596) the dividend of estimated ground
resolution at low contrast given in NASA (1974179-180) and 244Six lters plus colour and colour infrared lm (NASA 19747) 5Extracted from table 2
nadir point The look angle (the angle oV nadir that the camera is pointing) can becalculated trigonometrically by assuming a spherical earth and calculating the dis-tance between the coordinates of SN and PC ( gure 1) using the Great Circledistance haversine solution (Sinnott 1984 Snyder 198730ndash32 Chamberlain 1996)The diVerence between the spacecraft centre and nadir point latitudes Dlat=lat 2 shy lat 1 and the diVerence between the spacecraft centre and nadir pointlongitudes Dlon=lon 2 shy lon 1 enter the following equations
a=sin2ADlat
2 B+cos(lat 1) cos(lat 2) sin2ADlon
2 B (4)
c=2 arcsin(min[1 atilde a]) (5)
and oVset=R c with R the radius of a spherical Earth=6370 kmThe look angle (t ) is then given by
t=arctanAoffset
H B (6)
Assuming that the camera was positioned so that the imaginary line between thecentre and nadir points (the principal line) runs vertically through the centre of thephotograph the distance between the geometric centre of the photograph (principalpoint PP) and the top of the photograph is d2 ( gure 8(a)) The scale at any point
Astronaut photography as digital data for remote sensing 4423
(a) (b)
Figure 8 The geometric relationship between look angle lens focal length and the centrepoint and nadir points of a photograph (see also Estes and Simonett 1975) Note thatthe Earthrsquos surface and footprint represented in gure 1 are not shown here only arepresentation of the image area of the photograph Variables shown in the gurelook angle t focal length f PP centre of the image on the lm SN spacecraft nadird original image size (a) The special case where the camera is aligned squarely withthe isometric parallel In this case tilt occurs along a plane parallel with the centre ofthe photograph When the conditions are met calculations using Formulation 3 willgive the ground coordinates of the corner points of the photograph (b) The generalrelationship between these parameters and key photogrammetric points on the photo-graph For this more typical case coordinates of an additional point in the photographmust be input in order to adjust for twist of the camera and to nd the corner pointcoordinates
in the photograph varies as a function of the distance y along the principal linebetween the isocentre and the point according to the relationship
d
D=
f shy ysint
H(7)
(Wong 1980 equation (214) H amp the ground elevation h) Using gure 8(a) at thetop of the photo
y= f tanAt
2B+d
2(8)
and at the bottom of the photo
y= f tanAt
2Bshyd
2(9)
Thus for given PC and SN coordinates and assuming a photo orientation as in gure 8(a) and not gure 8(b) we can estimate a minimum D (using equations (7)and (8)) and a maximum D (using equations (7) and (9)) and then average the twoto determine the pixel size (P ) via equation (2)
J A Robinson et al4424
53 Formulation 3 T he L ow Oblique Space Photo Footprint CalculatorThe Low Oblique Space Photo Footprint Calculator was developed to provide
a more accurate estimation of the geographic coordinates for the footprint of a lowoblique photo of the Earthrsquos surface taken from a human-occupied spacecraft inorbit The calculator performs a series of three-dimensional coordinate transforma-tions to compute the location and orientation of the centre of the photo exposureplane relative to an earth referenced coordinate system The nominal camera focallength is then used to create a vector from the photorsquos perspective point througheach of eight points around the parameter of the image as de ned by the formatsize The geographical coordinates for the photo footprint are then computed byintersecting these photo vectors with a spherical earth model Although more sophist-icated projection algorithms are available no signi cant increase in the accuracy ofthe results would be produced by these algorithms due to inherent uncertainties inthe available input data (ie the spacecraft altitude photo centre location etc)
The calculations were initially implemented within a Microsoft Excel workbookwhich allowed one to embed the mathematical and graphical documentation nextto the actual calculations Thus interested users can inspect the mathematical pro-cessing A set of error traps was also built into the calculations to detect erroneousresults A summary of any errors generated is reported to the user with the resultsof the calculations Interested individuals are invited to download the Excel work-book from httpeoljscnasagovsseopFootprintCalculatorxls The calculations arecurrently being encoded in a high-level programming language and should soon beavailable alongside other background data provided for each photograph at OYceof Earth Sciences (2000)
For the purposes of these calculations a low oblique photograph was de ned as
one with the centre within 10 degrees of latitude and longitude of the spacecraftnadir point For the typical range of spacecraft altitudes to date this restricted thecalculations to photographs in which Earthrsquos horizon does not appear (the generalde nition of a low oblique photograph eg Campbell 199671 and gure 4)
531 Input data and resultsUpon opening the Excel workbook the user is presented with a program intro-
duction providing instructions for using the calculator The second worksheet tablsquoHow-To-Usersquo provides detailed step-by-step instructions for preparing the baselinedata for the program The third tab lsquoInput-Output rsquo contains the user input eldsand displays the results of the calculations The additional worksheets contain theactual calculations and program documentation Although users are welcome toreview these sheets an experienced user need only access the lsquoInput-Outputrsquo
spreadsheetTo begin a calculation the user enters the following information which is available
for each photo in the NASA Astronaut Photography Database (1) SN geographicalcoordinates of spacecraft nadir position at the time of photo (2) H spacecraftaltitude (3) PC the geographical coordinates of the centre of the photo (4) f nominal focal length and (5) d image format size The automatic implementationof the workbook on the web will automatically enter these values and completecalculations
For more accurate results the user may optionally enter the geographic co-ordinates and orientation for an auxiliary point on the photo which resolves thecamerarsquos rotation uncertainty about the optical axis The auxiliary point data must
Astronaut photography as digital data for remote sensing 4425
be computed by the user following the instruction contained in the lsquoHow-To-Usersquotab of the spreadsheet
After entering the input data the geographic coordinates of the photo footprint(ie four photo corner points four points at the bisector of each edge and the centreof the photo) are immediately displayed below the input elds along with any errormessages generated by the user input or by the calculations ( gure 7) Although
results are computed and displayed they should not be used when error messagesare produced by the program The program also computes the tilt angle for each ofthe photo vectors relative to the spacecraft nadir vector To the right of the photofootprint coordinates is displayed the arc distance along the surface of the sphere
between adjacent computed points
532 Calculation assumptions
The mathematical calculations implemented in the Low Oblique Space PhotoFootprint Calculator use the following assumptions
1 The SN location is used as exact even though the true value may vary by upto plusmn01 degree from the location provided with the photo
2 The spacecraft altitude is used as exact Although the determination of the
nadir point at the instant of a known spacecraft vector is relatively precise(plusmn115times10 Otilde 4 degrees) the propagator interpolates between sets of approxi-mately 10ndash40 known vectors per day and the time code recorded on the lmcan drift Thus the true value for SN may vary by up to plusmn01 degree fromthe value provided with the photo
3 The perspective centre of the camera is assumed to be at the given altitudeover the speci ed spacecraft nadir location at the time of photo exposure
4 The PC location is used as exact even though the true value may vary by upto plusmn05deg latitude and plusmn05deg longitude from the location provided with the
photo5 A spherical earth model is used with a nominal radius of 6 372 16154 m (a
common rst-order approximation for a spherical earth used in geodeticcomputations)
6 The nominal lens focal length of the camera lens is used in the computations(calibrated focal length values are not available)
7 The photo projection is based on the classic pin-hole camera model8 No correction for lens distortion or atmospheric re ection is made
9 If no auxiliary point data is provided the lsquoTop of the Imagersquo is orientedperpendicular to the vector from SN towards PC
533 T ransformation from Earth to photo coordinate systemsThe calculations begin by converting the geographic coordinates ( latitude and
longitude) of the SN and PC to a Rectangular Earth-Centred Coordinate System(R-Earth) de ned as shown in gure 9 (with the centre of the Earth at (0 0 0))
Using the vector from the Earthrsquos centre through SN and the spacecraft altitude thespacecraft location (SC) is also computed in R-Earth
For ease of computation a Rectangular Spacecraft-Centred Coordinate System(R-Spacecraft ) is de ned as shown in gure 9 The origin of R-Spacecraft is located
at SC with its +Z-axis aligned with vector from the centre of the Earth throughSN and its +X-axis aligned with the vector from SN to PC ( gure 9) The speci c
J A Robinson et al4426
Figure 9 Representation of the area included in a photograph as a geometric projectiononto the Earthrsquos surface Note that the focal length (the distance from the SpaceShuttle to the photo) is grossly overscale compared to the altitude (the distance fromthe Shuttle to the surface of the Earth
rotations and translation used to convert from the R-Earth to the R-Spacecraft arecomputed and documented in the spreadsheet
With the mathematical positions of the SC SN PC and the centre of the Earthcomputed in R-Spacecraft the program next computes the location of the camerarsquosprincipal point (PP) The principle point is the point of intersection of the opticalaxis of the lens with the image plane (ie the lm) It is nominally positioned at a
Astronaut photography as digital data for remote sensing 4427
distance equal to the focal length from the perspective centre of the camera (whichis assumed to be at SC) along the vector from PC through SC as shown in gure 9
A third coordinate system the Rectangular Photo Coordinate System (R-Photo)is created with its origin at PP its XndashY axial plane normal to the vector from PCthrough SC and its +X axis aligned with the +X axis of R-Spacecraft as shownin gure 9 The XndashY plane of this coordinate system represents the image plane ofthe photograph
534 Auxiliary point calculationsThe calculations above employ a critical assumption that all photos are taken
with the lsquotop of the imagersquo oriented perpendicular to the vector from the SN towardsPC as shown in gure 9 To avoid non-uniform solar heating of the external surfacemost orbiting spacecraft are slowly and continually rotated about one or more axesIn this condition a ight crew member taking a photo while oating in microgravitycould orient the photo with practically any orientation relative to the horizon (seealso gure 8(b)) Unfortunately since these photos are taken with conventionalhandheld cameras there is no other information available which can be used toresolve the photorsquos rotational ambiguity about the optical axis other then the photoitself This is why the above assumption is used and the footprint computed by thiscalculator is actually a lsquolocator ellipsersquo which estimates the area on the ground whatis likely to be included in a speci c photograph (see gure 6) This locator ellipse ismost accurate for square image formats and subject to additional distortion as thephotograph format becomes more rectangular
If the user wants a more precise calculation of the image footprint the photorsquosrotational ambiguity about the optical axis must be resolved This can be done inthe calculator by adding data for an auxiliary point Detailed instructions regardinghow to prepare and use auxiliary point data in the computations are included in thelsquoHow-To-Usersquo tab of the spreadsheet Basically the user determines which side ofthe photograph is top and then measures the angle between the line from PP to thetop of the photo and from PP to the auxiliary point on the photo ( gure 9)
If the user includes data for an auxiliary point a series of computations arecompleted to resolve the photo rotation ambiguity about the optical axis (ie the
+Z axis in R-Photo) A vector from the Auxiliary Point on the Earth (AE) throughthe photograph perspective centre ( located at SC) is intersected with the photo imageplane (XndashY plane of R-Photo) to compute the coordinates of the Auxiliary Point onthe photo (AP) in R-Photo A two-dimensional angle in the XndashY plane of R-Photofrom the shy X axis to a line from PP to AP is calculated as shown in gure 9 The
shy X axis is used as the origin of the angle since it represents the top of the photoonce it passes through the perspective centre The diVerence between the computedangle and the angle measured by the user on the photo resolves the ambiguity inthe rotation of the photo relative to the principal line ( gures 7 and 9) Thetransformations from R-Spacecraft and R-Photo are then modi ed to include anadditional rotation angle about the +Z axis in R-Photo
535 lsquoFootprint rsquo calculationsThe program next computes the coordinates of eight points about the perimeter
of the image format (ie located at the four photo corners plus a bisector pointalong each edge of the image) These points are identi ed in R-Photo based uponthe photograph format size and then converted to R-Spacecraft Since all computa-tions are done in orthogonal coordinate systems the R-Spacecraft to R-Photo
J A Robinson et al4428
rotation matrix is transposed to produce an R-Photo to R-Spacecraft rotation matrixOnce in R-Spacecraft a unit vector from each of the eight perimeter points throughthe photo perspective centre (the same point as SC) is computed This provides thecoordinates for points about the perimeter of the image format with their directionvectors in a common coordinate system with other key points needed to computethe photo footprint
The next step is to compute the point of intersection between the spherical earthmodel and each of the eight perimeter point vectors The scalar value for eachperimeter point unit vector is computed using two-dimensional planar trigonometryAn angle c is computed using the formula for the cosine of an angle between twoincluded vectors (the perimeter point unit vector and the vector from SC to thecentre of the Earth) Angle y is computed using the Law of Sines Angle e=180degrees shy c shy y The scalar value of the perimeter point vector is computed using eand the Law of Cosines The scalar value is then multiplied by the perimeter pointunit vector to produce the three-dimensional point of intersection of the vector withEarthrsquos surface in R-Spacecraft The process is repeated independently for each ofthe eight perimeter point vectors Aside from its mathematical simplicity the valueof arcsine (computed in step 2) will exceed the normal range for the sin of an anglewhen the perimeter point vectors fail to intersect with the surface of the earth Asimple test based on this principle allows the program to correctly handle obliquephotos which image a portion of the horizon (see results for high oblique photographsin table 5)
The nal step in the process converts the eight earth intersection points from theR-Spacecraft to R-Earth The results are then converted to the standard geographiccoordinate system and displayed on the lsquoInput-Outputrsquo page of the spreadsheet
54 ExamplesAll three formulations were applied to the photographs included in this paper
and the results are compared in table 5 Formulation 1 gives too small a value fordistance across the photograph (D) for all but the most nadir shots and thus servesas an indicator of the best theoretical case but is not a good measure for a speci cphotograph For example the photograph of Lake Eyre taken from 276 km altitude( gure 2(a)) and the photograph of Limmen Bight ( gure 7) were closest to beingnadir views (oVsetslt68 km or tlt15deg table 5) For these photographs D calculatedusing Formulation 1 was similar to the minimum D calculated using Formulations2 and 3 For almost all other more oblique photographs Formulation 1 gave asigni cant underestimate of the distance covered in the photograph For gure 5(A)(the picture of Houston taken with a 40 mm lens) Formulation 1 did not give anunderestimate for D This is because Formulation 1 does not account for curvatureof the Earth in any way With this large eld of view assuming a at Earth in atedthe value of D above the minimum from calculations that included Earth curvature
A major diVerence between Formulations 2 and 3 is the ability to estimate pixelsizes (P) in both directions (along the principal line and perpendicular to the principalline) For the more oblique photographs the vertical estimate of D and pixel sizes ismuch larger than in the horizontal direction (eg the low oblique and high obliquephotographs of Hawaii gure 4 table 5)
For the area of Limmen Bight ( gure 7) table 5 illustrates the improvement inthe estimate of distance and pixel size that can be obtained by re-estimating thelocation of the PC with greater accuracy Centre points in the catalogued data are
Astronaut photography as digital data for remote sensing 4429
plusmn05deg of latitude and longitude When the centre point was re-estimated to plusmn002degwe determined that the photograph was not taken as obliquely as rst thought(change in the estimate of the look angle t from 169 to 128deg table 5) When theauxiliary point was added to the calculations of Formula 3 the calculated look angleshrank further to 110deg indicating that this photograph was taken at very close toa nadir view Of course this improvement in accuracy could have also led to estimatesof greater obliquity and corresponding larger pixel sizes
We also tested the performance of the scale calculator with auxiliary point byestimating the corner and point locations on the photograph using a 11 000 000Operational Navigational Chart For this test we estimated our ability to readcoordinates from the map as plusmn002degand our error in nding the locations of thecorner points as plusmn015deg (this error varies among photographs depending on thedetail that can be matched between photo and map) For Limmen Bight ( gure 7)the mean diVerence between map estimates and calculator estimates for four pointswas 031deg (SD=018 n=8) For a photograph of San Francisco Bay (STS062-151-291) the mean diVerence between map estimates and calculator estimates forfour points was 0064deg (SD=018 n=8) For a photograph of San Francisco Bay(STS062-151-291) the mean diVerence between map estimates and calculator estim-ates for eight points was 0196deg (SD=0146 n=16) Thus in one case the calculatorestimates were better than our estimate of the error in locating corner points on themap It is reasonable to expect that for nadir-viewing photographs the calculatorused with an auxiliary point can estimate locations of the edges of a photograph towithin plusmn03deg
55 Empirical con rmation of spatial resolution estimatesAs stated previously a challenge to estimating system-AWAR for astronaut
photography of Earth is the lack of suitable targets Small features in an image cansometimes be used as a check on the size of objects that can be successfully resolvedgiving an approximate value for GRD Similarly the number of pixels that make upthose features in the digitized image can be used to make an independent calculationof pixel size We have successfully used features such as roads and airport runwaysto make estimates of spatial scale and resolution (eg Robinson et al 2000c) Whilerecognizing that the use of linear detail in an image is a poor approximation to abar target and that linear objects smaller than the resolving power can often bedetected (Charman 1965) few objects other than roads could be found to make anydirect estimates of GRD Thus roads and runways were used in the images ofHouston (where one can readily conduct ground veri cations and where a numberof higher-contrast concrete roadways were available) to obtain empirical estimatesof GRD and pixel size for comparison with table 5
In the all-digital ESC image of Houston ( gure 3(d )) we examined EllingtonField runway 4-22 (centre left of the image) which is 24384 mtimes457 m This runwayis approximately 6ndash7 pixels in width and 304ndash3094 pixels in length so pixelsrepresent an distance on the ground 7ndash8 m Using a lower contrast measure of astreet length between two intersections (2125 m=211 pixels) a pixel width of 101 mis estimated These results compare favourably with the minimum estimate of 81 mpixels using Formulation 1 (table 5) For an estimate of GRD that would be morecomparable to aerial photography of a line target the smallest street without treecover that could be clearly distinguished on the photograph was 792 m wide Thesmallest non-street object (a gap between stages of the Saturn rocket on display in
J A Robinson et al4430
Tab
le5
App
lica
tio
nof
met
hod
sfo
res
tim
ati
ng
spati
alre
solu
tio
no
fast
ron
aut
ph
oto
gra
ph
sin
clu
din
gF
orm
ula
tion
s12
and
3Im
pro
ved
cen
tre
po
int
esti
mat
esan
dauxilia
ryp
oin
tca
lcu
lati
ons
are
sho
wn
for
gure
7V
aria
ble
sas
use
din
the
text
fle
ns
foca
lle
ngt
hH
al
titu
de
PC
ph
oto
gra
ph
cen
tre
poin
tSN
sp
ace
craft
nadir
po
int
D
dis
tance
cover
edo
nth
egr
oun
d
Pd
ista
nce
on
the
grou
nd
repre
sen
ted
by
apix
el
OVse
td
ista
nce
bet
wee
nP
Cand
SNusi
ng
the
grea
tci
rcle
form
ula
t
look
angle
away
from
SN
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
2L
ake
Eyre
ST
S093
-717
-80
250
276
shy29
0137
0shy
28
413
71
607
11
767
413
7609
ndash64
212
013
760
9ndash64
7121
124
ST
S082
-754
-61
250
617
shy28
5137
0shy
28
113
50
135
726
1201
18
013
8ndash14
827
517
9138ndash15
3277
294
Fig
ure
3H
ou
sto
nS
TS
062
-97-
151
4029
829
5shy
95
530
5shy
95
4410
78
8112
20
5348ndash58
990
1205
351
ndash63
8951
10
36
ST
S094
-737
-28
100
294
295
shy
95
528
3shy
95
5162
31
1133
24
415
8ndash20
334
724
3158ndash20
7351
390
ST
S055
-71-
41250
302
295
shy
95
028
2shy
93
766
412
8192
32
5736
ndash84
715
232
273
9ndash85
7154
185
S73
E51
45300
2685
295
shy
95
024
78
0F
igu
re4
Haw
aii
ST
S069
-701
-65
250
370
195
shy
155
520
4shy
153
881
415
7204
28
9876
ndash98
917
928
688
0ndash10
9181
210
ST
S069
-701
-70
250
370
210
shy
157
519
2shy
150
781
415
7738
63
414
9ndash23
336
760
7148ndash55
6394
10
63
ST
S069
-729
-27
100
398
200
shy
156
620
5shy
148
2219
42
1867
65
432
8ndash13
10
158
62
4323+
311
N
A
Astronaut photography as digital data for remote sensing 4431
Tab
le5
(Cont
inued
)
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
7L
imm
enB
igh
tS
TS
093
-704
-50
250
283
shy15
0135
0shy
15
013
58
623
120
859
16
963
0ndash67
3125
16
863
0ndash67
612
613
2P
CR
e-es
tim
ated
shy14
733
1352
67
64
5128
623
ndash65
5123
128
623
ndash65
712
3127
Auxilia
ryP
oin
t611
062
3ndash65
112
312
5F
igu
re10
N
ear
Au
stin
T
XS
TS
095
-716
-46
250
543
30
5shy
975
28
6shy
1007
120
230
375
34
613
5ndash157
281
34
1136ndash16
128
636
0
1 All
pho
togr
aphs
exce
pt
on
eta
ken
wit
hH
ass
elbla
dca
mer
aso
dis
tan
ceacr
oss
the
image
d=
55m
m
for
S73
E51
45
taken
wit
hel
ectr
onic
still
cam
erad=
276
5m
mho
rizo
nta
l2 P
Cand
SN
are
giv
enin
the
form
(lati
tud
elo
ngit
ude)
deg
rees
w
ith
neg
ativ
en
um
ber
sre
pre
senti
ng
south
and
wes
tA
ccura
cyo
fP
Cis
plusmn0
5and
SN
isplusmn
01
as
reco
rded
inth
eA
stro
naut
Ph
oto
gra
phy
Data
base
ex
cep
tw
her
en
ote
dth
atce
ntr
ep
oin
tw
asre
-est
imate
dw
ith
grea
ter
accu
racy
3 C
alc
ula
ted
usi
ng
the
mea
nofth
em
inim
um
an
dm
axim
um
esti
mate
sof
DF
orm
ula
tion
2co
nsi
der
sD
inth
ed
irec
tion
per
pen
dic
ula
rto
the
Pri
nci
pal
Lin
eo
nly
4 C
alc
ula
ted
inho
rizo
nta
lan
dve
rtic
aldir
ecti
on
sb
ut
still
ass
um
ing
geom
etry
of
gure
6(B
)F
irst
nu
mb
eris
the
mea
no
fth
em
inim
um
Dat
the
bott
om
of
the
ph
oto
and
the
maxi
mum
Dat
the
top
of
the
ph
oto
(range
show
nin
the
tab
le)
and
isco
mpara
ble
toF
orm
ula
tio
n2
Sec
ond
num
ber
isth
em
ean
of
Des
tim
ated
alon
gth
eP
rin
cip
alline
and
alo
ng
the
ou
tsid
eed
ge
(range
not
show
n)
5 Alt
itu
de
isap
pro
xim
ate
for
mis
sion
as
exac
tti
me
pho
togr
ap
hw
asta
ken
and
corr
esp
on
din
gn
adir
data
are
no
tava
ilab
le
Lack
of
nad
irdata
pre
ven
tsap
plica
tion
of
Form
ula
tion
s2
an
d3
6 Au
xilliar
ypo
int
add
edw
as
shy14
80
lati
tud
e13
51
2lo
ngit
ude
shy46
5degro
tati
on
angle
in
stru
ctio
ns
for
esti
mati
ng
the
rota
tio
nan
gle
para
met
erare
conta
ined
as
par
tof
the
scal
eca
lcu
lato
rsp
read
shee
tdo
cum
enta
tio
n
J A Robinson et al4432
a park at Johnson Space Center) that could clearly be distinguished on the photo-graph was 853 m wide
For the photograph of Houston taken with a 250 mm lens ( gure 3(C)) anddigitized from second generation lm at 2400 ppi (106 mm pixel Otilde 1 ) Ellington Fieldrunway 4-22 is 3 pixels in width and 1613 pixels in length so pixels represent151ndash152 m on the ground These results compare favourably with the minimumestimate of 152 m using Formulation 2 and 154ndash185 m pixels using Formulation 3(table 5) For an estimate of GRD using an 8times8 inch print (1369 enlargement) and4times magni cation the smallest street that could clearly be distinguished was 822 mwide the same feature could barely be distinguished on the digitized image
We also made an empirical estimate of spatial resolution for lower contrastvegetation boundaries By clearing forest so that a pattern would be visible to landingaircraft a landowner outside Austin Texas (see also aerial photo in Lisheron 2000)created a target that is also useful for evaluating spatial resolution of astronautphotographs The forest was selectively cleared in order to spell the landownerrsquosname lsquoLUECKErsquo with the remaining trees ( gure 10) According to local surveyorswho planned the clearing the plan was to create letters that were 3100 fttimes1700 ft(9449 mtimes5182 m) Photographed at a high altitude relative to most Shuttle missions(543 km) with a 250 mm lens Formula 3 predicts that each pixel would represent anarea 286 mtimes360 m on the ground (table 5) When original lm was digitized at2400 ppi (106 mm pixel Otilde 1 ) letters correspond to 294times188 pixels for a comparablepixel size of 27ndash32 m
6 Summary and conclusionsAstronaut photographs can be an excellent source of data for remote sensing
applications Best-case resolutions are similar to that for Landsat or SPOT withpixels as small as lt10 m It was estimated that of images taken to date 504 hadlens and obliquity characteristics that make them potential candidates for remotesensing information Digitized astronaut photographs can be overlaid with othersatellite data using GIS (Eckardt et al 2000) or used to ll in gaps in time serieswhen other imagery is not available As a source of public-domain information theycan be very useful for scientists who do not have access to satellite imagery eitherbecause of the expense of image acquisition (to the end user) or the computersystems needed for processing satellite images These diVerences in image acquisitioncosts (to the user) are summarized in table 6
Searching of the complete database of NASA astronaut photography includinglow-spatial-resolution browse images is available via the Web (OYce of EarthSciences 2000) This provides nearly global access for identifying images that willcontribute to a speci c scienti c project For the most detailed studies digitalproducts posted to the web will not be of suYcient quality To date we have madeit a practice of digitizing small numbers of images when requested by scientists atno charge Such requests (including a description of the project involved) can bemade through the authors or using the contact information listed on the websiteInvestigators with needs for larger numbers of digital images have also been servedthrough collaborative agreements
In our experience scientists that nd the data most valuable are those in develop-ing countries those studying areas that have not usually been targets for the majorsatellites those having diYculty in nding low-cloud images those interested inconstructing time series or those interested in using a large number of images
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
J A Robinson et al4422
Table 4 Comparison of pixel sizes (instantaneous elds of view) for satellite remote sensorswith pixel sizes for near vertical viewing photography from spacecraft Note that alldata returned from the automated satellites have the same eld of view while indicatedpixel sizes for astronaut photography are best-case for near vertical look angles See gure 5 for distributions of astronaut photographs of various look angles High-altitude aerial photography is also included for reference
Altitude PixelInstrument (km) width (mtimesm) Bands1
Landsat Multipectral Scanner2 (MSS 1974-) 880ndash940 79times56 4ndash5Landsat Thematic Mapper (TM 1982-) ~705 30times30 7Satellite pour lrsquoObservation de la Terre (SPOT 1986-) ~832
SPOT 1-3 Panchromatic 10times10 1SPOT 1-3 Multispectral (XS) 20times20 3
Astronaut Photography (1961-)Skylab3 (1973-1974 ) ~435
Multispectral Camera (6 camera stations) 17ndash19times17ndash19 3ndash8Earth Terrain Camera (hi-res colour) 875times875 3
Space Shuttle5 (1981-) 222ndash611Hasselblad 70 mm (250mm lens colour) vertical view 9ndash26times9ndash26 3Hasselblad 70 mm (100mm lens colour) vertical view 23ndash65times23ndash65 3Nikon 35 mm (400mm lens colour lm or ESC) vertical 5ndash16times5ndash16 3
High Altitude Aerial Photograph (1100 000) ~12 2times2 3
1Three bands listed for aerial photography obtainable by high-resolution colour digitizingFor high-resolution colour lm at 65 lp mmOtilde 1 (K Teitelbaum Eastman Kodak Co personalcommunication) the maximum information contained would be 3302 ppi
2Data from Simonett (1983Table 1-2)3Calculated as recommended by Jensen (19831596) the dividend of estimated ground
resolution at low contrast given in NASA (1974179-180) and 244Six lters plus colour and colour infrared lm (NASA 19747) 5Extracted from table 2
nadir point The look angle (the angle oV nadir that the camera is pointing) can becalculated trigonometrically by assuming a spherical earth and calculating the dis-tance between the coordinates of SN and PC ( gure 1) using the Great Circledistance haversine solution (Sinnott 1984 Snyder 198730ndash32 Chamberlain 1996)The diVerence between the spacecraft centre and nadir point latitudes Dlat=lat 2 shy lat 1 and the diVerence between the spacecraft centre and nadir pointlongitudes Dlon=lon 2 shy lon 1 enter the following equations
a=sin2ADlat
2 B+cos(lat 1) cos(lat 2) sin2ADlon
2 B (4)
c=2 arcsin(min[1 atilde a]) (5)
and oVset=R c with R the radius of a spherical Earth=6370 kmThe look angle (t ) is then given by
t=arctanAoffset
H B (6)
Assuming that the camera was positioned so that the imaginary line between thecentre and nadir points (the principal line) runs vertically through the centre of thephotograph the distance between the geometric centre of the photograph (principalpoint PP) and the top of the photograph is d2 ( gure 8(a)) The scale at any point
Astronaut photography as digital data for remote sensing 4423
(a) (b)
Figure 8 The geometric relationship between look angle lens focal length and the centrepoint and nadir points of a photograph (see also Estes and Simonett 1975) Note thatthe Earthrsquos surface and footprint represented in gure 1 are not shown here only arepresentation of the image area of the photograph Variables shown in the gurelook angle t focal length f PP centre of the image on the lm SN spacecraft nadird original image size (a) The special case where the camera is aligned squarely withthe isometric parallel In this case tilt occurs along a plane parallel with the centre ofthe photograph When the conditions are met calculations using Formulation 3 willgive the ground coordinates of the corner points of the photograph (b) The generalrelationship between these parameters and key photogrammetric points on the photo-graph For this more typical case coordinates of an additional point in the photographmust be input in order to adjust for twist of the camera and to nd the corner pointcoordinates
in the photograph varies as a function of the distance y along the principal linebetween the isocentre and the point according to the relationship
d
D=
f shy ysint
H(7)
(Wong 1980 equation (214) H amp the ground elevation h) Using gure 8(a) at thetop of the photo
y= f tanAt
2B+d
2(8)
and at the bottom of the photo
y= f tanAt
2Bshyd
2(9)
Thus for given PC and SN coordinates and assuming a photo orientation as in gure 8(a) and not gure 8(b) we can estimate a minimum D (using equations (7)and (8)) and a maximum D (using equations (7) and (9)) and then average the twoto determine the pixel size (P ) via equation (2)
J A Robinson et al4424
53 Formulation 3 T he L ow Oblique Space Photo Footprint CalculatorThe Low Oblique Space Photo Footprint Calculator was developed to provide
a more accurate estimation of the geographic coordinates for the footprint of a lowoblique photo of the Earthrsquos surface taken from a human-occupied spacecraft inorbit The calculator performs a series of three-dimensional coordinate transforma-tions to compute the location and orientation of the centre of the photo exposureplane relative to an earth referenced coordinate system The nominal camera focallength is then used to create a vector from the photorsquos perspective point througheach of eight points around the parameter of the image as de ned by the formatsize The geographical coordinates for the photo footprint are then computed byintersecting these photo vectors with a spherical earth model Although more sophist-icated projection algorithms are available no signi cant increase in the accuracy ofthe results would be produced by these algorithms due to inherent uncertainties inthe available input data (ie the spacecraft altitude photo centre location etc)
The calculations were initially implemented within a Microsoft Excel workbookwhich allowed one to embed the mathematical and graphical documentation nextto the actual calculations Thus interested users can inspect the mathematical pro-cessing A set of error traps was also built into the calculations to detect erroneousresults A summary of any errors generated is reported to the user with the resultsof the calculations Interested individuals are invited to download the Excel work-book from httpeoljscnasagovsseopFootprintCalculatorxls The calculations arecurrently being encoded in a high-level programming language and should soon beavailable alongside other background data provided for each photograph at OYceof Earth Sciences (2000)
For the purposes of these calculations a low oblique photograph was de ned as
one with the centre within 10 degrees of latitude and longitude of the spacecraftnadir point For the typical range of spacecraft altitudes to date this restricted thecalculations to photographs in which Earthrsquos horizon does not appear (the generalde nition of a low oblique photograph eg Campbell 199671 and gure 4)
531 Input data and resultsUpon opening the Excel workbook the user is presented with a program intro-
duction providing instructions for using the calculator The second worksheet tablsquoHow-To-Usersquo provides detailed step-by-step instructions for preparing the baselinedata for the program The third tab lsquoInput-Output rsquo contains the user input eldsand displays the results of the calculations The additional worksheets contain theactual calculations and program documentation Although users are welcome toreview these sheets an experienced user need only access the lsquoInput-Outputrsquo
spreadsheetTo begin a calculation the user enters the following information which is available
for each photo in the NASA Astronaut Photography Database (1) SN geographicalcoordinates of spacecraft nadir position at the time of photo (2) H spacecraftaltitude (3) PC the geographical coordinates of the centre of the photo (4) f nominal focal length and (5) d image format size The automatic implementationof the workbook on the web will automatically enter these values and completecalculations
For more accurate results the user may optionally enter the geographic co-ordinates and orientation for an auxiliary point on the photo which resolves thecamerarsquos rotation uncertainty about the optical axis The auxiliary point data must
Astronaut photography as digital data for remote sensing 4425
be computed by the user following the instruction contained in the lsquoHow-To-Usersquotab of the spreadsheet
After entering the input data the geographic coordinates of the photo footprint(ie four photo corner points four points at the bisector of each edge and the centreof the photo) are immediately displayed below the input elds along with any errormessages generated by the user input or by the calculations ( gure 7) Although
results are computed and displayed they should not be used when error messagesare produced by the program The program also computes the tilt angle for each ofthe photo vectors relative to the spacecraft nadir vector To the right of the photofootprint coordinates is displayed the arc distance along the surface of the sphere
between adjacent computed points
532 Calculation assumptions
The mathematical calculations implemented in the Low Oblique Space PhotoFootprint Calculator use the following assumptions
1 The SN location is used as exact even though the true value may vary by upto plusmn01 degree from the location provided with the photo
2 The spacecraft altitude is used as exact Although the determination of the
nadir point at the instant of a known spacecraft vector is relatively precise(plusmn115times10 Otilde 4 degrees) the propagator interpolates between sets of approxi-mately 10ndash40 known vectors per day and the time code recorded on the lmcan drift Thus the true value for SN may vary by up to plusmn01 degree fromthe value provided with the photo
3 The perspective centre of the camera is assumed to be at the given altitudeover the speci ed spacecraft nadir location at the time of photo exposure
4 The PC location is used as exact even though the true value may vary by upto plusmn05deg latitude and plusmn05deg longitude from the location provided with the
photo5 A spherical earth model is used with a nominal radius of 6 372 16154 m (a
common rst-order approximation for a spherical earth used in geodeticcomputations)
6 The nominal lens focal length of the camera lens is used in the computations(calibrated focal length values are not available)
7 The photo projection is based on the classic pin-hole camera model8 No correction for lens distortion or atmospheric re ection is made
9 If no auxiliary point data is provided the lsquoTop of the Imagersquo is orientedperpendicular to the vector from SN towards PC
533 T ransformation from Earth to photo coordinate systemsThe calculations begin by converting the geographic coordinates ( latitude and
longitude) of the SN and PC to a Rectangular Earth-Centred Coordinate System(R-Earth) de ned as shown in gure 9 (with the centre of the Earth at (0 0 0))
Using the vector from the Earthrsquos centre through SN and the spacecraft altitude thespacecraft location (SC) is also computed in R-Earth
For ease of computation a Rectangular Spacecraft-Centred Coordinate System(R-Spacecraft ) is de ned as shown in gure 9 The origin of R-Spacecraft is located
at SC with its +Z-axis aligned with vector from the centre of the Earth throughSN and its +X-axis aligned with the vector from SN to PC ( gure 9) The speci c
J A Robinson et al4426
Figure 9 Representation of the area included in a photograph as a geometric projectiononto the Earthrsquos surface Note that the focal length (the distance from the SpaceShuttle to the photo) is grossly overscale compared to the altitude (the distance fromthe Shuttle to the surface of the Earth
rotations and translation used to convert from the R-Earth to the R-Spacecraft arecomputed and documented in the spreadsheet
With the mathematical positions of the SC SN PC and the centre of the Earthcomputed in R-Spacecraft the program next computes the location of the camerarsquosprincipal point (PP) The principle point is the point of intersection of the opticalaxis of the lens with the image plane (ie the lm) It is nominally positioned at a
Astronaut photography as digital data for remote sensing 4427
distance equal to the focal length from the perspective centre of the camera (whichis assumed to be at SC) along the vector from PC through SC as shown in gure 9
A third coordinate system the Rectangular Photo Coordinate System (R-Photo)is created with its origin at PP its XndashY axial plane normal to the vector from PCthrough SC and its +X axis aligned with the +X axis of R-Spacecraft as shownin gure 9 The XndashY plane of this coordinate system represents the image plane ofthe photograph
534 Auxiliary point calculationsThe calculations above employ a critical assumption that all photos are taken
with the lsquotop of the imagersquo oriented perpendicular to the vector from the SN towardsPC as shown in gure 9 To avoid non-uniform solar heating of the external surfacemost orbiting spacecraft are slowly and continually rotated about one or more axesIn this condition a ight crew member taking a photo while oating in microgravitycould orient the photo with practically any orientation relative to the horizon (seealso gure 8(b)) Unfortunately since these photos are taken with conventionalhandheld cameras there is no other information available which can be used toresolve the photorsquos rotational ambiguity about the optical axis other then the photoitself This is why the above assumption is used and the footprint computed by thiscalculator is actually a lsquolocator ellipsersquo which estimates the area on the ground whatis likely to be included in a speci c photograph (see gure 6) This locator ellipse ismost accurate for square image formats and subject to additional distortion as thephotograph format becomes more rectangular
If the user wants a more precise calculation of the image footprint the photorsquosrotational ambiguity about the optical axis must be resolved This can be done inthe calculator by adding data for an auxiliary point Detailed instructions regardinghow to prepare and use auxiliary point data in the computations are included in thelsquoHow-To-Usersquo tab of the spreadsheet Basically the user determines which side ofthe photograph is top and then measures the angle between the line from PP to thetop of the photo and from PP to the auxiliary point on the photo ( gure 9)
If the user includes data for an auxiliary point a series of computations arecompleted to resolve the photo rotation ambiguity about the optical axis (ie the
+Z axis in R-Photo) A vector from the Auxiliary Point on the Earth (AE) throughthe photograph perspective centre ( located at SC) is intersected with the photo imageplane (XndashY plane of R-Photo) to compute the coordinates of the Auxiliary Point onthe photo (AP) in R-Photo A two-dimensional angle in the XndashY plane of R-Photofrom the shy X axis to a line from PP to AP is calculated as shown in gure 9 The
shy X axis is used as the origin of the angle since it represents the top of the photoonce it passes through the perspective centre The diVerence between the computedangle and the angle measured by the user on the photo resolves the ambiguity inthe rotation of the photo relative to the principal line ( gures 7 and 9) Thetransformations from R-Spacecraft and R-Photo are then modi ed to include anadditional rotation angle about the +Z axis in R-Photo
535 lsquoFootprint rsquo calculationsThe program next computes the coordinates of eight points about the perimeter
of the image format (ie located at the four photo corners plus a bisector pointalong each edge of the image) These points are identi ed in R-Photo based uponthe photograph format size and then converted to R-Spacecraft Since all computa-tions are done in orthogonal coordinate systems the R-Spacecraft to R-Photo
J A Robinson et al4428
rotation matrix is transposed to produce an R-Photo to R-Spacecraft rotation matrixOnce in R-Spacecraft a unit vector from each of the eight perimeter points throughthe photo perspective centre (the same point as SC) is computed This provides thecoordinates for points about the perimeter of the image format with their directionvectors in a common coordinate system with other key points needed to computethe photo footprint
The next step is to compute the point of intersection between the spherical earthmodel and each of the eight perimeter point vectors The scalar value for eachperimeter point unit vector is computed using two-dimensional planar trigonometryAn angle c is computed using the formula for the cosine of an angle between twoincluded vectors (the perimeter point unit vector and the vector from SC to thecentre of the Earth) Angle y is computed using the Law of Sines Angle e=180degrees shy c shy y The scalar value of the perimeter point vector is computed using eand the Law of Cosines The scalar value is then multiplied by the perimeter pointunit vector to produce the three-dimensional point of intersection of the vector withEarthrsquos surface in R-Spacecraft The process is repeated independently for each ofthe eight perimeter point vectors Aside from its mathematical simplicity the valueof arcsine (computed in step 2) will exceed the normal range for the sin of an anglewhen the perimeter point vectors fail to intersect with the surface of the earth Asimple test based on this principle allows the program to correctly handle obliquephotos which image a portion of the horizon (see results for high oblique photographsin table 5)
The nal step in the process converts the eight earth intersection points from theR-Spacecraft to R-Earth The results are then converted to the standard geographiccoordinate system and displayed on the lsquoInput-Outputrsquo page of the spreadsheet
54 ExamplesAll three formulations were applied to the photographs included in this paper
and the results are compared in table 5 Formulation 1 gives too small a value fordistance across the photograph (D) for all but the most nadir shots and thus servesas an indicator of the best theoretical case but is not a good measure for a speci cphotograph For example the photograph of Lake Eyre taken from 276 km altitude( gure 2(a)) and the photograph of Limmen Bight ( gure 7) were closest to beingnadir views (oVsetslt68 km or tlt15deg table 5) For these photographs D calculatedusing Formulation 1 was similar to the minimum D calculated using Formulations2 and 3 For almost all other more oblique photographs Formulation 1 gave asigni cant underestimate of the distance covered in the photograph For gure 5(A)(the picture of Houston taken with a 40 mm lens) Formulation 1 did not give anunderestimate for D This is because Formulation 1 does not account for curvatureof the Earth in any way With this large eld of view assuming a at Earth in atedthe value of D above the minimum from calculations that included Earth curvature
A major diVerence between Formulations 2 and 3 is the ability to estimate pixelsizes (P) in both directions (along the principal line and perpendicular to the principalline) For the more oblique photographs the vertical estimate of D and pixel sizes ismuch larger than in the horizontal direction (eg the low oblique and high obliquephotographs of Hawaii gure 4 table 5)
For the area of Limmen Bight ( gure 7) table 5 illustrates the improvement inthe estimate of distance and pixel size that can be obtained by re-estimating thelocation of the PC with greater accuracy Centre points in the catalogued data are
Astronaut photography as digital data for remote sensing 4429
plusmn05deg of latitude and longitude When the centre point was re-estimated to plusmn002degwe determined that the photograph was not taken as obliquely as rst thought(change in the estimate of the look angle t from 169 to 128deg table 5) When theauxiliary point was added to the calculations of Formula 3 the calculated look angleshrank further to 110deg indicating that this photograph was taken at very close toa nadir view Of course this improvement in accuracy could have also led to estimatesof greater obliquity and corresponding larger pixel sizes
We also tested the performance of the scale calculator with auxiliary point byestimating the corner and point locations on the photograph using a 11 000 000Operational Navigational Chart For this test we estimated our ability to readcoordinates from the map as plusmn002degand our error in nding the locations of thecorner points as plusmn015deg (this error varies among photographs depending on thedetail that can be matched between photo and map) For Limmen Bight ( gure 7)the mean diVerence between map estimates and calculator estimates for four pointswas 031deg (SD=018 n=8) For a photograph of San Francisco Bay (STS062-151-291) the mean diVerence between map estimates and calculator estimates forfour points was 0064deg (SD=018 n=8) For a photograph of San Francisco Bay(STS062-151-291) the mean diVerence between map estimates and calculator estim-ates for eight points was 0196deg (SD=0146 n=16) Thus in one case the calculatorestimates were better than our estimate of the error in locating corner points on themap It is reasonable to expect that for nadir-viewing photographs the calculatorused with an auxiliary point can estimate locations of the edges of a photograph towithin plusmn03deg
55 Empirical con rmation of spatial resolution estimatesAs stated previously a challenge to estimating system-AWAR for astronaut
photography of Earth is the lack of suitable targets Small features in an image cansometimes be used as a check on the size of objects that can be successfully resolvedgiving an approximate value for GRD Similarly the number of pixels that make upthose features in the digitized image can be used to make an independent calculationof pixel size We have successfully used features such as roads and airport runwaysto make estimates of spatial scale and resolution (eg Robinson et al 2000c) Whilerecognizing that the use of linear detail in an image is a poor approximation to abar target and that linear objects smaller than the resolving power can often bedetected (Charman 1965) few objects other than roads could be found to make anydirect estimates of GRD Thus roads and runways were used in the images ofHouston (where one can readily conduct ground veri cations and where a numberof higher-contrast concrete roadways were available) to obtain empirical estimatesof GRD and pixel size for comparison with table 5
In the all-digital ESC image of Houston ( gure 3(d )) we examined EllingtonField runway 4-22 (centre left of the image) which is 24384 mtimes457 m This runwayis approximately 6ndash7 pixels in width and 304ndash3094 pixels in length so pixelsrepresent an distance on the ground 7ndash8 m Using a lower contrast measure of astreet length between two intersections (2125 m=211 pixels) a pixel width of 101 mis estimated These results compare favourably with the minimum estimate of 81 mpixels using Formulation 1 (table 5) For an estimate of GRD that would be morecomparable to aerial photography of a line target the smallest street without treecover that could be clearly distinguished on the photograph was 792 m wide Thesmallest non-street object (a gap between stages of the Saturn rocket on display in
J A Robinson et al4430
Tab
le5
App
lica
tio
nof
met
hod
sfo
res
tim
ati
ng
spati
alre
solu
tio
no
fast
ron
aut
ph
oto
gra
ph
sin
clu
din
gF
orm
ula
tion
s12
and
3Im
pro
ved
cen
tre
po
int
esti
mat
esan
dauxilia
ryp
oin
tca
lcu
lati
ons
are
sho
wn
for
gure
7V
aria
ble
sas
use
din
the
text
fle
ns
foca
lle
ngt
hH
al
titu
de
PC
ph
oto
gra
ph
cen
tre
poin
tSN
sp
ace
craft
nadir
po
int
D
dis
tance
cover
edo
nth
egr
oun
d
Pd
ista
nce
on
the
grou
nd
repre
sen
ted
by
apix
el
OVse
td
ista
nce
bet
wee
nP
Cand
SNusi
ng
the
grea
tci
rcle
form
ula
t
look
angle
away
from
SN
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
2L
ake
Eyre
ST
S093
-717
-80
250
276
shy29
0137
0shy
28
413
71
607
11
767
413
7609
ndash64
212
013
760
9ndash64
7121
124
ST
S082
-754
-61
250
617
shy28
5137
0shy
28
113
50
135
726
1201
18
013
8ndash14
827
517
9138ndash15
3277
294
Fig
ure
3H
ou
sto
nS
TS
062
-97-
151
4029
829
5shy
95
530
5shy
95
4410
78
8112
20
5348ndash58
990
1205
351
ndash63
8951
10
36
ST
S094
-737
-28
100
294
295
shy
95
528
3shy
95
5162
31
1133
24
415
8ndash20
334
724
3158ndash20
7351
390
ST
S055
-71-
41250
302
295
shy
95
028
2shy
93
766
412
8192
32
5736
ndash84
715
232
273
9ndash85
7154
185
S73
E51
45300
2685
295
shy
95
024
78
0F
igu
re4
Haw
aii
ST
S069
-701
-65
250
370
195
shy
155
520
4shy
153
881
415
7204
28
9876
ndash98
917
928
688
0ndash10
9181
210
ST
S069
-701
-70
250
370
210
shy
157
519
2shy
150
781
415
7738
63
414
9ndash23
336
760
7148ndash55
6394
10
63
ST
S069
-729
-27
100
398
200
shy
156
620
5shy
148
2219
42
1867
65
432
8ndash13
10
158
62
4323+
311
N
A
Astronaut photography as digital data for remote sensing 4431
Tab
le5
(Cont
inued
)
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
7L
imm
enB
igh
tS
TS
093
-704
-50
250
283
shy15
0135
0shy
15
013
58
623
120
859
16
963
0ndash67
3125
16
863
0ndash67
612
613
2P
CR
e-es
tim
ated
shy14
733
1352
67
64
5128
623
ndash65
5123
128
623
ndash65
712
3127
Auxilia
ryP
oin
t611
062
3ndash65
112
312
5F
igu
re10
N
ear
Au
stin
T
XS
TS
095
-716
-46
250
543
30
5shy
975
28
6shy
1007
120
230
375
34
613
5ndash157
281
34
1136ndash16
128
636
0
1 All
pho
togr
aphs
exce
pt
on
eta
ken
wit
hH
ass
elbla
dca
mer
aso
dis
tan
ceacr
oss
the
image
d=
55m
m
for
S73
E51
45
taken
wit
hel
ectr
onic
still
cam
erad=
276
5m
mho
rizo
nta
l2 P
Cand
SN
are
giv
enin
the
form
(lati
tud
elo
ngit
ude)
deg
rees
w
ith
neg
ativ
en
um
ber
sre
pre
senti
ng
south
and
wes
tA
ccura
cyo
fP
Cis
plusmn0
5and
SN
isplusmn
01
as
reco
rded
inth
eA
stro
naut
Ph
oto
gra
phy
Data
base
ex
cep
tw
her
en
ote
dth
atce
ntr
ep
oin
tw
asre
-est
imate
dw
ith
grea
ter
accu
racy
3 C
alc
ula
ted
usi
ng
the
mea
nofth
em
inim
um
an
dm
axim
um
esti
mate
sof
DF
orm
ula
tion
2co
nsi
der
sD
inth
ed
irec
tion
per
pen
dic
ula
rto
the
Pri
nci
pal
Lin
eo
nly
4 C
alc
ula
ted
inho
rizo
nta
lan
dve
rtic
aldir
ecti
on
sb
ut
still
ass
um
ing
geom
etry
of
gure
6(B
)F
irst
nu
mb
eris
the
mea
no
fth
em
inim
um
Dat
the
bott
om
of
the
ph
oto
and
the
maxi
mum
Dat
the
top
of
the
ph
oto
(range
show
nin
the
tab
le)
and
isco
mpara
ble
toF
orm
ula
tio
n2
Sec
ond
num
ber
isth
em
ean
of
Des
tim
ated
alon
gth
eP
rin
cip
alline
and
alo
ng
the
ou
tsid
eed
ge
(range
not
show
n)
5 Alt
itu
de
isap
pro
xim
ate
for
mis
sion
as
exac
tti
me
pho
togr
ap
hw
asta
ken
and
corr
esp
on
din
gn
adir
data
are
no
tava
ilab
le
Lack
of
nad
irdata
pre
ven
tsap
plica
tion
of
Form
ula
tion
s2
an
d3
6 Au
xilliar
ypo
int
add
edw
as
shy14
80
lati
tud
e13
51
2lo
ngit
ude
shy46
5degro
tati
on
angle
in
stru
ctio
ns
for
esti
mati
ng
the
rota
tio
nan
gle
para
met
erare
conta
ined
as
par
tof
the
scal
eca
lcu
lato
rsp
read
shee
tdo
cum
enta
tio
n
J A Robinson et al4432
a park at Johnson Space Center) that could clearly be distinguished on the photo-graph was 853 m wide
For the photograph of Houston taken with a 250 mm lens ( gure 3(C)) anddigitized from second generation lm at 2400 ppi (106 mm pixel Otilde 1 ) Ellington Fieldrunway 4-22 is 3 pixels in width and 1613 pixels in length so pixels represent151ndash152 m on the ground These results compare favourably with the minimumestimate of 152 m using Formulation 2 and 154ndash185 m pixels using Formulation 3(table 5) For an estimate of GRD using an 8times8 inch print (1369 enlargement) and4times magni cation the smallest street that could clearly be distinguished was 822 mwide the same feature could barely be distinguished on the digitized image
We also made an empirical estimate of spatial resolution for lower contrastvegetation boundaries By clearing forest so that a pattern would be visible to landingaircraft a landowner outside Austin Texas (see also aerial photo in Lisheron 2000)created a target that is also useful for evaluating spatial resolution of astronautphotographs The forest was selectively cleared in order to spell the landownerrsquosname lsquoLUECKErsquo with the remaining trees ( gure 10) According to local surveyorswho planned the clearing the plan was to create letters that were 3100 fttimes1700 ft(9449 mtimes5182 m) Photographed at a high altitude relative to most Shuttle missions(543 km) with a 250 mm lens Formula 3 predicts that each pixel would represent anarea 286 mtimes360 m on the ground (table 5) When original lm was digitized at2400 ppi (106 mm pixel Otilde 1 ) letters correspond to 294times188 pixels for a comparablepixel size of 27ndash32 m
6 Summary and conclusionsAstronaut photographs can be an excellent source of data for remote sensing
applications Best-case resolutions are similar to that for Landsat or SPOT withpixels as small as lt10 m It was estimated that of images taken to date 504 hadlens and obliquity characteristics that make them potential candidates for remotesensing information Digitized astronaut photographs can be overlaid with othersatellite data using GIS (Eckardt et al 2000) or used to ll in gaps in time serieswhen other imagery is not available As a source of public-domain information theycan be very useful for scientists who do not have access to satellite imagery eitherbecause of the expense of image acquisition (to the end user) or the computersystems needed for processing satellite images These diVerences in image acquisitioncosts (to the user) are summarized in table 6
Searching of the complete database of NASA astronaut photography includinglow-spatial-resolution browse images is available via the Web (OYce of EarthSciences 2000) This provides nearly global access for identifying images that willcontribute to a speci c scienti c project For the most detailed studies digitalproducts posted to the web will not be of suYcient quality To date we have madeit a practice of digitizing small numbers of images when requested by scientists atno charge Such requests (including a description of the project involved) can bemade through the authors or using the contact information listed on the websiteInvestigators with needs for larger numbers of digital images have also been servedthrough collaborative agreements
In our experience scientists that nd the data most valuable are those in develop-ing countries those studying areas that have not usually been targets for the majorsatellites those having diYculty in nding low-cloud images those interested inconstructing time series or those interested in using a large number of images
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
Astronaut photography as digital data for remote sensing 4423
(a) (b)
Figure 8 The geometric relationship between look angle lens focal length and the centrepoint and nadir points of a photograph (see also Estes and Simonett 1975) Note thatthe Earthrsquos surface and footprint represented in gure 1 are not shown here only arepresentation of the image area of the photograph Variables shown in the gurelook angle t focal length f PP centre of the image on the lm SN spacecraft nadird original image size (a) The special case where the camera is aligned squarely withthe isometric parallel In this case tilt occurs along a plane parallel with the centre ofthe photograph When the conditions are met calculations using Formulation 3 willgive the ground coordinates of the corner points of the photograph (b) The generalrelationship between these parameters and key photogrammetric points on the photo-graph For this more typical case coordinates of an additional point in the photographmust be input in order to adjust for twist of the camera and to nd the corner pointcoordinates
in the photograph varies as a function of the distance y along the principal linebetween the isocentre and the point according to the relationship
d
D=
f shy ysint
H(7)
(Wong 1980 equation (214) H amp the ground elevation h) Using gure 8(a) at thetop of the photo
y= f tanAt
2B+d
2(8)
and at the bottom of the photo
y= f tanAt
2Bshyd
2(9)
Thus for given PC and SN coordinates and assuming a photo orientation as in gure 8(a) and not gure 8(b) we can estimate a minimum D (using equations (7)and (8)) and a maximum D (using equations (7) and (9)) and then average the twoto determine the pixel size (P ) via equation (2)
J A Robinson et al4424
53 Formulation 3 T he L ow Oblique Space Photo Footprint CalculatorThe Low Oblique Space Photo Footprint Calculator was developed to provide
a more accurate estimation of the geographic coordinates for the footprint of a lowoblique photo of the Earthrsquos surface taken from a human-occupied spacecraft inorbit The calculator performs a series of three-dimensional coordinate transforma-tions to compute the location and orientation of the centre of the photo exposureplane relative to an earth referenced coordinate system The nominal camera focallength is then used to create a vector from the photorsquos perspective point througheach of eight points around the parameter of the image as de ned by the formatsize The geographical coordinates for the photo footprint are then computed byintersecting these photo vectors with a spherical earth model Although more sophist-icated projection algorithms are available no signi cant increase in the accuracy ofthe results would be produced by these algorithms due to inherent uncertainties inthe available input data (ie the spacecraft altitude photo centre location etc)
The calculations were initially implemented within a Microsoft Excel workbookwhich allowed one to embed the mathematical and graphical documentation nextto the actual calculations Thus interested users can inspect the mathematical pro-cessing A set of error traps was also built into the calculations to detect erroneousresults A summary of any errors generated is reported to the user with the resultsof the calculations Interested individuals are invited to download the Excel work-book from httpeoljscnasagovsseopFootprintCalculatorxls The calculations arecurrently being encoded in a high-level programming language and should soon beavailable alongside other background data provided for each photograph at OYceof Earth Sciences (2000)
For the purposes of these calculations a low oblique photograph was de ned as
one with the centre within 10 degrees of latitude and longitude of the spacecraftnadir point For the typical range of spacecraft altitudes to date this restricted thecalculations to photographs in which Earthrsquos horizon does not appear (the generalde nition of a low oblique photograph eg Campbell 199671 and gure 4)
531 Input data and resultsUpon opening the Excel workbook the user is presented with a program intro-
duction providing instructions for using the calculator The second worksheet tablsquoHow-To-Usersquo provides detailed step-by-step instructions for preparing the baselinedata for the program The third tab lsquoInput-Output rsquo contains the user input eldsand displays the results of the calculations The additional worksheets contain theactual calculations and program documentation Although users are welcome toreview these sheets an experienced user need only access the lsquoInput-Outputrsquo
spreadsheetTo begin a calculation the user enters the following information which is available
for each photo in the NASA Astronaut Photography Database (1) SN geographicalcoordinates of spacecraft nadir position at the time of photo (2) H spacecraftaltitude (3) PC the geographical coordinates of the centre of the photo (4) f nominal focal length and (5) d image format size The automatic implementationof the workbook on the web will automatically enter these values and completecalculations
For more accurate results the user may optionally enter the geographic co-ordinates and orientation for an auxiliary point on the photo which resolves thecamerarsquos rotation uncertainty about the optical axis The auxiliary point data must
Astronaut photography as digital data for remote sensing 4425
be computed by the user following the instruction contained in the lsquoHow-To-Usersquotab of the spreadsheet
After entering the input data the geographic coordinates of the photo footprint(ie four photo corner points four points at the bisector of each edge and the centreof the photo) are immediately displayed below the input elds along with any errormessages generated by the user input or by the calculations ( gure 7) Although
results are computed and displayed they should not be used when error messagesare produced by the program The program also computes the tilt angle for each ofthe photo vectors relative to the spacecraft nadir vector To the right of the photofootprint coordinates is displayed the arc distance along the surface of the sphere
between adjacent computed points
532 Calculation assumptions
The mathematical calculations implemented in the Low Oblique Space PhotoFootprint Calculator use the following assumptions
1 The SN location is used as exact even though the true value may vary by upto plusmn01 degree from the location provided with the photo
2 The spacecraft altitude is used as exact Although the determination of the
nadir point at the instant of a known spacecraft vector is relatively precise(plusmn115times10 Otilde 4 degrees) the propagator interpolates between sets of approxi-mately 10ndash40 known vectors per day and the time code recorded on the lmcan drift Thus the true value for SN may vary by up to plusmn01 degree fromthe value provided with the photo
3 The perspective centre of the camera is assumed to be at the given altitudeover the speci ed spacecraft nadir location at the time of photo exposure
4 The PC location is used as exact even though the true value may vary by upto plusmn05deg latitude and plusmn05deg longitude from the location provided with the
photo5 A spherical earth model is used with a nominal radius of 6 372 16154 m (a
common rst-order approximation for a spherical earth used in geodeticcomputations)
6 The nominal lens focal length of the camera lens is used in the computations(calibrated focal length values are not available)
7 The photo projection is based on the classic pin-hole camera model8 No correction for lens distortion or atmospheric re ection is made
9 If no auxiliary point data is provided the lsquoTop of the Imagersquo is orientedperpendicular to the vector from SN towards PC
533 T ransformation from Earth to photo coordinate systemsThe calculations begin by converting the geographic coordinates ( latitude and
longitude) of the SN and PC to a Rectangular Earth-Centred Coordinate System(R-Earth) de ned as shown in gure 9 (with the centre of the Earth at (0 0 0))
Using the vector from the Earthrsquos centre through SN and the spacecraft altitude thespacecraft location (SC) is also computed in R-Earth
For ease of computation a Rectangular Spacecraft-Centred Coordinate System(R-Spacecraft ) is de ned as shown in gure 9 The origin of R-Spacecraft is located
at SC with its +Z-axis aligned with vector from the centre of the Earth throughSN and its +X-axis aligned with the vector from SN to PC ( gure 9) The speci c
J A Robinson et al4426
Figure 9 Representation of the area included in a photograph as a geometric projectiononto the Earthrsquos surface Note that the focal length (the distance from the SpaceShuttle to the photo) is grossly overscale compared to the altitude (the distance fromthe Shuttle to the surface of the Earth
rotations and translation used to convert from the R-Earth to the R-Spacecraft arecomputed and documented in the spreadsheet
With the mathematical positions of the SC SN PC and the centre of the Earthcomputed in R-Spacecraft the program next computes the location of the camerarsquosprincipal point (PP) The principle point is the point of intersection of the opticalaxis of the lens with the image plane (ie the lm) It is nominally positioned at a
Astronaut photography as digital data for remote sensing 4427
distance equal to the focal length from the perspective centre of the camera (whichis assumed to be at SC) along the vector from PC through SC as shown in gure 9
A third coordinate system the Rectangular Photo Coordinate System (R-Photo)is created with its origin at PP its XndashY axial plane normal to the vector from PCthrough SC and its +X axis aligned with the +X axis of R-Spacecraft as shownin gure 9 The XndashY plane of this coordinate system represents the image plane ofthe photograph
534 Auxiliary point calculationsThe calculations above employ a critical assumption that all photos are taken
with the lsquotop of the imagersquo oriented perpendicular to the vector from the SN towardsPC as shown in gure 9 To avoid non-uniform solar heating of the external surfacemost orbiting spacecraft are slowly and continually rotated about one or more axesIn this condition a ight crew member taking a photo while oating in microgravitycould orient the photo with practically any orientation relative to the horizon (seealso gure 8(b)) Unfortunately since these photos are taken with conventionalhandheld cameras there is no other information available which can be used toresolve the photorsquos rotational ambiguity about the optical axis other then the photoitself This is why the above assumption is used and the footprint computed by thiscalculator is actually a lsquolocator ellipsersquo which estimates the area on the ground whatis likely to be included in a speci c photograph (see gure 6) This locator ellipse ismost accurate for square image formats and subject to additional distortion as thephotograph format becomes more rectangular
If the user wants a more precise calculation of the image footprint the photorsquosrotational ambiguity about the optical axis must be resolved This can be done inthe calculator by adding data for an auxiliary point Detailed instructions regardinghow to prepare and use auxiliary point data in the computations are included in thelsquoHow-To-Usersquo tab of the spreadsheet Basically the user determines which side ofthe photograph is top and then measures the angle between the line from PP to thetop of the photo and from PP to the auxiliary point on the photo ( gure 9)
If the user includes data for an auxiliary point a series of computations arecompleted to resolve the photo rotation ambiguity about the optical axis (ie the
+Z axis in R-Photo) A vector from the Auxiliary Point on the Earth (AE) throughthe photograph perspective centre ( located at SC) is intersected with the photo imageplane (XndashY plane of R-Photo) to compute the coordinates of the Auxiliary Point onthe photo (AP) in R-Photo A two-dimensional angle in the XndashY plane of R-Photofrom the shy X axis to a line from PP to AP is calculated as shown in gure 9 The
shy X axis is used as the origin of the angle since it represents the top of the photoonce it passes through the perspective centre The diVerence between the computedangle and the angle measured by the user on the photo resolves the ambiguity inthe rotation of the photo relative to the principal line ( gures 7 and 9) Thetransformations from R-Spacecraft and R-Photo are then modi ed to include anadditional rotation angle about the +Z axis in R-Photo
535 lsquoFootprint rsquo calculationsThe program next computes the coordinates of eight points about the perimeter
of the image format (ie located at the four photo corners plus a bisector pointalong each edge of the image) These points are identi ed in R-Photo based uponthe photograph format size and then converted to R-Spacecraft Since all computa-tions are done in orthogonal coordinate systems the R-Spacecraft to R-Photo
J A Robinson et al4428
rotation matrix is transposed to produce an R-Photo to R-Spacecraft rotation matrixOnce in R-Spacecraft a unit vector from each of the eight perimeter points throughthe photo perspective centre (the same point as SC) is computed This provides thecoordinates for points about the perimeter of the image format with their directionvectors in a common coordinate system with other key points needed to computethe photo footprint
The next step is to compute the point of intersection between the spherical earthmodel and each of the eight perimeter point vectors The scalar value for eachperimeter point unit vector is computed using two-dimensional planar trigonometryAn angle c is computed using the formula for the cosine of an angle between twoincluded vectors (the perimeter point unit vector and the vector from SC to thecentre of the Earth) Angle y is computed using the Law of Sines Angle e=180degrees shy c shy y The scalar value of the perimeter point vector is computed using eand the Law of Cosines The scalar value is then multiplied by the perimeter pointunit vector to produce the three-dimensional point of intersection of the vector withEarthrsquos surface in R-Spacecraft The process is repeated independently for each ofthe eight perimeter point vectors Aside from its mathematical simplicity the valueof arcsine (computed in step 2) will exceed the normal range for the sin of an anglewhen the perimeter point vectors fail to intersect with the surface of the earth Asimple test based on this principle allows the program to correctly handle obliquephotos which image a portion of the horizon (see results for high oblique photographsin table 5)
The nal step in the process converts the eight earth intersection points from theR-Spacecraft to R-Earth The results are then converted to the standard geographiccoordinate system and displayed on the lsquoInput-Outputrsquo page of the spreadsheet
54 ExamplesAll three formulations were applied to the photographs included in this paper
and the results are compared in table 5 Formulation 1 gives too small a value fordistance across the photograph (D) for all but the most nadir shots and thus servesas an indicator of the best theoretical case but is not a good measure for a speci cphotograph For example the photograph of Lake Eyre taken from 276 km altitude( gure 2(a)) and the photograph of Limmen Bight ( gure 7) were closest to beingnadir views (oVsetslt68 km or tlt15deg table 5) For these photographs D calculatedusing Formulation 1 was similar to the minimum D calculated using Formulations2 and 3 For almost all other more oblique photographs Formulation 1 gave asigni cant underestimate of the distance covered in the photograph For gure 5(A)(the picture of Houston taken with a 40 mm lens) Formulation 1 did not give anunderestimate for D This is because Formulation 1 does not account for curvatureof the Earth in any way With this large eld of view assuming a at Earth in atedthe value of D above the minimum from calculations that included Earth curvature
A major diVerence between Formulations 2 and 3 is the ability to estimate pixelsizes (P) in both directions (along the principal line and perpendicular to the principalline) For the more oblique photographs the vertical estimate of D and pixel sizes ismuch larger than in the horizontal direction (eg the low oblique and high obliquephotographs of Hawaii gure 4 table 5)
For the area of Limmen Bight ( gure 7) table 5 illustrates the improvement inthe estimate of distance and pixel size that can be obtained by re-estimating thelocation of the PC with greater accuracy Centre points in the catalogued data are
Astronaut photography as digital data for remote sensing 4429
plusmn05deg of latitude and longitude When the centre point was re-estimated to plusmn002degwe determined that the photograph was not taken as obliquely as rst thought(change in the estimate of the look angle t from 169 to 128deg table 5) When theauxiliary point was added to the calculations of Formula 3 the calculated look angleshrank further to 110deg indicating that this photograph was taken at very close toa nadir view Of course this improvement in accuracy could have also led to estimatesof greater obliquity and corresponding larger pixel sizes
We also tested the performance of the scale calculator with auxiliary point byestimating the corner and point locations on the photograph using a 11 000 000Operational Navigational Chart For this test we estimated our ability to readcoordinates from the map as plusmn002degand our error in nding the locations of thecorner points as plusmn015deg (this error varies among photographs depending on thedetail that can be matched between photo and map) For Limmen Bight ( gure 7)the mean diVerence between map estimates and calculator estimates for four pointswas 031deg (SD=018 n=8) For a photograph of San Francisco Bay (STS062-151-291) the mean diVerence between map estimates and calculator estimates forfour points was 0064deg (SD=018 n=8) For a photograph of San Francisco Bay(STS062-151-291) the mean diVerence between map estimates and calculator estim-ates for eight points was 0196deg (SD=0146 n=16) Thus in one case the calculatorestimates were better than our estimate of the error in locating corner points on themap It is reasonable to expect that for nadir-viewing photographs the calculatorused with an auxiliary point can estimate locations of the edges of a photograph towithin plusmn03deg
55 Empirical con rmation of spatial resolution estimatesAs stated previously a challenge to estimating system-AWAR for astronaut
photography of Earth is the lack of suitable targets Small features in an image cansometimes be used as a check on the size of objects that can be successfully resolvedgiving an approximate value for GRD Similarly the number of pixels that make upthose features in the digitized image can be used to make an independent calculationof pixel size We have successfully used features such as roads and airport runwaysto make estimates of spatial scale and resolution (eg Robinson et al 2000c) Whilerecognizing that the use of linear detail in an image is a poor approximation to abar target and that linear objects smaller than the resolving power can often bedetected (Charman 1965) few objects other than roads could be found to make anydirect estimates of GRD Thus roads and runways were used in the images ofHouston (where one can readily conduct ground veri cations and where a numberof higher-contrast concrete roadways were available) to obtain empirical estimatesof GRD and pixel size for comparison with table 5
In the all-digital ESC image of Houston ( gure 3(d )) we examined EllingtonField runway 4-22 (centre left of the image) which is 24384 mtimes457 m This runwayis approximately 6ndash7 pixels in width and 304ndash3094 pixels in length so pixelsrepresent an distance on the ground 7ndash8 m Using a lower contrast measure of astreet length between two intersections (2125 m=211 pixels) a pixel width of 101 mis estimated These results compare favourably with the minimum estimate of 81 mpixels using Formulation 1 (table 5) For an estimate of GRD that would be morecomparable to aerial photography of a line target the smallest street without treecover that could be clearly distinguished on the photograph was 792 m wide Thesmallest non-street object (a gap between stages of the Saturn rocket on display in
J A Robinson et al4430
Tab
le5
App
lica
tio
nof
met
hod
sfo
res
tim
ati
ng
spati
alre
solu
tio
no
fast
ron
aut
ph
oto
gra
ph
sin
clu
din
gF
orm
ula
tion
s12
and
3Im
pro
ved
cen
tre
po
int
esti
mat
esan
dauxilia
ryp
oin
tca
lcu
lati
ons
are
sho
wn
for
gure
7V
aria
ble
sas
use
din
the
text
fle
ns
foca
lle
ngt
hH
al
titu
de
PC
ph
oto
gra
ph
cen
tre
poin
tSN
sp
ace
craft
nadir
po
int
D
dis
tance
cover
edo
nth
egr
oun
d
Pd
ista
nce
on
the
grou
nd
repre
sen
ted
by
apix
el
OVse
td
ista
nce
bet
wee
nP
Cand
SNusi
ng
the
grea
tci
rcle
form
ula
t
look
angle
away
from
SN
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
2L
ake
Eyre
ST
S093
-717
-80
250
276
shy29
0137
0shy
28
413
71
607
11
767
413
7609
ndash64
212
013
760
9ndash64
7121
124
ST
S082
-754
-61
250
617
shy28
5137
0shy
28
113
50
135
726
1201
18
013
8ndash14
827
517
9138ndash15
3277
294
Fig
ure
3H
ou
sto
nS
TS
062
-97-
151
4029
829
5shy
95
530
5shy
95
4410
78
8112
20
5348ndash58
990
1205
351
ndash63
8951
10
36
ST
S094
-737
-28
100
294
295
shy
95
528
3shy
95
5162
31
1133
24
415
8ndash20
334
724
3158ndash20
7351
390
ST
S055
-71-
41250
302
295
shy
95
028
2shy
93
766
412
8192
32
5736
ndash84
715
232
273
9ndash85
7154
185
S73
E51
45300
2685
295
shy
95
024
78
0F
igu
re4
Haw
aii
ST
S069
-701
-65
250
370
195
shy
155
520
4shy
153
881
415
7204
28
9876
ndash98
917
928
688
0ndash10
9181
210
ST
S069
-701
-70
250
370
210
shy
157
519
2shy
150
781
415
7738
63
414
9ndash23
336
760
7148ndash55
6394
10
63
ST
S069
-729
-27
100
398
200
shy
156
620
5shy
148
2219
42
1867
65
432
8ndash13
10
158
62
4323+
311
N
A
Astronaut photography as digital data for remote sensing 4431
Tab
le5
(Cont
inued
)
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
7L
imm
enB
igh
tS
TS
093
-704
-50
250
283
shy15
0135
0shy
15
013
58
623
120
859
16
963
0ndash67
3125
16
863
0ndash67
612
613
2P
CR
e-es
tim
ated
shy14
733
1352
67
64
5128
623
ndash65
5123
128
623
ndash65
712
3127
Auxilia
ryP
oin
t611
062
3ndash65
112
312
5F
igu
re10
N
ear
Au
stin
T
XS
TS
095
-716
-46
250
543
30
5shy
975
28
6shy
1007
120
230
375
34
613
5ndash157
281
34
1136ndash16
128
636
0
1 All
pho
togr
aphs
exce
pt
on
eta
ken
wit
hH
ass
elbla
dca
mer
aso
dis
tan
ceacr
oss
the
image
d=
55m
m
for
S73
E51
45
taken
wit
hel
ectr
onic
still
cam
erad=
276
5m
mho
rizo
nta
l2 P
Cand
SN
are
giv
enin
the
form
(lati
tud
elo
ngit
ude)
deg
rees
w
ith
neg
ativ
en
um
ber
sre
pre
senti
ng
south
and
wes
tA
ccura
cyo
fP
Cis
plusmn0
5and
SN
isplusmn
01
as
reco
rded
inth
eA
stro
naut
Ph
oto
gra
phy
Data
base
ex
cep
tw
her
en
ote
dth
atce
ntr
ep
oin
tw
asre
-est
imate
dw
ith
grea
ter
accu
racy
3 C
alc
ula
ted
usi
ng
the
mea
nofth
em
inim
um
an
dm
axim
um
esti
mate
sof
DF
orm
ula
tion
2co
nsi
der
sD
inth
ed
irec
tion
per
pen
dic
ula
rto
the
Pri
nci
pal
Lin
eo
nly
4 C
alc
ula
ted
inho
rizo
nta
lan
dve
rtic
aldir
ecti
on
sb
ut
still
ass
um
ing
geom
etry
of
gure
6(B
)F
irst
nu
mb
eris
the
mea
no
fth
em
inim
um
Dat
the
bott
om
of
the
ph
oto
and
the
maxi
mum
Dat
the
top
of
the
ph
oto
(range
show
nin
the
tab
le)
and
isco
mpara
ble
toF
orm
ula
tio
n2
Sec
ond
num
ber
isth
em
ean
of
Des
tim
ated
alon
gth
eP
rin
cip
alline
and
alo
ng
the
ou
tsid
eed
ge
(range
not
show
n)
5 Alt
itu
de
isap
pro
xim
ate
for
mis
sion
as
exac
tti
me
pho
togr
ap
hw
asta
ken
and
corr
esp
on
din
gn
adir
data
are
no
tava
ilab
le
Lack
of
nad
irdata
pre
ven
tsap
plica
tion
of
Form
ula
tion
s2
an
d3
6 Au
xilliar
ypo
int
add
edw
as
shy14
80
lati
tud
e13
51
2lo
ngit
ude
shy46
5degro
tati
on
angle
in
stru
ctio
ns
for
esti
mati
ng
the
rota
tio
nan
gle
para
met
erare
conta
ined
as
par
tof
the
scal
eca
lcu
lato
rsp
read
shee
tdo
cum
enta
tio
n
J A Robinson et al4432
a park at Johnson Space Center) that could clearly be distinguished on the photo-graph was 853 m wide
For the photograph of Houston taken with a 250 mm lens ( gure 3(C)) anddigitized from second generation lm at 2400 ppi (106 mm pixel Otilde 1 ) Ellington Fieldrunway 4-22 is 3 pixels in width and 1613 pixels in length so pixels represent151ndash152 m on the ground These results compare favourably with the minimumestimate of 152 m using Formulation 2 and 154ndash185 m pixels using Formulation 3(table 5) For an estimate of GRD using an 8times8 inch print (1369 enlargement) and4times magni cation the smallest street that could clearly be distinguished was 822 mwide the same feature could barely be distinguished on the digitized image
We also made an empirical estimate of spatial resolution for lower contrastvegetation boundaries By clearing forest so that a pattern would be visible to landingaircraft a landowner outside Austin Texas (see also aerial photo in Lisheron 2000)created a target that is also useful for evaluating spatial resolution of astronautphotographs The forest was selectively cleared in order to spell the landownerrsquosname lsquoLUECKErsquo with the remaining trees ( gure 10) According to local surveyorswho planned the clearing the plan was to create letters that were 3100 fttimes1700 ft(9449 mtimes5182 m) Photographed at a high altitude relative to most Shuttle missions(543 km) with a 250 mm lens Formula 3 predicts that each pixel would represent anarea 286 mtimes360 m on the ground (table 5) When original lm was digitized at2400 ppi (106 mm pixel Otilde 1 ) letters correspond to 294times188 pixels for a comparablepixel size of 27ndash32 m
6 Summary and conclusionsAstronaut photographs can be an excellent source of data for remote sensing
applications Best-case resolutions are similar to that for Landsat or SPOT withpixels as small as lt10 m It was estimated that of images taken to date 504 hadlens and obliquity characteristics that make them potential candidates for remotesensing information Digitized astronaut photographs can be overlaid with othersatellite data using GIS (Eckardt et al 2000) or used to ll in gaps in time serieswhen other imagery is not available As a source of public-domain information theycan be very useful for scientists who do not have access to satellite imagery eitherbecause of the expense of image acquisition (to the end user) or the computersystems needed for processing satellite images These diVerences in image acquisitioncosts (to the user) are summarized in table 6
Searching of the complete database of NASA astronaut photography includinglow-spatial-resolution browse images is available via the Web (OYce of EarthSciences 2000) This provides nearly global access for identifying images that willcontribute to a speci c scienti c project For the most detailed studies digitalproducts posted to the web will not be of suYcient quality To date we have madeit a practice of digitizing small numbers of images when requested by scientists atno charge Such requests (including a description of the project involved) can bemade through the authors or using the contact information listed on the websiteInvestigators with needs for larger numbers of digital images have also been servedthrough collaborative agreements
In our experience scientists that nd the data most valuable are those in develop-ing countries those studying areas that have not usually been targets for the majorsatellites those having diYculty in nding low-cloud images those interested inconstructing time series or those interested in using a large number of images
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
J A Robinson et al4424
53 Formulation 3 T he L ow Oblique Space Photo Footprint CalculatorThe Low Oblique Space Photo Footprint Calculator was developed to provide
a more accurate estimation of the geographic coordinates for the footprint of a lowoblique photo of the Earthrsquos surface taken from a human-occupied spacecraft inorbit The calculator performs a series of three-dimensional coordinate transforma-tions to compute the location and orientation of the centre of the photo exposureplane relative to an earth referenced coordinate system The nominal camera focallength is then used to create a vector from the photorsquos perspective point througheach of eight points around the parameter of the image as de ned by the formatsize The geographical coordinates for the photo footprint are then computed byintersecting these photo vectors with a spherical earth model Although more sophist-icated projection algorithms are available no signi cant increase in the accuracy ofthe results would be produced by these algorithms due to inherent uncertainties inthe available input data (ie the spacecraft altitude photo centre location etc)
The calculations were initially implemented within a Microsoft Excel workbookwhich allowed one to embed the mathematical and graphical documentation nextto the actual calculations Thus interested users can inspect the mathematical pro-cessing A set of error traps was also built into the calculations to detect erroneousresults A summary of any errors generated is reported to the user with the resultsof the calculations Interested individuals are invited to download the Excel work-book from httpeoljscnasagovsseopFootprintCalculatorxls The calculations arecurrently being encoded in a high-level programming language and should soon beavailable alongside other background data provided for each photograph at OYceof Earth Sciences (2000)
For the purposes of these calculations a low oblique photograph was de ned as
one with the centre within 10 degrees of latitude and longitude of the spacecraftnadir point For the typical range of spacecraft altitudes to date this restricted thecalculations to photographs in which Earthrsquos horizon does not appear (the generalde nition of a low oblique photograph eg Campbell 199671 and gure 4)
531 Input data and resultsUpon opening the Excel workbook the user is presented with a program intro-
duction providing instructions for using the calculator The second worksheet tablsquoHow-To-Usersquo provides detailed step-by-step instructions for preparing the baselinedata for the program The third tab lsquoInput-Output rsquo contains the user input eldsand displays the results of the calculations The additional worksheets contain theactual calculations and program documentation Although users are welcome toreview these sheets an experienced user need only access the lsquoInput-Outputrsquo
spreadsheetTo begin a calculation the user enters the following information which is available
for each photo in the NASA Astronaut Photography Database (1) SN geographicalcoordinates of spacecraft nadir position at the time of photo (2) H spacecraftaltitude (3) PC the geographical coordinates of the centre of the photo (4) f nominal focal length and (5) d image format size The automatic implementationof the workbook on the web will automatically enter these values and completecalculations
For more accurate results the user may optionally enter the geographic co-ordinates and orientation for an auxiliary point on the photo which resolves thecamerarsquos rotation uncertainty about the optical axis The auxiliary point data must
Astronaut photography as digital data for remote sensing 4425
be computed by the user following the instruction contained in the lsquoHow-To-Usersquotab of the spreadsheet
After entering the input data the geographic coordinates of the photo footprint(ie four photo corner points four points at the bisector of each edge and the centreof the photo) are immediately displayed below the input elds along with any errormessages generated by the user input or by the calculations ( gure 7) Although
results are computed and displayed they should not be used when error messagesare produced by the program The program also computes the tilt angle for each ofthe photo vectors relative to the spacecraft nadir vector To the right of the photofootprint coordinates is displayed the arc distance along the surface of the sphere
between adjacent computed points
532 Calculation assumptions
The mathematical calculations implemented in the Low Oblique Space PhotoFootprint Calculator use the following assumptions
1 The SN location is used as exact even though the true value may vary by upto plusmn01 degree from the location provided with the photo
2 The spacecraft altitude is used as exact Although the determination of the
nadir point at the instant of a known spacecraft vector is relatively precise(plusmn115times10 Otilde 4 degrees) the propagator interpolates between sets of approxi-mately 10ndash40 known vectors per day and the time code recorded on the lmcan drift Thus the true value for SN may vary by up to plusmn01 degree fromthe value provided with the photo
3 The perspective centre of the camera is assumed to be at the given altitudeover the speci ed spacecraft nadir location at the time of photo exposure
4 The PC location is used as exact even though the true value may vary by upto plusmn05deg latitude and plusmn05deg longitude from the location provided with the
photo5 A spherical earth model is used with a nominal radius of 6 372 16154 m (a
common rst-order approximation for a spherical earth used in geodeticcomputations)
6 The nominal lens focal length of the camera lens is used in the computations(calibrated focal length values are not available)
7 The photo projection is based on the classic pin-hole camera model8 No correction for lens distortion or atmospheric re ection is made
9 If no auxiliary point data is provided the lsquoTop of the Imagersquo is orientedperpendicular to the vector from SN towards PC
533 T ransformation from Earth to photo coordinate systemsThe calculations begin by converting the geographic coordinates ( latitude and
longitude) of the SN and PC to a Rectangular Earth-Centred Coordinate System(R-Earth) de ned as shown in gure 9 (with the centre of the Earth at (0 0 0))
Using the vector from the Earthrsquos centre through SN and the spacecraft altitude thespacecraft location (SC) is also computed in R-Earth
For ease of computation a Rectangular Spacecraft-Centred Coordinate System(R-Spacecraft ) is de ned as shown in gure 9 The origin of R-Spacecraft is located
at SC with its +Z-axis aligned with vector from the centre of the Earth throughSN and its +X-axis aligned with the vector from SN to PC ( gure 9) The speci c
J A Robinson et al4426
Figure 9 Representation of the area included in a photograph as a geometric projectiononto the Earthrsquos surface Note that the focal length (the distance from the SpaceShuttle to the photo) is grossly overscale compared to the altitude (the distance fromthe Shuttle to the surface of the Earth
rotations and translation used to convert from the R-Earth to the R-Spacecraft arecomputed and documented in the spreadsheet
With the mathematical positions of the SC SN PC and the centre of the Earthcomputed in R-Spacecraft the program next computes the location of the camerarsquosprincipal point (PP) The principle point is the point of intersection of the opticalaxis of the lens with the image plane (ie the lm) It is nominally positioned at a
Astronaut photography as digital data for remote sensing 4427
distance equal to the focal length from the perspective centre of the camera (whichis assumed to be at SC) along the vector from PC through SC as shown in gure 9
A third coordinate system the Rectangular Photo Coordinate System (R-Photo)is created with its origin at PP its XndashY axial plane normal to the vector from PCthrough SC and its +X axis aligned with the +X axis of R-Spacecraft as shownin gure 9 The XndashY plane of this coordinate system represents the image plane ofthe photograph
534 Auxiliary point calculationsThe calculations above employ a critical assumption that all photos are taken
with the lsquotop of the imagersquo oriented perpendicular to the vector from the SN towardsPC as shown in gure 9 To avoid non-uniform solar heating of the external surfacemost orbiting spacecraft are slowly and continually rotated about one or more axesIn this condition a ight crew member taking a photo while oating in microgravitycould orient the photo with practically any orientation relative to the horizon (seealso gure 8(b)) Unfortunately since these photos are taken with conventionalhandheld cameras there is no other information available which can be used toresolve the photorsquos rotational ambiguity about the optical axis other then the photoitself This is why the above assumption is used and the footprint computed by thiscalculator is actually a lsquolocator ellipsersquo which estimates the area on the ground whatis likely to be included in a speci c photograph (see gure 6) This locator ellipse ismost accurate for square image formats and subject to additional distortion as thephotograph format becomes more rectangular
If the user wants a more precise calculation of the image footprint the photorsquosrotational ambiguity about the optical axis must be resolved This can be done inthe calculator by adding data for an auxiliary point Detailed instructions regardinghow to prepare and use auxiliary point data in the computations are included in thelsquoHow-To-Usersquo tab of the spreadsheet Basically the user determines which side ofthe photograph is top and then measures the angle between the line from PP to thetop of the photo and from PP to the auxiliary point on the photo ( gure 9)
If the user includes data for an auxiliary point a series of computations arecompleted to resolve the photo rotation ambiguity about the optical axis (ie the
+Z axis in R-Photo) A vector from the Auxiliary Point on the Earth (AE) throughthe photograph perspective centre ( located at SC) is intersected with the photo imageplane (XndashY plane of R-Photo) to compute the coordinates of the Auxiliary Point onthe photo (AP) in R-Photo A two-dimensional angle in the XndashY plane of R-Photofrom the shy X axis to a line from PP to AP is calculated as shown in gure 9 The
shy X axis is used as the origin of the angle since it represents the top of the photoonce it passes through the perspective centre The diVerence between the computedangle and the angle measured by the user on the photo resolves the ambiguity inthe rotation of the photo relative to the principal line ( gures 7 and 9) Thetransformations from R-Spacecraft and R-Photo are then modi ed to include anadditional rotation angle about the +Z axis in R-Photo
535 lsquoFootprint rsquo calculationsThe program next computes the coordinates of eight points about the perimeter
of the image format (ie located at the four photo corners plus a bisector pointalong each edge of the image) These points are identi ed in R-Photo based uponthe photograph format size and then converted to R-Spacecraft Since all computa-tions are done in orthogonal coordinate systems the R-Spacecraft to R-Photo
J A Robinson et al4428
rotation matrix is transposed to produce an R-Photo to R-Spacecraft rotation matrixOnce in R-Spacecraft a unit vector from each of the eight perimeter points throughthe photo perspective centre (the same point as SC) is computed This provides thecoordinates for points about the perimeter of the image format with their directionvectors in a common coordinate system with other key points needed to computethe photo footprint
The next step is to compute the point of intersection between the spherical earthmodel and each of the eight perimeter point vectors The scalar value for eachperimeter point unit vector is computed using two-dimensional planar trigonometryAn angle c is computed using the formula for the cosine of an angle between twoincluded vectors (the perimeter point unit vector and the vector from SC to thecentre of the Earth) Angle y is computed using the Law of Sines Angle e=180degrees shy c shy y The scalar value of the perimeter point vector is computed using eand the Law of Cosines The scalar value is then multiplied by the perimeter pointunit vector to produce the three-dimensional point of intersection of the vector withEarthrsquos surface in R-Spacecraft The process is repeated independently for each ofthe eight perimeter point vectors Aside from its mathematical simplicity the valueof arcsine (computed in step 2) will exceed the normal range for the sin of an anglewhen the perimeter point vectors fail to intersect with the surface of the earth Asimple test based on this principle allows the program to correctly handle obliquephotos which image a portion of the horizon (see results for high oblique photographsin table 5)
The nal step in the process converts the eight earth intersection points from theR-Spacecraft to R-Earth The results are then converted to the standard geographiccoordinate system and displayed on the lsquoInput-Outputrsquo page of the spreadsheet
54 ExamplesAll three formulations were applied to the photographs included in this paper
and the results are compared in table 5 Formulation 1 gives too small a value fordistance across the photograph (D) for all but the most nadir shots and thus servesas an indicator of the best theoretical case but is not a good measure for a speci cphotograph For example the photograph of Lake Eyre taken from 276 km altitude( gure 2(a)) and the photograph of Limmen Bight ( gure 7) were closest to beingnadir views (oVsetslt68 km or tlt15deg table 5) For these photographs D calculatedusing Formulation 1 was similar to the minimum D calculated using Formulations2 and 3 For almost all other more oblique photographs Formulation 1 gave asigni cant underestimate of the distance covered in the photograph For gure 5(A)(the picture of Houston taken with a 40 mm lens) Formulation 1 did not give anunderestimate for D This is because Formulation 1 does not account for curvatureof the Earth in any way With this large eld of view assuming a at Earth in atedthe value of D above the minimum from calculations that included Earth curvature
A major diVerence between Formulations 2 and 3 is the ability to estimate pixelsizes (P) in both directions (along the principal line and perpendicular to the principalline) For the more oblique photographs the vertical estimate of D and pixel sizes ismuch larger than in the horizontal direction (eg the low oblique and high obliquephotographs of Hawaii gure 4 table 5)
For the area of Limmen Bight ( gure 7) table 5 illustrates the improvement inthe estimate of distance and pixel size that can be obtained by re-estimating thelocation of the PC with greater accuracy Centre points in the catalogued data are
Astronaut photography as digital data for remote sensing 4429
plusmn05deg of latitude and longitude When the centre point was re-estimated to plusmn002degwe determined that the photograph was not taken as obliquely as rst thought(change in the estimate of the look angle t from 169 to 128deg table 5) When theauxiliary point was added to the calculations of Formula 3 the calculated look angleshrank further to 110deg indicating that this photograph was taken at very close toa nadir view Of course this improvement in accuracy could have also led to estimatesof greater obliquity and corresponding larger pixel sizes
We also tested the performance of the scale calculator with auxiliary point byestimating the corner and point locations on the photograph using a 11 000 000Operational Navigational Chart For this test we estimated our ability to readcoordinates from the map as plusmn002degand our error in nding the locations of thecorner points as plusmn015deg (this error varies among photographs depending on thedetail that can be matched between photo and map) For Limmen Bight ( gure 7)the mean diVerence between map estimates and calculator estimates for four pointswas 031deg (SD=018 n=8) For a photograph of San Francisco Bay (STS062-151-291) the mean diVerence between map estimates and calculator estimates forfour points was 0064deg (SD=018 n=8) For a photograph of San Francisco Bay(STS062-151-291) the mean diVerence between map estimates and calculator estim-ates for eight points was 0196deg (SD=0146 n=16) Thus in one case the calculatorestimates were better than our estimate of the error in locating corner points on themap It is reasonable to expect that for nadir-viewing photographs the calculatorused with an auxiliary point can estimate locations of the edges of a photograph towithin plusmn03deg
55 Empirical con rmation of spatial resolution estimatesAs stated previously a challenge to estimating system-AWAR for astronaut
photography of Earth is the lack of suitable targets Small features in an image cansometimes be used as a check on the size of objects that can be successfully resolvedgiving an approximate value for GRD Similarly the number of pixels that make upthose features in the digitized image can be used to make an independent calculationof pixel size We have successfully used features such as roads and airport runwaysto make estimates of spatial scale and resolution (eg Robinson et al 2000c) Whilerecognizing that the use of linear detail in an image is a poor approximation to abar target and that linear objects smaller than the resolving power can often bedetected (Charman 1965) few objects other than roads could be found to make anydirect estimates of GRD Thus roads and runways were used in the images ofHouston (where one can readily conduct ground veri cations and where a numberof higher-contrast concrete roadways were available) to obtain empirical estimatesof GRD and pixel size for comparison with table 5
In the all-digital ESC image of Houston ( gure 3(d )) we examined EllingtonField runway 4-22 (centre left of the image) which is 24384 mtimes457 m This runwayis approximately 6ndash7 pixels in width and 304ndash3094 pixels in length so pixelsrepresent an distance on the ground 7ndash8 m Using a lower contrast measure of astreet length between two intersections (2125 m=211 pixels) a pixel width of 101 mis estimated These results compare favourably with the minimum estimate of 81 mpixels using Formulation 1 (table 5) For an estimate of GRD that would be morecomparable to aerial photography of a line target the smallest street without treecover that could be clearly distinguished on the photograph was 792 m wide Thesmallest non-street object (a gap between stages of the Saturn rocket on display in
J A Robinson et al4430
Tab
le5
App
lica
tio
nof
met
hod
sfo
res
tim
ati
ng
spati
alre
solu
tio
no
fast
ron
aut
ph
oto
gra
ph
sin
clu
din
gF
orm
ula
tion
s12
and
3Im
pro
ved
cen
tre
po
int
esti
mat
esan
dauxilia
ryp
oin
tca
lcu
lati
ons
are
sho
wn
for
gure
7V
aria
ble
sas
use
din
the
text
fle
ns
foca
lle
ngt
hH
al
titu
de
PC
ph
oto
gra
ph
cen
tre
poin
tSN
sp
ace
craft
nadir
po
int
D
dis
tance
cover
edo
nth
egr
oun
d
Pd
ista
nce
on
the
grou
nd
repre
sen
ted
by
apix
el
OVse
td
ista
nce
bet
wee
nP
Cand
SNusi
ng
the
grea
tci
rcle
form
ula
t
look
angle
away
from
SN
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
2L
ake
Eyre
ST
S093
-717
-80
250
276
shy29
0137
0shy
28
413
71
607
11
767
413
7609
ndash64
212
013
760
9ndash64
7121
124
ST
S082
-754
-61
250
617
shy28
5137
0shy
28
113
50
135
726
1201
18
013
8ndash14
827
517
9138ndash15
3277
294
Fig
ure
3H
ou
sto
nS
TS
062
-97-
151
4029
829
5shy
95
530
5shy
95
4410
78
8112
20
5348ndash58
990
1205
351
ndash63
8951
10
36
ST
S094
-737
-28
100
294
295
shy
95
528
3shy
95
5162
31
1133
24
415
8ndash20
334
724
3158ndash20
7351
390
ST
S055
-71-
41250
302
295
shy
95
028
2shy
93
766
412
8192
32
5736
ndash84
715
232
273
9ndash85
7154
185
S73
E51
45300
2685
295
shy
95
024
78
0F
igu
re4
Haw
aii
ST
S069
-701
-65
250
370
195
shy
155
520
4shy
153
881
415
7204
28
9876
ndash98
917
928
688
0ndash10
9181
210
ST
S069
-701
-70
250
370
210
shy
157
519
2shy
150
781
415
7738
63
414
9ndash23
336
760
7148ndash55
6394
10
63
ST
S069
-729
-27
100
398
200
shy
156
620
5shy
148
2219
42
1867
65
432
8ndash13
10
158
62
4323+
311
N
A
Astronaut photography as digital data for remote sensing 4431
Tab
le5
(Cont
inued
)
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
7L
imm
enB
igh
tS
TS
093
-704
-50
250
283
shy15
0135
0shy
15
013
58
623
120
859
16
963
0ndash67
3125
16
863
0ndash67
612
613
2P
CR
e-es
tim
ated
shy14
733
1352
67
64
5128
623
ndash65
5123
128
623
ndash65
712
3127
Auxilia
ryP
oin
t611
062
3ndash65
112
312
5F
igu
re10
N
ear
Au
stin
T
XS
TS
095
-716
-46
250
543
30
5shy
975
28
6shy
1007
120
230
375
34
613
5ndash157
281
34
1136ndash16
128
636
0
1 All
pho
togr
aphs
exce
pt
on
eta
ken
wit
hH
ass
elbla
dca
mer
aso
dis
tan
ceacr
oss
the
image
d=
55m
m
for
S73
E51
45
taken
wit
hel
ectr
onic
still
cam
erad=
276
5m
mho
rizo
nta
l2 P
Cand
SN
are
giv
enin
the
form
(lati
tud
elo
ngit
ude)
deg
rees
w
ith
neg
ativ
en
um
ber
sre
pre
senti
ng
south
and
wes
tA
ccura
cyo
fP
Cis
plusmn0
5and
SN
isplusmn
01
as
reco
rded
inth
eA
stro
naut
Ph
oto
gra
phy
Data
base
ex
cep
tw
her
en
ote
dth
atce
ntr
ep
oin
tw
asre
-est
imate
dw
ith
grea
ter
accu
racy
3 C
alc
ula
ted
usi
ng
the
mea
nofth
em
inim
um
an
dm
axim
um
esti
mate
sof
DF
orm
ula
tion
2co
nsi
der
sD
inth
ed
irec
tion
per
pen
dic
ula
rto
the
Pri
nci
pal
Lin
eo
nly
4 C
alc
ula
ted
inho
rizo
nta
lan
dve
rtic
aldir
ecti
on
sb
ut
still
ass
um
ing
geom
etry
of
gure
6(B
)F
irst
nu
mb
eris
the
mea
no
fth
em
inim
um
Dat
the
bott
om
of
the
ph
oto
and
the
maxi
mum
Dat
the
top
of
the
ph
oto
(range
show
nin
the
tab
le)
and
isco
mpara
ble
toF
orm
ula
tio
n2
Sec
ond
num
ber
isth
em
ean
of
Des
tim
ated
alon
gth
eP
rin
cip
alline
and
alo
ng
the
ou
tsid
eed
ge
(range
not
show
n)
5 Alt
itu
de
isap
pro
xim
ate
for
mis
sion
as
exac
tti
me
pho
togr
ap
hw
asta
ken
and
corr
esp
on
din
gn
adir
data
are
no
tava
ilab
le
Lack
of
nad
irdata
pre
ven
tsap
plica
tion
of
Form
ula
tion
s2
an
d3
6 Au
xilliar
ypo
int
add
edw
as
shy14
80
lati
tud
e13
51
2lo
ngit
ude
shy46
5degro
tati
on
angle
in
stru
ctio
ns
for
esti
mati
ng
the
rota
tio
nan
gle
para
met
erare
conta
ined
as
par
tof
the
scal
eca
lcu
lato
rsp
read
shee
tdo
cum
enta
tio
n
J A Robinson et al4432
a park at Johnson Space Center) that could clearly be distinguished on the photo-graph was 853 m wide
For the photograph of Houston taken with a 250 mm lens ( gure 3(C)) anddigitized from second generation lm at 2400 ppi (106 mm pixel Otilde 1 ) Ellington Fieldrunway 4-22 is 3 pixels in width and 1613 pixels in length so pixels represent151ndash152 m on the ground These results compare favourably with the minimumestimate of 152 m using Formulation 2 and 154ndash185 m pixels using Formulation 3(table 5) For an estimate of GRD using an 8times8 inch print (1369 enlargement) and4times magni cation the smallest street that could clearly be distinguished was 822 mwide the same feature could barely be distinguished on the digitized image
We also made an empirical estimate of spatial resolution for lower contrastvegetation boundaries By clearing forest so that a pattern would be visible to landingaircraft a landowner outside Austin Texas (see also aerial photo in Lisheron 2000)created a target that is also useful for evaluating spatial resolution of astronautphotographs The forest was selectively cleared in order to spell the landownerrsquosname lsquoLUECKErsquo with the remaining trees ( gure 10) According to local surveyorswho planned the clearing the plan was to create letters that were 3100 fttimes1700 ft(9449 mtimes5182 m) Photographed at a high altitude relative to most Shuttle missions(543 km) with a 250 mm lens Formula 3 predicts that each pixel would represent anarea 286 mtimes360 m on the ground (table 5) When original lm was digitized at2400 ppi (106 mm pixel Otilde 1 ) letters correspond to 294times188 pixels for a comparablepixel size of 27ndash32 m
6 Summary and conclusionsAstronaut photographs can be an excellent source of data for remote sensing
applications Best-case resolutions are similar to that for Landsat or SPOT withpixels as small as lt10 m It was estimated that of images taken to date 504 hadlens and obliquity characteristics that make them potential candidates for remotesensing information Digitized astronaut photographs can be overlaid with othersatellite data using GIS (Eckardt et al 2000) or used to ll in gaps in time serieswhen other imagery is not available As a source of public-domain information theycan be very useful for scientists who do not have access to satellite imagery eitherbecause of the expense of image acquisition (to the end user) or the computersystems needed for processing satellite images These diVerences in image acquisitioncosts (to the user) are summarized in table 6
Searching of the complete database of NASA astronaut photography includinglow-spatial-resolution browse images is available via the Web (OYce of EarthSciences 2000) This provides nearly global access for identifying images that willcontribute to a speci c scienti c project For the most detailed studies digitalproducts posted to the web will not be of suYcient quality To date we have madeit a practice of digitizing small numbers of images when requested by scientists atno charge Such requests (including a description of the project involved) can bemade through the authors or using the contact information listed on the websiteInvestigators with needs for larger numbers of digital images have also been servedthrough collaborative agreements
In our experience scientists that nd the data most valuable are those in develop-ing countries those studying areas that have not usually been targets for the majorsatellites those having diYculty in nding low-cloud images those interested inconstructing time series or those interested in using a large number of images
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
Astronaut photography as digital data for remote sensing 4425
be computed by the user following the instruction contained in the lsquoHow-To-Usersquotab of the spreadsheet
After entering the input data the geographic coordinates of the photo footprint(ie four photo corner points four points at the bisector of each edge and the centreof the photo) are immediately displayed below the input elds along with any errormessages generated by the user input or by the calculations ( gure 7) Although
results are computed and displayed they should not be used when error messagesare produced by the program The program also computes the tilt angle for each ofthe photo vectors relative to the spacecraft nadir vector To the right of the photofootprint coordinates is displayed the arc distance along the surface of the sphere
between adjacent computed points
532 Calculation assumptions
The mathematical calculations implemented in the Low Oblique Space PhotoFootprint Calculator use the following assumptions
1 The SN location is used as exact even though the true value may vary by upto plusmn01 degree from the location provided with the photo
2 The spacecraft altitude is used as exact Although the determination of the
nadir point at the instant of a known spacecraft vector is relatively precise(plusmn115times10 Otilde 4 degrees) the propagator interpolates between sets of approxi-mately 10ndash40 known vectors per day and the time code recorded on the lmcan drift Thus the true value for SN may vary by up to plusmn01 degree fromthe value provided with the photo
3 The perspective centre of the camera is assumed to be at the given altitudeover the speci ed spacecraft nadir location at the time of photo exposure
4 The PC location is used as exact even though the true value may vary by upto plusmn05deg latitude and plusmn05deg longitude from the location provided with the
photo5 A spherical earth model is used with a nominal radius of 6 372 16154 m (a
common rst-order approximation for a spherical earth used in geodeticcomputations)
6 The nominal lens focal length of the camera lens is used in the computations(calibrated focal length values are not available)
7 The photo projection is based on the classic pin-hole camera model8 No correction for lens distortion or atmospheric re ection is made
9 If no auxiliary point data is provided the lsquoTop of the Imagersquo is orientedperpendicular to the vector from SN towards PC
533 T ransformation from Earth to photo coordinate systemsThe calculations begin by converting the geographic coordinates ( latitude and
longitude) of the SN and PC to a Rectangular Earth-Centred Coordinate System(R-Earth) de ned as shown in gure 9 (with the centre of the Earth at (0 0 0))
Using the vector from the Earthrsquos centre through SN and the spacecraft altitude thespacecraft location (SC) is also computed in R-Earth
For ease of computation a Rectangular Spacecraft-Centred Coordinate System(R-Spacecraft ) is de ned as shown in gure 9 The origin of R-Spacecraft is located
at SC with its +Z-axis aligned with vector from the centre of the Earth throughSN and its +X-axis aligned with the vector from SN to PC ( gure 9) The speci c
J A Robinson et al4426
Figure 9 Representation of the area included in a photograph as a geometric projectiononto the Earthrsquos surface Note that the focal length (the distance from the SpaceShuttle to the photo) is grossly overscale compared to the altitude (the distance fromthe Shuttle to the surface of the Earth
rotations and translation used to convert from the R-Earth to the R-Spacecraft arecomputed and documented in the spreadsheet
With the mathematical positions of the SC SN PC and the centre of the Earthcomputed in R-Spacecraft the program next computes the location of the camerarsquosprincipal point (PP) The principle point is the point of intersection of the opticalaxis of the lens with the image plane (ie the lm) It is nominally positioned at a
Astronaut photography as digital data for remote sensing 4427
distance equal to the focal length from the perspective centre of the camera (whichis assumed to be at SC) along the vector from PC through SC as shown in gure 9
A third coordinate system the Rectangular Photo Coordinate System (R-Photo)is created with its origin at PP its XndashY axial plane normal to the vector from PCthrough SC and its +X axis aligned with the +X axis of R-Spacecraft as shownin gure 9 The XndashY plane of this coordinate system represents the image plane ofthe photograph
534 Auxiliary point calculationsThe calculations above employ a critical assumption that all photos are taken
with the lsquotop of the imagersquo oriented perpendicular to the vector from the SN towardsPC as shown in gure 9 To avoid non-uniform solar heating of the external surfacemost orbiting spacecraft are slowly and continually rotated about one or more axesIn this condition a ight crew member taking a photo while oating in microgravitycould orient the photo with practically any orientation relative to the horizon (seealso gure 8(b)) Unfortunately since these photos are taken with conventionalhandheld cameras there is no other information available which can be used toresolve the photorsquos rotational ambiguity about the optical axis other then the photoitself This is why the above assumption is used and the footprint computed by thiscalculator is actually a lsquolocator ellipsersquo which estimates the area on the ground whatis likely to be included in a speci c photograph (see gure 6) This locator ellipse ismost accurate for square image formats and subject to additional distortion as thephotograph format becomes more rectangular
If the user wants a more precise calculation of the image footprint the photorsquosrotational ambiguity about the optical axis must be resolved This can be done inthe calculator by adding data for an auxiliary point Detailed instructions regardinghow to prepare and use auxiliary point data in the computations are included in thelsquoHow-To-Usersquo tab of the spreadsheet Basically the user determines which side ofthe photograph is top and then measures the angle between the line from PP to thetop of the photo and from PP to the auxiliary point on the photo ( gure 9)
If the user includes data for an auxiliary point a series of computations arecompleted to resolve the photo rotation ambiguity about the optical axis (ie the
+Z axis in R-Photo) A vector from the Auxiliary Point on the Earth (AE) throughthe photograph perspective centre ( located at SC) is intersected with the photo imageplane (XndashY plane of R-Photo) to compute the coordinates of the Auxiliary Point onthe photo (AP) in R-Photo A two-dimensional angle in the XndashY plane of R-Photofrom the shy X axis to a line from PP to AP is calculated as shown in gure 9 The
shy X axis is used as the origin of the angle since it represents the top of the photoonce it passes through the perspective centre The diVerence between the computedangle and the angle measured by the user on the photo resolves the ambiguity inthe rotation of the photo relative to the principal line ( gures 7 and 9) Thetransformations from R-Spacecraft and R-Photo are then modi ed to include anadditional rotation angle about the +Z axis in R-Photo
535 lsquoFootprint rsquo calculationsThe program next computes the coordinates of eight points about the perimeter
of the image format (ie located at the four photo corners plus a bisector pointalong each edge of the image) These points are identi ed in R-Photo based uponthe photograph format size and then converted to R-Spacecraft Since all computa-tions are done in orthogonal coordinate systems the R-Spacecraft to R-Photo
J A Robinson et al4428
rotation matrix is transposed to produce an R-Photo to R-Spacecraft rotation matrixOnce in R-Spacecraft a unit vector from each of the eight perimeter points throughthe photo perspective centre (the same point as SC) is computed This provides thecoordinates for points about the perimeter of the image format with their directionvectors in a common coordinate system with other key points needed to computethe photo footprint
The next step is to compute the point of intersection between the spherical earthmodel and each of the eight perimeter point vectors The scalar value for eachperimeter point unit vector is computed using two-dimensional planar trigonometryAn angle c is computed using the formula for the cosine of an angle between twoincluded vectors (the perimeter point unit vector and the vector from SC to thecentre of the Earth) Angle y is computed using the Law of Sines Angle e=180degrees shy c shy y The scalar value of the perimeter point vector is computed using eand the Law of Cosines The scalar value is then multiplied by the perimeter pointunit vector to produce the three-dimensional point of intersection of the vector withEarthrsquos surface in R-Spacecraft The process is repeated independently for each ofthe eight perimeter point vectors Aside from its mathematical simplicity the valueof arcsine (computed in step 2) will exceed the normal range for the sin of an anglewhen the perimeter point vectors fail to intersect with the surface of the earth Asimple test based on this principle allows the program to correctly handle obliquephotos which image a portion of the horizon (see results for high oblique photographsin table 5)
The nal step in the process converts the eight earth intersection points from theR-Spacecraft to R-Earth The results are then converted to the standard geographiccoordinate system and displayed on the lsquoInput-Outputrsquo page of the spreadsheet
54 ExamplesAll three formulations were applied to the photographs included in this paper
and the results are compared in table 5 Formulation 1 gives too small a value fordistance across the photograph (D) for all but the most nadir shots and thus servesas an indicator of the best theoretical case but is not a good measure for a speci cphotograph For example the photograph of Lake Eyre taken from 276 km altitude( gure 2(a)) and the photograph of Limmen Bight ( gure 7) were closest to beingnadir views (oVsetslt68 km or tlt15deg table 5) For these photographs D calculatedusing Formulation 1 was similar to the minimum D calculated using Formulations2 and 3 For almost all other more oblique photographs Formulation 1 gave asigni cant underestimate of the distance covered in the photograph For gure 5(A)(the picture of Houston taken with a 40 mm lens) Formulation 1 did not give anunderestimate for D This is because Formulation 1 does not account for curvatureof the Earth in any way With this large eld of view assuming a at Earth in atedthe value of D above the minimum from calculations that included Earth curvature
A major diVerence between Formulations 2 and 3 is the ability to estimate pixelsizes (P) in both directions (along the principal line and perpendicular to the principalline) For the more oblique photographs the vertical estimate of D and pixel sizes ismuch larger than in the horizontal direction (eg the low oblique and high obliquephotographs of Hawaii gure 4 table 5)
For the area of Limmen Bight ( gure 7) table 5 illustrates the improvement inthe estimate of distance and pixel size that can be obtained by re-estimating thelocation of the PC with greater accuracy Centre points in the catalogued data are
Astronaut photography as digital data for remote sensing 4429
plusmn05deg of latitude and longitude When the centre point was re-estimated to plusmn002degwe determined that the photograph was not taken as obliquely as rst thought(change in the estimate of the look angle t from 169 to 128deg table 5) When theauxiliary point was added to the calculations of Formula 3 the calculated look angleshrank further to 110deg indicating that this photograph was taken at very close toa nadir view Of course this improvement in accuracy could have also led to estimatesof greater obliquity and corresponding larger pixel sizes
We also tested the performance of the scale calculator with auxiliary point byestimating the corner and point locations on the photograph using a 11 000 000Operational Navigational Chart For this test we estimated our ability to readcoordinates from the map as plusmn002degand our error in nding the locations of thecorner points as plusmn015deg (this error varies among photographs depending on thedetail that can be matched between photo and map) For Limmen Bight ( gure 7)the mean diVerence between map estimates and calculator estimates for four pointswas 031deg (SD=018 n=8) For a photograph of San Francisco Bay (STS062-151-291) the mean diVerence between map estimates and calculator estimates forfour points was 0064deg (SD=018 n=8) For a photograph of San Francisco Bay(STS062-151-291) the mean diVerence between map estimates and calculator estim-ates for eight points was 0196deg (SD=0146 n=16) Thus in one case the calculatorestimates were better than our estimate of the error in locating corner points on themap It is reasonable to expect that for nadir-viewing photographs the calculatorused with an auxiliary point can estimate locations of the edges of a photograph towithin plusmn03deg
55 Empirical con rmation of spatial resolution estimatesAs stated previously a challenge to estimating system-AWAR for astronaut
photography of Earth is the lack of suitable targets Small features in an image cansometimes be used as a check on the size of objects that can be successfully resolvedgiving an approximate value for GRD Similarly the number of pixels that make upthose features in the digitized image can be used to make an independent calculationof pixel size We have successfully used features such as roads and airport runwaysto make estimates of spatial scale and resolution (eg Robinson et al 2000c) Whilerecognizing that the use of linear detail in an image is a poor approximation to abar target and that linear objects smaller than the resolving power can often bedetected (Charman 1965) few objects other than roads could be found to make anydirect estimates of GRD Thus roads and runways were used in the images ofHouston (where one can readily conduct ground veri cations and where a numberof higher-contrast concrete roadways were available) to obtain empirical estimatesof GRD and pixel size for comparison with table 5
In the all-digital ESC image of Houston ( gure 3(d )) we examined EllingtonField runway 4-22 (centre left of the image) which is 24384 mtimes457 m This runwayis approximately 6ndash7 pixels in width and 304ndash3094 pixels in length so pixelsrepresent an distance on the ground 7ndash8 m Using a lower contrast measure of astreet length between two intersections (2125 m=211 pixels) a pixel width of 101 mis estimated These results compare favourably with the minimum estimate of 81 mpixels using Formulation 1 (table 5) For an estimate of GRD that would be morecomparable to aerial photography of a line target the smallest street without treecover that could be clearly distinguished on the photograph was 792 m wide Thesmallest non-street object (a gap between stages of the Saturn rocket on display in
J A Robinson et al4430
Tab
le5
App
lica
tio
nof
met
hod
sfo
res
tim
ati
ng
spati
alre
solu
tio
no
fast
ron
aut
ph
oto
gra
ph
sin
clu
din
gF
orm
ula
tion
s12
and
3Im
pro
ved
cen
tre
po
int
esti
mat
esan
dauxilia
ryp
oin
tca
lcu
lati
ons
are
sho
wn
for
gure
7V
aria
ble
sas
use
din
the
text
fle
ns
foca
lle
ngt
hH
al
titu
de
PC
ph
oto
gra
ph
cen
tre
poin
tSN
sp
ace
craft
nadir
po
int
D
dis
tance
cover
edo
nth
egr
oun
d
Pd
ista
nce
on
the
grou
nd
repre
sen
ted
by
apix
el
OVse
td
ista
nce
bet
wee
nP
Cand
SNusi
ng
the
grea
tci
rcle
form
ula
t
look
angle
away
from
SN
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
2L
ake
Eyre
ST
S093
-717
-80
250
276
shy29
0137
0shy
28
413
71
607
11
767
413
7609
ndash64
212
013
760
9ndash64
7121
124
ST
S082
-754
-61
250
617
shy28
5137
0shy
28
113
50
135
726
1201
18
013
8ndash14
827
517
9138ndash15
3277
294
Fig
ure
3H
ou
sto
nS
TS
062
-97-
151
4029
829
5shy
95
530
5shy
95
4410
78
8112
20
5348ndash58
990
1205
351
ndash63
8951
10
36
ST
S094
-737
-28
100
294
295
shy
95
528
3shy
95
5162
31
1133
24
415
8ndash20
334
724
3158ndash20
7351
390
ST
S055
-71-
41250
302
295
shy
95
028
2shy
93
766
412
8192
32
5736
ndash84
715
232
273
9ndash85
7154
185
S73
E51
45300
2685
295
shy
95
024
78
0F
igu
re4
Haw
aii
ST
S069
-701
-65
250
370
195
shy
155
520
4shy
153
881
415
7204
28
9876
ndash98
917
928
688
0ndash10
9181
210
ST
S069
-701
-70
250
370
210
shy
157
519
2shy
150
781
415
7738
63
414
9ndash23
336
760
7148ndash55
6394
10
63
ST
S069
-729
-27
100
398
200
shy
156
620
5shy
148
2219
42
1867
65
432
8ndash13
10
158
62
4323+
311
N
A
Astronaut photography as digital data for remote sensing 4431
Tab
le5
(Cont
inued
)
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
7L
imm
enB
igh
tS
TS
093
-704
-50
250
283
shy15
0135
0shy
15
013
58
623
120
859
16
963
0ndash67
3125
16
863
0ndash67
612
613
2P
CR
e-es
tim
ated
shy14
733
1352
67
64
5128
623
ndash65
5123
128
623
ndash65
712
3127
Auxilia
ryP
oin
t611
062
3ndash65
112
312
5F
igu
re10
N
ear
Au
stin
T
XS
TS
095
-716
-46
250
543
30
5shy
975
28
6shy
1007
120
230
375
34
613
5ndash157
281
34
1136ndash16
128
636
0
1 All
pho
togr
aphs
exce
pt
on
eta
ken
wit
hH
ass
elbla
dca
mer
aso
dis
tan
ceacr
oss
the
image
d=
55m
m
for
S73
E51
45
taken
wit
hel
ectr
onic
still
cam
erad=
276
5m
mho
rizo
nta
l2 P
Cand
SN
are
giv
enin
the
form
(lati
tud
elo
ngit
ude)
deg
rees
w
ith
neg
ativ
en
um
ber
sre
pre
senti
ng
south
and
wes
tA
ccura
cyo
fP
Cis
plusmn0
5and
SN
isplusmn
01
as
reco
rded
inth
eA
stro
naut
Ph
oto
gra
phy
Data
base
ex
cep
tw
her
en
ote
dth
atce
ntr
ep
oin
tw
asre
-est
imate
dw
ith
grea
ter
accu
racy
3 C
alc
ula
ted
usi
ng
the
mea
nofth
em
inim
um
an
dm
axim
um
esti
mate
sof
DF
orm
ula
tion
2co
nsi
der
sD
inth
ed
irec
tion
per
pen
dic
ula
rto
the
Pri
nci
pal
Lin
eo
nly
4 C
alc
ula
ted
inho
rizo
nta
lan
dve
rtic
aldir
ecti
on
sb
ut
still
ass
um
ing
geom
etry
of
gure
6(B
)F
irst
nu
mb
eris
the
mea
no
fth
em
inim
um
Dat
the
bott
om
of
the
ph
oto
and
the
maxi
mum
Dat
the
top
of
the
ph
oto
(range
show
nin
the
tab
le)
and
isco
mpara
ble
toF
orm
ula
tio
n2
Sec
ond
num
ber
isth
em
ean
of
Des
tim
ated
alon
gth
eP
rin
cip
alline
and
alo
ng
the
ou
tsid
eed
ge
(range
not
show
n)
5 Alt
itu
de
isap
pro
xim
ate
for
mis
sion
as
exac
tti
me
pho
togr
ap
hw
asta
ken
and
corr
esp
on
din
gn
adir
data
are
no
tava
ilab
le
Lack
of
nad
irdata
pre
ven
tsap
plica
tion
of
Form
ula
tion
s2
an
d3
6 Au
xilliar
ypo
int
add
edw
as
shy14
80
lati
tud
e13
51
2lo
ngit
ude
shy46
5degro
tati
on
angle
in
stru
ctio
ns
for
esti
mati
ng
the
rota
tio
nan
gle
para
met
erare
conta
ined
as
par
tof
the
scal
eca
lcu
lato
rsp
read
shee
tdo
cum
enta
tio
n
J A Robinson et al4432
a park at Johnson Space Center) that could clearly be distinguished on the photo-graph was 853 m wide
For the photograph of Houston taken with a 250 mm lens ( gure 3(C)) anddigitized from second generation lm at 2400 ppi (106 mm pixel Otilde 1 ) Ellington Fieldrunway 4-22 is 3 pixels in width and 1613 pixels in length so pixels represent151ndash152 m on the ground These results compare favourably with the minimumestimate of 152 m using Formulation 2 and 154ndash185 m pixels using Formulation 3(table 5) For an estimate of GRD using an 8times8 inch print (1369 enlargement) and4times magni cation the smallest street that could clearly be distinguished was 822 mwide the same feature could barely be distinguished on the digitized image
We also made an empirical estimate of spatial resolution for lower contrastvegetation boundaries By clearing forest so that a pattern would be visible to landingaircraft a landowner outside Austin Texas (see also aerial photo in Lisheron 2000)created a target that is also useful for evaluating spatial resolution of astronautphotographs The forest was selectively cleared in order to spell the landownerrsquosname lsquoLUECKErsquo with the remaining trees ( gure 10) According to local surveyorswho planned the clearing the plan was to create letters that were 3100 fttimes1700 ft(9449 mtimes5182 m) Photographed at a high altitude relative to most Shuttle missions(543 km) with a 250 mm lens Formula 3 predicts that each pixel would represent anarea 286 mtimes360 m on the ground (table 5) When original lm was digitized at2400 ppi (106 mm pixel Otilde 1 ) letters correspond to 294times188 pixels for a comparablepixel size of 27ndash32 m
6 Summary and conclusionsAstronaut photographs can be an excellent source of data for remote sensing
applications Best-case resolutions are similar to that for Landsat or SPOT withpixels as small as lt10 m It was estimated that of images taken to date 504 hadlens and obliquity characteristics that make them potential candidates for remotesensing information Digitized astronaut photographs can be overlaid with othersatellite data using GIS (Eckardt et al 2000) or used to ll in gaps in time serieswhen other imagery is not available As a source of public-domain information theycan be very useful for scientists who do not have access to satellite imagery eitherbecause of the expense of image acquisition (to the end user) or the computersystems needed for processing satellite images These diVerences in image acquisitioncosts (to the user) are summarized in table 6
Searching of the complete database of NASA astronaut photography includinglow-spatial-resolution browse images is available via the Web (OYce of EarthSciences 2000) This provides nearly global access for identifying images that willcontribute to a speci c scienti c project For the most detailed studies digitalproducts posted to the web will not be of suYcient quality To date we have madeit a practice of digitizing small numbers of images when requested by scientists atno charge Such requests (including a description of the project involved) can bemade through the authors or using the contact information listed on the websiteInvestigators with needs for larger numbers of digital images have also been servedthrough collaborative agreements
In our experience scientists that nd the data most valuable are those in develop-ing countries those studying areas that have not usually been targets for the majorsatellites those having diYculty in nding low-cloud images those interested inconstructing time series or those interested in using a large number of images
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
J A Robinson et al4426
Figure 9 Representation of the area included in a photograph as a geometric projectiononto the Earthrsquos surface Note that the focal length (the distance from the SpaceShuttle to the photo) is grossly overscale compared to the altitude (the distance fromthe Shuttle to the surface of the Earth
rotations and translation used to convert from the R-Earth to the R-Spacecraft arecomputed and documented in the spreadsheet
With the mathematical positions of the SC SN PC and the centre of the Earthcomputed in R-Spacecraft the program next computes the location of the camerarsquosprincipal point (PP) The principle point is the point of intersection of the opticalaxis of the lens with the image plane (ie the lm) It is nominally positioned at a
Astronaut photography as digital data for remote sensing 4427
distance equal to the focal length from the perspective centre of the camera (whichis assumed to be at SC) along the vector from PC through SC as shown in gure 9
A third coordinate system the Rectangular Photo Coordinate System (R-Photo)is created with its origin at PP its XndashY axial plane normal to the vector from PCthrough SC and its +X axis aligned with the +X axis of R-Spacecraft as shownin gure 9 The XndashY plane of this coordinate system represents the image plane ofthe photograph
534 Auxiliary point calculationsThe calculations above employ a critical assumption that all photos are taken
with the lsquotop of the imagersquo oriented perpendicular to the vector from the SN towardsPC as shown in gure 9 To avoid non-uniform solar heating of the external surfacemost orbiting spacecraft are slowly and continually rotated about one or more axesIn this condition a ight crew member taking a photo while oating in microgravitycould orient the photo with practically any orientation relative to the horizon (seealso gure 8(b)) Unfortunately since these photos are taken with conventionalhandheld cameras there is no other information available which can be used toresolve the photorsquos rotational ambiguity about the optical axis other then the photoitself This is why the above assumption is used and the footprint computed by thiscalculator is actually a lsquolocator ellipsersquo which estimates the area on the ground whatis likely to be included in a speci c photograph (see gure 6) This locator ellipse ismost accurate for square image formats and subject to additional distortion as thephotograph format becomes more rectangular
If the user wants a more precise calculation of the image footprint the photorsquosrotational ambiguity about the optical axis must be resolved This can be done inthe calculator by adding data for an auxiliary point Detailed instructions regardinghow to prepare and use auxiliary point data in the computations are included in thelsquoHow-To-Usersquo tab of the spreadsheet Basically the user determines which side ofthe photograph is top and then measures the angle between the line from PP to thetop of the photo and from PP to the auxiliary point on the photo ( gure 9)
If the user includes data for an auxiliary point a series of computations arecompleted to resolve the photo rotation ambiguity about the optical axis (ie the
+Z axis in R-Photo) A vector from the Auxiliary Point on the Earth (AE) throughthe photograph perspective centre ( located at SC) is intersected with the photo imageplane (XndashY plane of R-Photo) to compute the coordinates of the Auxiliary Point onthe photo (AP) in R-Photo A two-dimensional angle in the XndashY plane of R-Photofrom the shy X axis to a line from PP to AP is calculated as shown in gure 9 The
shy X axis is used as the origin of the angle since it represents the top of the photoonce it passes through the perspective centre The diVerence between the computedangle and the angle measured by the user on the photo resolves the ambiguity inthe rotation of the photo relative to the principal line ( gures 7 and 9) Thetransformations from R-Spacecraft and R-Photo are then modi ed to include anadditional rotation angle about the +Z axis in R-Photo
535 lsquoFootprint rsquo calculationsThe program next computes the coordinates of eight points about the perimeter
of the image format (ie located at the four photo corners plus a bisector pointalong each edge of the image) These points are identi ed in R-Photo based uponthe photograph format size and then converted to R-Spacecraft Since all computa-tions are done in orthogonal coordinate systems the R-Spacecraft to R-Photo
J A Robinson et al4428
rotation matrix is transposed to produce an R-Photo to R-Spacecraft rotation matrixOnce in R-Spacecraft a unit vector from each of the eight perimeter points throughthe photo perspective centre (the same point as SC) is computed This provides thecoordinates for points about the perimeter of the image format with their directionvectors in a common coordinate system with other key points needed to computethe photo footprint
The next step is to compute the point of intersection between the spherical earthmodel and each of the eight perimeter point vectors The scalar value for eachperimeter point unit vector is computed using two-dimensional planar trigonometryAn angle c is computed using the formula for the cosine of an angle between twoincluded vectors (the perimeter point unit vector and the vector from SC to thecentre of the Earth) Angle y is computed using the Law of Sines Angle e=180degrees shy c shy y The scalar value of the perimeter point vector is computed using eand the Law of Cosines The scalar value is then multiplied by the perimeter pointunit vector to produce the three-dimensional point of intersection of the vector withEarthrsquos surface in R-Spacecraft The process is repeated independently for each ofthe eight perimeter point vectors Aside from its mathematical simplicity the valueof arcsine (computed in step 2) will exceed the normal range for the sin of an anglewhen the perimeter point vectors fail to intersect with the surface of the earth Asimple test based on this principle allows the program to correctly handle obliquephotos which image a portion of the horizon (see results for high oblique photographsin table 5)
The nal step in the process converts the eight earth intersection points from theR-Spacecraft to R-Earth The results are then converted to the standard geographiccoordinate system and displayed on the lsquoInput-Outputrsquo page of the spreadsheet
54 ExamplesAll three formulations were applied to the photographs included in this paper
and the results are compared in table 5 Formulation 1 gives too small a value fordistance across the photograph (D) for all but the most nadir shots and thus servesas an indicator of the best theoretical case but is not a good measure for a speci cphotograph For example the photograph of Lake Eyre taken from 276 km altitude( gure 2(a)) and the photograph of Limmen Bight ( gure 7) were closest to beingnadir views (oVsetslt68 km or tlt15deg table 5) For these photographs D calculatedusing Formulation 1 was similar to the minimum D calculated using Formulations2 and 3 For almost all other more oblique photographs Formulation 1 gave asigni cant underestimate of the distance covered in the photograph For gure 5(A)(the picture of Houston taken with a 40 mm lens) Formulation 1 did not give anunderestimate for D This is because Formulation 1 does not account for curvatureof the Earth in any way With this large eld of view assuming a at Earth in atedthe value of D above the minimum from calculations that included Earth curvature
A major diVerence between Formulations 2 and 3 is the ability to estimate pixelsizes (P) in both directions (along the principal line and perpendicular to the principalline) For the more oblique photographs the vertical estimate of D and pixel sizes ismuch larger than in the horizontal direction (eg the low oblique and high obliquephotographs of Hawaii gure 4 table 5)
For the area of Limmen Bight ( gure 7) table 5 illustrates the improvement inthe estimate of distance and pixel size that can be obtained by re-estimating thelocation of the PC with greater accuracy Centre points in the catalogued data are
Astronaut photography as digital data for remote sensing 4429
plusmn05deg of latitude and longitude When the centre point was re-estimated to plusmn002degwe determined that the photograph was not taken as obliquely as rst thought(change in the estimate of the look angle t from 169 to 128deg table 5) When theauxiliary point was added to the calculations of Formula 3 the calculated look angleshrank further to 110deg indicating that this photograph was taken at very close toa nadir view Of course this improvement in accuracy could have also led to estimatesof greater obliquity and corresponding larger pixel sizes
We also tested the performance of the scale calculator with auxiliary point byestimating the corner and point locations on the photograph using a 11 000 000Operational Navigational Chart For this test we estimated our ability to readcoordinates from the map as plusmn002degand our error in nding the locations of thecorner points as plusmn015deg (this error varies among photographs depending on thedetail that can be matched between photo and map) For Limmen Bight ( gure 7)the mean diVerence between map estimates and calculator estimates for four pointswas 031deg (SD=018 n=8) For a photograph of San Francisco Bay (STS062-151-291) the mean diVerence between map estimates and calculator estimates forfour points was 0064deg (SD=018 n=8) For a photograph of San Francisco Bay(STS062-151-291) the mean diVerence between map estimates and calculator estim-ates for eight points was 0196deg (SD=0146 n=16) Thus in one case the calculatorestimates were better than our estimate of the error in locating corner points on themap It is reasonable to expect that for nadir-viewing photographs the calculatorused with an auxiliary point can estimate locations of the edges of a photograph towithin plusmn03deg
55 Empirical con rmation of spatial resolution estimatesAs stated previously a challenge to estimating system-AWAR for astronaut
photography of Earth is the lack of suitable targets Small features in an image cansometimes be used as a check on the size of objects that can be successfully resolvedgiving an approximate value for GRD Similarly the number of pixels that make upthose features in the digitized image can be used to make an independent calculationof pixel size We have successfully used features such as roads and airport runwaysto make estimates of spatial scale and resolution (eg Robinson et al 2000c) Whilerecognizing that the use of linear detail in an image is a poor approximation to abar target and that linear objects smaller than the resolving power can often bedetected (Charman 1965) few objects other than roads could be found to make anydirect estimates of GRD Thus roads and runways were used in the images ofHouston (where one can readily conduct ground veri cations and where a numberof higher-contrast concrete roadways were available) to obtain empirical estimatesof GRD and pixel size for comparison with table 5
In the all-digital ESC image of Houston ( gure 3(d )) we examined EllingtonField runway 4-22 (centre left of the image) which is 24384 mtimes457 m This runwayis approximately 6ndash7 pixels in width and 304ndash3094 pixels in length so pixelsrepresent an distance on the ground 7ndash8 m Using a lower contrast measure of astreet length between two intersections (2125 m=211 pixels) a pixel width of 101 mis estimated These results compare favourably with the minimum estimate of 81 mpixels using Formulation 1 (table 5) For an estimate of GRD that would be morecomparable to aerial photography of a line target the smallest street without treecover that could be clearly distinguished on the photograph was 792 m wide Thesmallest non-street object (a gap between stages of the Saturn rocket on display in
J A Robinson et al4430
Tab
le5
App
lica
tio
nof
met
hod
sfo
res
tim
ati
ng
spati
alre
solu
tio
no
fast
ron
aut
ph
oto
gra
ph
sin
clu
din
gF
orm
ula
tion
s12
and
3Im
pro
ved
cen
tre
po
int
esti
mat
esan
dauxilia
ryp
oin
tca
lcu
lati
ons
are
sho
wn
for
gure
7V
aria
ble
sas
use
din
the
text
fle
ns
foca
lle
ngt
hH
al
titu
de
PC
ph
oto
gra
ph
cen
tre
poin
tSN
sp
ace
craft
nadir
po
int
D
dis
tance
cover
edo
nth
egr
oun
d
Pd
ista
nce
on
the
grou
nd
repre
sen
ted
by
apix
el
OVse
td
ista
nce
bet
wee
nP
Cand
SNusi
ng
the
grea
tci
rcle
form
ula
t
look
angle
away
from
SN
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
2L
ake
Eyre
ST
S093
-717
-80
250
276
shy29
0137
0shy
28
413
71
607
11
767
413
7609
ndash64
212
013
760
9ndash64
7121
124
ST
S082
-754
-61
250
617
shy28
5137
0shy
28
113
50
135
726
1201
18
013
8ndash14
827
517
9138ndash15
3277
294
Fig
ure
3H
ou
sto
nS
TS
062
-97-
151
4029
829
5shy
95
530
5shy
95
4410
78
8112
20
5348ndash58
990
1205
351
ndash63
8951
10
36
ST
S094
-737
-28
100
294
295
shy
95
528
3shy
95
5162
31
1133
24
415
8ndash20
334
724
3158ndash20
7351
390
ST
S055
-71-
41250
302
295
shy
95
028
2shy
93
766
412
8192
32
5736
ndash84
715
232
273
9ndash85
7154
185
S73
E51
45300
2685
295
shy
95
024
78
0F
igu
re4
Haw
aii
ST
S069
-701
-65
250
370
195
shy
155
520
4shy
153
881
415
7204
28
9876
ndash98
917
928
688
0ndash10
9181
210
ST
S069
-701
-70
250
370
210
shy
157
519
2shy
150
781
415
7738
63
414
9ndash23
336
760
7148ndash55
6394
10
63
ST
S069
-729
-27
100
398
200
shy
156
620
5shy
148
2219
42
1867
65
432
8ndash13
10
158
62
4323+
311
N
A
Astronaut photography as digital data for remote sensing 4431
Tab
le5
(Cont
inued
)
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
7L
imm
enB
igh
tS
TS
093
-704
-50
250
283
shy15
0135
0shy
15
013
58
623
120
859
16
963
0ndash67
3125
16
863
0ndash67
612
613
2P
CR
e-es
tim
ated
shy14
733
1352
67
64
5128
623
ndash65
5123
128
623
ndash65
712
3127
Auxilia
ryP
oin
t611
062
3ndash65
112
312
5F
igu
re10
N
ear
Au
stin
T
XS
TS
095
-716
-46
250
543
30
5shy
975
28
6shy
1007
120
230
375
34
613
5ndash157
281
34
1136ndash16
128
636
0
1 All
pho
togr
aphs
exce
pt
on
eta
ken
wit
hH
ass
elbla
dca
mer
aso
dis
tan
ceacr
oss
the
image
d=
55m
m
for
S73
E51
45
taken
wit
hel
ectr
onic
still
cam
erad=
276
5m
mho
rizo
nta
l2 P
Cand
SN
are
giv
enin
the
form
(lati
tud
elo
ngit
ude)
deg
rees
w
ith
neg
ativ
en
um
ber
sre
pre
senti
ng
south
and
wes
tA
ccura
cyo
fP
Cis
plusmn0
5and
SN
isplusmn
01
as
reco
rded
inth
eA
stro
naut
Ph
oto
gra
phy
Data
base
ex
cep
tw
her
en
ote
dth
atce
ntr
ep
oin
tw
asre
-est
imate
dw
ith
grea
ter
accu
racy
3 C
alc
ula
ted
usi
ng
the
mea
nofth
em
inim
um
an
dm
axim
um
esti
mate
sof
DF
orm
ula
tion
2co
nsi
der
sD
inth
ed
irec
tion
per
pen
dic
ula
rto
the
Pri
nci
pal
Lin
eo
nly
4 C
alc
ula
ted
inho
rizo
nta
lan
dve
rtic
aldir
ecti
on
sb
ut
still
ass
um
ing
geom
etry
of
gure
6(B
)F
irst
nu
mb
eris
the
mea
no
fth
em
inim
um
Dat
the
bott
om
of
the
ph
oto
and
the
maxi
mum
Dat
the
top
of
the
ph
oto
(range
show
nin
the
tab
le)
and
isco
mpara
ble
toF
orm
ula
tio
n2
Sec
ond
num
ber
isth
em
ean
of
Des
tim
ated
alon
gth
eP
rin
cip
alline
and
alo
ng
the
ou
tsid
eed
ge
(range
not
show
n)
5 Alt
itu
de
isap
pro
xim
ate
for
mis
sion
as
exac
tti
me
pho
togr
ap
hw
asta
ken
and
corr
esp
on
din
gn
adir
data
are
no
tava
ilab
le
Lack
of
nad
irdata
pre
ven
tsap
plica
tion
of
Form
ula
tion
s2
an
d3
6 Au
xilliar
ypo
int
add
edw
as
shy14
80
lati
tud
e13
51
2lo
ngit
ude
shy46
5degro
tati
on
angle
in
stru
ctio
ns
for
esti
mati
ng
the
rota
tio
nan
gle
para
met
erare
conta
ined
as
par
tof
the
scal
eca
lcu
lato
rsp
read
shee
tdo
cum
enta
tio
n
J A Robinson et al4432
a park at Johnson Space Center) that could clearly be distinguished on the photo-graph was 853 m wide
For the photograph of Houston taken with a 250 mm lens ( gure 3(C)) anddigitized from second generation lm at 2400 ppi (106 mm pixel Otilde 1 ) Ellington Fieldrunway 4-22 is 3 pixels in width and 1613 pixels in length so pixels represent151ndash152 m on the ground These results compare favourably with the minimumestimate of 152 m using Formulation 2 and 154ndash185 m pixels using Formulation 3(table 5) For an estimate of GRD using an 8times8 inch print (1369 enlargement) and4times magni cation the smallest street that could clearly be distinguished was 822 mwide the same feature could barely be distinguished on the digitized image
We also made an empirical estimate of spatial resolution for lower contrastvegetation boundaries By clearing forest so that a pattern would be visible to landingaircraft a landowner outside Austin Texas (see also aerial photo in Lisheron 2000)created a target that is also useful for evaluating spatial resolution of astronautphotographs The forest was selectively cleared in order to spell the landownerrsquosname lsquoLUECKErsquo with the remaining trees ( gure 10) According to local surveyorswho planned the clearing the plan was to create letters that were 3100 fttimes1700 ft(9449 mtimes5182 m) Photographed at a high altitude relative to most Shuttle missions(543 km) with a 250 mm lens Formula 3 predicts that each pixel would represent anarea 286 mtimes360 m on the ground (table 5) When original lm was digitized at2400 ppi (106 mm pixel Otilde 1 ) letters correspond to 294times188 pixels for a comparablepixel size of 27ndash32 m
6 Summary and conclusionsAstronaut photographs can be an excellent source of data for remote sensing
applications Best-case resolutions are similar to that for Landsat or SPOT withpixels as small as lt10 m It was estimated that of images taken to date 504 hadlens and obliquity characteristics that make them potential candidates for remotesensing information Digitized astronaut photographs can be overlaid with othersatellite data using GIS (Eckardt et al 2000) or used to ll in gaps in time serieswhen other imagery is not available As a source of public-domain information theycan be very useful for scientists who do not have access to satellite imagery eitherbecause of the expense of image acquisition (to the end user) or the computersystems needed for processing satellite images These diVerences in image acquisitioncosts (to the user) are summarized in table 6
Searching of the complete database of NASA astronaut photography includinglow-spatial-resolution browse images is available via the Web (OYce of EarthSciences 2000) This provides nearly global access for identifying images that willcontribute to a speci c scienti c project For the most detailed studies digitalproducts posted to the web will not be of suYcient quality To date we have madeit a practice of digitizing small numbers of images when requested by scientists atno charge Such requests (including a description of the project involved) can bemade through the authors or using the contact information listed on the websiteInvestigators with needs for larger numbers of digital images have also been servedthrough collaborative agreements
In our experience scientists that nd the data most valuable are those in develop-ing countries those studying areas that have not usually been targets for the majorsatellites those having diYculty in nding low-cloud images those interested inconstructing time series or those interested in using a large number of images
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
Astronaut photography as digital data for remote sensing 4427
distance equal to the focal length from the perspective centre of the camera (whichis assumed to be at SC) along the vector from PC through SC as shown in gure 9
A third coordinate system the Rectangular Photo Coordinate System (R-Photo)is created with its origin at PP its XndashY axial plane normal to the vector from PCthrough SC and its +X axis aligned with the +X axis of R-Spacecraft as shownin gure 9 The XndashY plane of this coordinate system represents the image plane ofthe photograph
534 Auxiliary point calculationsThe calculations above employ a critical assumption that all photos are taken
with the lsquotop of the imagersquo oriented perpendicular to the vector from the SN towardsPC as shown in gure 9 To avoid non-uniform solar heating of the external surfacemost orbiting spacecraft are slowly and continually rotated about one or more axesIn this condition a ight crew member taking a photo while oating in microgravitycould orient the photo with practically any orientation relative to the horizon (seealso gure 8(b)) Unfortunately since these photos are taken with conventionalhandheld cameras there is no other information available which can be used toresolve the photorsquos rotational ambiguity about the optical axis other then the photoitself This is why the above assumption is used and the footprint computed by thiscalculator is actually a lsquolocator ellipsersquo which estimates the area on the ground whatis likely to be included in a speci c photograph (see gure 6) This locator ellipse ismost accurate for square image formats and subject to additional distortion as thephotograph format becomes more rectangular
If the user wants a more precise calculation of the image footprint the photorsquosrotational ambiguity about the optical axis must be resolved This can be done inthe calculator by adding data for an auxiliary point Detailed instructions regardinghow to prepare and use auxiliary point data in the computations are included in thelsquoHow-To-Usersquo tab of the spreadsheet Basically the user determines which side ofthe photograph is top and then measures the angle between the line from PP to thetop of the photo and from PP to the auxiliary point on the photo ( gure 9)
If the user includes data for an auxiliary point a series of computations arecompleted to resolve the photo rotation ambiguity about the optical axis (ie the
+Z axis in R-Photo) A vector from the Auxiliary Point on the Earth (AE) throughthe photograph perspective centre ( located at SC) is intersected with the photo imageplane (XndashY plane of R-Photo) to compute the coordinates of the Auxiliary Point onthe photo (AP) in R-Photo A two-dimensional angle in the XndashY plane of R-Photofrom the shy X axis to a line from PP to AP is calculated as shown in gure 9 The
shy X axis is used as the origin of the angle since it represents the top of the photoonce it passes through the perspective centre The diVerence between the computedangle and the angle measured by the user on the photo resolves the ambiguity inthe rotation of the photo relative to the principal line ( gures 7 and 9) Thetransformations from R-Spacecraft and R-Photo are then modi ed to include anadditional rotation angle about the +Z axis in R-Photo
535 lsquoFootprint rsquo calculationsThe program next computes the coordinates of eight points about the perimeter
of the image format (ie located at the four photo corners plus a bisector pointalong each edge of the image) These points are identi ed in R-Photo based uponthe photograph format size and then converted to R-Spacecraft Since all computa-tions are done in orthogonal coordinate systems the R-Spacecraft to R-Photo
J A Robinson et al4428
rotation matrix is transposed to produce an R-Photo to R-Spacecraft rotation matrixOnce in R-Spacecraft a unit vector from each of the eight perimeter points throughthe photo perspective centre (the same point as SC) is computed This provides thecoordinates for points about the perimeter of the image format with their directionvectors in a common coordinate system with other key points needed to computethe photo footprint
The next step is to compute the point of intersection between the spherical earthmodel and each of the eight perimeter point vectors The scalar value for eachperimeter point unit vector is computed using two-dimensional planar trigonometryAn angle c is computed using the formula for the cosine of an angle between twoincluded vectors (the perimeter point unit vector and the vector from SC to thecentre of the Earth) Angle y is computed using the Law of Sines Angle e=180degrees shy c shy y The scalar value of the perimeter point vector is computed using eand the Law of Cosines The scalar value is then multiplied by the perimeter pointunit vector to produce the three-dimensional point of intersection of the vector withEarthrsquos surface in R-Spacecraft The process is repeated independently for each ofthe eight perimeter point vectors Aside from its mathematical simplicity the valueof arcsine (computed in step 2) will exceed the normal range for the sin of an anglewhen the perimeter point vectors fail to intersect with the surface of the earth Asimple test based on this principle allows the program to correctly handle obliquephotos which image a portion of the horizon (see results for high oblique photographsin table 5)
The nal step in the process converts the eight earth intersection points from theR-Spacecraft to R-Earth The results are then converted to the standard geographiccoordinate system and displayed on the lsquoInput-Outputrsquo page of the spreadsheet
54 ExamplesAll three formulations were applied to the photographs included in this paper
and the results are compared in table 5 Formulation 1 gives too small a value fordistance across the photograph (D) for all but the most nadir shots and thus servesas an indicator of the best theoretical case but is not a good measure for a speci cphotograph For example the photograph of Lake Eyre taken from 276 km altitude( gure 2(a)) and the photograph of Limmen Bight ( gure 7) were closest to beingnadir views (oVsetslt68 km or tlt15deg table 5) For these photographs D calculatedusing Formulation 1 was similar to the minimum D calculated using Formulations2 and 3 For almost all other more oblique photographs Formulation 1 gave asigni cant underestimate of the distance covered in the photograph For gure 5(A)(the picture of Houston taken with a 40 mm lens) Formulation 1 did not give anunderestimate for D This is because Formulation 1 does not account for curvatureof the Earth in any way With this large eld of view assuming a at Earth in atedthe value of D above the minimum from calculations that included Earth curvature
A major diVerence between Formulations 2 and 3 is the ability to estimate pixelsizes (P) in both directions (along the principal line and perpendicular to the principalline) For the more oblique photographs the vertical estimate of D and pixel sizes ismuch larger than in the horizontal direction (eg the low oblique and high obliquephotographs of Hawaii gure 4 table 5)
For the area of Limmen Bight ( gure 7) table 5 illustrates the improvement inthe estimate of distance and pixel size that can be obtained by re-estimating thelocation of the PC with greater accuracy Centre points in the catalogued data are
Astronaut photography as digital data for remote sensing 4429
plusmn05deg of latitude and longitude When the centre point was re-estimated to plusmn002degwe determined that the photograph was not taken as obliquely as rst thought(change in the estimate of the look angle t from 169 to 128deg table 5) When theauxiliary point was added to the calculations of Formula 3 the calculated look angleshrank further to 110deg indicating that this photograph was taken at very close toa nadir view Of course this improvement in accuracy could have also led to estimatesof greater obliquity and corresponding larger pixel sizes
We also tested the performance of the scale calculator with auxiliary point byestimating the corner and point locations on the photograph using a 11 000 000Operational Navigational Chart For this test we estimated our ability to readcoordinates from the map as plusmn002degand our error in nding the locations of thecorner points as plusmn015deg (this error varies among photographs depending on thedetail that can be matched between photo and map) For Limmen Bight ( gure 7)the mean diVerence between map estimates and calculator estimates for four pointswas 031deg (SD=018 n=8) For a photograph of San Francisco Bay (STS062-151-291) the mean diVerence between map estimates and calculator estimates forfour points was 0064deg (SD=018 n=8) For a photograph of San Francisco Bay(STS062-151-291) the mean diVerence between map estimates and calculator estim-ates for eight points was 0196deg (SD=0146 n=16) Thus in one case the calculatorestimates were better than our estimate of the error in locating corner points on themap It is reasonable to expect that for nadir-viewing photographs the calculatorused with an auxiliary point can estimate locations of the edges of a photograph towithin plusmn03deg
55 Empirical con rmation of spatial resolution estimatesAs stated previously a challenge to estimating system-AWAR for astronaut
photography of Earth is the lack of suitable targets Small features in an image cansometimes be used as a check on the size of objects that can be successfully resolvedgiving an approximate value for GRD Similarly the number of pixels that make upthose features in the digitized image can be used to make an independent calculationof pixel size We have successfully used features such as roads and airport runwaysto make estimates of spatial scale and resolution (eg Robinson et al 2000c) Whilerecognizing that the use of linear detail in an image is a poor approximation to abar target and that linear objects smaller than the resolving power can often bedetected (Charman 1965) few objects other than roads could be found to make anydirect estimates of GRD Thus roads and runways were used in the images ofHouston (where one can readily conduct ground veri cations and where a numberof higher-contrast concrete roadways were available) to obtain empirical estimatesof GRD and pixel size for comparison with table 5
In the all-digital ESC image of Houston ( gure 3(d )) we examined EllingtonField runway 4-22 (centre left of the image) which is 24384 mtimes457 m This runwayis approximately 6ndash7 pixels in width and 304ndash3094 pixels in length so pixelsrepresent an distance on the ground 7ndash8 m Using a lower contrast measure of astreet length between two intersections (2125 m=211 pixels) a pixel width of 101 mis estimated These results compare favourably with the minimum estimate of 81 mpixels using Formulation 1 (table 5) For an estimate of GRD that would be morecomparable to aerial photography of a line target the smallest street without treecover that could be clearly distinguished on the photograph was 792 m wide Thesmallest non-street object (a gap between stages of the Saturn rocket on display in
J A Robinson et al4430
Tab
le5
App
lica
tio
nof
met
hod
sfo
res
tim
ati
ng
spati
alre
solu
tio
no
fast
ron
aut
ph
oto
gra
ph
sin
clu
din
gF
orm
ula
tion
s12
and
3Im
pro
ved
cen
tre
po
int
esti
mat
esan
dauxilia
ryp
oin
tca
lcu
lati
ons
are
sho
wn
for
gure
7V
aria
ble
sas
use
din
the
text
fle
ns
foca
lle
ngt
hH
al
titu
de
PC
ph
oto
gra
ph
cen
tre
poin
tSN
sp
ace
craft
nadir
po
int
D
dis
tance
cover
edo
nth
egr
oun
d
Pd
ista
nce
on
the
grou
nd
repre
sen
ted
by
apix
el
OVse
td
ista
nce
bet
wee
nP
Cand
SNusi
ng
the
grea
tci
rcle
form
ula
t
look
angle
away
from
SN
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
2L
ake
Eyre
ST
S093
-717
-80
250
276
shy29
0137
0shy
28
413
71
607
11
767
413
7609
ndash64
212
013
760
9ndash64
7121
124
ST
S082
-754
-61
250
617
shy28
5137
0shy
28
113
50
135
726
1201
18
013
8ndash14
827
517
9138ndash15
3277
294
Fig
ure
3H
ou
sto
nS
TS
062
-97-
151
4029
829
5shy
95
530
5shy
95
4410
78
8112
20
5348ndash58
990
1205
351
ndash63
8951
10
36
ST
S094
-737
-28
100
294
295
shy
95
528
3shy
95
5162
31
1133
24
415
8ndash20
334
724
3158ndash20
7351
390
ST
S055
-71-
41250
302
295
shy
95
028
2shy
93
766
412
8192
32
5736
ndash84
715
232
273
9ndash85
7154
185
S73
E51
45300
2685
295
shy
95
024
78
0F
igu
re4
Haw
aii
ST
S069
-701
-65
250
370
195
shy
155
520
4shy
153
881
415
7204
28
9876
ndash98
917
928
688
0ndash10
9181
210
ST
S069
-701
-70
250
370
210
shy
157
519
2shy
150
781
415
7738
63
414
9ndash23
336
760
7148ndash55
6394
10
63
ST
S069
-729
-27
100
398
200
shy
156
620
5shy
148
2219
42
1867
65
432
8ndash13
10
158
62
4323+
311
N
A
Astronaut photography as digital data for remote sensing 4431
Tab
le5
(Cont
inued
)
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
7L
imm
enB
igh
tS
TS
093
-704
-50
250
283
shy15
0135
0shy
15
013
58
623
120
859
16
963
0ndash67
3125
16
863
0ndash67
612
613
2P
CR
e-es
tim
ated
shy14
733
1352
67
64
5128
623
ndash65
5123
128
623
ndash65
712
3127
Auxilia
ryP
oin
t611
062
3ndash65
112
312
5F
igu
re10
N
ear
Au
stin
T
XS
TS
095
-716
-46
250
543
30
5shy
975
28
6shy
1007
120
230
375
34
613
5ndash157
281
34
1136ndash16
128
636
0
1 All
pho
togr
aphs
exce
pt
on
eta
ken
wit
hH
ass
elbla
dca
mer
aso
dis
tan
ceacr
oss
the
image
d=
55m
m
for
S73
E51
45
taken
wit
hel
ectr
onic
still
cam
erad=
276
5m
mho
rizo
nta
l2 P
Cand
SN
are
giv
enin
the
form
(lati
tud
elo
ngit
ude)
deg
rees
w
ith
neg
ativ
en
um
ber
sre
pre
senti
ng
south
and
wes
tA
ccura
cyo
fP
Cis
plusmn0
5and
SN
isplusmn
01
as
reco
rded
inth
eA
stro
naut
Ph
oto
gra
phy
Data
base
ex
cep
tw
her
en
ote
dth
atce
ntr
ep
oin
tw
asre
-est
imate
dw
ith
grea
ter
accu
racy
3 C
alc
ula
ted
usi
ng
the
mea
nofth
em
inim
um
an
dm
axim
um
esti
mate
sof
DF
orm
ula
tion
2co
nsi
der
sD
inth
ed
irec
tion
per
pen
dic
ula
rto
the
Pri
nci
pal
Lin
eo
nly
4 C
alc
ula
ted
inho
rizo
nta
lan
dve
rtic
aldir
ecti
on
sb
ut
still
ass
um
ing
geom
etry
of
gure
6(B
)F
irst
nu
mb
eris
the
mea
no
fth
em
inim
um
Dat
the
bott
om
of
the
ph
oto
and
the
maxi
mum
Dat
the
top
of
the
ph
oto
(range
show
nin
the
tab
le)
and
isco
mpara
ble
toF
orm
ula
tio
n2
Sec
ond
num
ber
isth
em
ean
of
Des
tim
ated
alon
gth
eP
rin
cip
alline
and
alo
ng
the
ou
tsid
eed
ge
(range
not
show
n)
5 Alt
itu
de
isap
pro
xim
ate
for
mis
sion
as
exac
tti
me
pho
togr
ap
hw
asta
ken
and
corr
esp
on
din
gn
adir
data
are
no
tava
ilab
le
Lack
of
nad
irdata
pre
ven
tsap
plica
tion
of
Form
ula
tion
s2
an
d3
6 Au
xilliar
ypo
int
add
edw
as
shy14
80
lati
tud
e13
51
2lo
ngit
ude
shy46
5degro
tati
on
angle
in
stru
ctio
ns
for
esti
mati
ng
the
rota
tio
nan
gle
para
met
erare
conta
ined
as
par
tof
the
scal
eca
lcu
lato
rsp
read
shee
tdo
cum
enta
tio
n
J A Robinson et al4432
a park at Johnson Space Center) that could clearly be distinguished on the photo-graph was 853 m wide
For the photograph of Houston taken with a 250 mm lens ( gure 3(C)) anddigitized from second generation lm at 2400 ppi (106 mm pixel Otilde 1 ) Ellington Fieldrunway 4-22 is 3 pixels in width and 1613 pixels in length so pixels represent151ndash152 m on the ground These results compare favourably with the minimumestimate of 152 m using Formulation 2 and 154ndash185 m pixels using Formulation 3(table 5) For an estimate of GRD using an 8times8 inch print (1369 enlargement) and4times magni cation the smallest street that could clearly be distinguished was 822 mwide the same feature could barely be distinguished on the digitized image
We also made an empirical estimate of spatial resolution for lower contrastvegetation boundaries By clearing forest so that a pattern would be visible to landingaircraft a landowner outside Austin Texas (see also aerial photo in Lisheron 2000)created a target that is also useful for evaluating spatial resolution of astronautphotographs The forest was selectively cleared in order to spell the landownerrsquosname lsquoLUECKErsquo with the remaining trees ( gure 10) According to local surveyorswho planned the clearing the plan was to create letters that were 3100 fttimes1700 ft(9449 mtimes5182 m) Photographed at a high altitude relative to most Shuttle missions(543 km) with a 250 mm lens Formula 3 predicts that each pixel would represent anarea 286 mtimes360 m on the ground (table 5) When original lm was digitized at2400 ppi (106 mm pixel Otilde 1 ) letters correspond to 294times188 pixels for a comparablepixel size of 27ndash32 m
6 Summary and conclusionsAstronaut photographs can be an excellent source of data for remote sensing
applications Best-case resolutions are similar to that for Landsat or SPOT withpixels as small as lt10 m It was estimated that of images taken to date 504 hadlens and obliquity characteristics that make them potential candidates for remotesensing information Digitized astronaut photographs can be overlaid with othersatellite data using GIS (Eckardt et al 2000) or used to ll in gaps in time serieswhen other imagery is not available As a source of public-domain information theycan be very useful for scientists who do not have access to satellite imagery eitherbecause of the expense of image acquisition (to the end user) or the computersystems needed for processing satellite images These diVerences in image acquisitioncosts (to the user) are summarized in table 6
Searching of the complete database of NASA astronaut photography includinglow-spatial-resolution browse images is available via the Web (OYce of EarthSciences 2000) This provides nearly global access for identifying images that willcontribute to a speci c scienti c project For the most detailed studies digitalproducts posted to the web will not be of suYcient quality To date we have madeit a practice of digitizing small numbers of images when requested by scientists atno charge Such requests (including a description of the project involved) can bemade through the authors or using the contact information listed on the websiteInvestigators with needs for larger numbers of digital images have also been servedthrough collaborative agreements
In our experience scientists that nd the data most valuable are those in develop-ing countries those studying areas that have not usually been targets for the majorsatellites those having diYculty in nding low-cloud images those interested inconstructing time series or those interested in using a large number of images
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
J A Robinson et al4428
rotation matrix is transposed to produce an R-Photo to R-Spacecraft rotation matrixOnce in R-Spacecraft a unit vector from each of the eight perimeter points throughthe photo perspective centre (the same point as SC) is computed This provides thecoordinates for points about the perimeter of the image format with their directionvectors in a common coordinate system with other key points needed to computethe photo footprint
The next step is to compute the point of intersection between the spherical earthmodel and each of the eight perimeter point vectors The scalar value for eachperimeter point unit vector is computed using two-dimensional planar trigonometryAn angle c is computed using the formula for the cosine of an angle between twoincluded vectors (the perimeter point unit vector and the vector from SC to thecentre of the Earth) Angle y is computed using the Law of Sines Angle e=180degrees shy c shy y The scalar value of the perimeter point vector is computed using eand the Law of Cosines The scalar value is then multiplied by the perimeter pointunit vector to produce the three-dimensional point of intersection of the vector withEarthrsquos surface in R-Spacecraft The process is repeated independently for each ofthe eight perimeter point vectors Aside from its mathematical simplicity the valueof arcsine (computed in step 2) will exceed the normal range for the sin of an anglewhen the perimeter point vectors fail to intersect with the surface of the earth Asimple test based on this principle allows the program to correctly handle obliquephotos which image a portion of the horizon (see results for high oblique photographsin table 5)
The nal step in the process converts the eight earth intersection points from theR-Spacecraft to R-Earth The results are then converted to the standard geographiccoordinate system and displayed on the lsquoInput-Outputrsquo page of the spreadsheet
54 ExamplesAll three formulations were applied to the photographs included in this paper
and the results are compared in table 5 Formulation 1 gives too small a value fordistance across the photograph (D) for all but the most nadir shots and thus servesas an indicator of the best theoretical case but is not a good measure for a speci cphotograph For example the photograph of Lake Eyre taken from 276 km altitude( gure 2(a)) and the photograph of Limmen Bight ( gure 7) were closest to beingnadir views (oVsetslt68 km or tlt15deg table 5) For these photographs D calculatedusing Formulation 1 was similar to the minimum D calculated using Formulations2 and 3 For almost all other more oblique photographs Formulation 1 gave asigni cant underestimate of the distance covered in the photograph For gure 5(A)(the picture of Houston taken with a 40 mm lens) Formulation 1 did not give anunderestimate for D This is because Formulation 1 does not account for curvatureof the Earth in any way With this large eld of view assuming a at Earth in atedthe value of D above the minimum from calculations that included Earth curvature
A major diVerence between Formulations 2 and 3 is the ability to estimate pixelsizes (P) in both directions (along the principal line and perpendicular to the principalline) For the more oblique photographs the vertical estimate of D and pixel sizes ismuch larger than in the horizontal direction (eg the low oblique and high obliquephotographs of Hawaii gure 4 table 5)
For the area of Limmen Bight ( gure 7) table 5 illustrates the improvement inthe estimate of distance and pixel size that can be obtained by re-estimating thelocation of the PC with greater accuracy Centre points in the catalogued data are
Astronaut photography as digital data for remote sensing 4429
plusmn05deg of latitude and longitude When the centre point was re-estimated to plusmn002degwe determined that the photograph was not taken as obliquely as rst thought(change in the estimate of the look angle t from 169 to 128deg table 5) When theauxiliary point was added to the calculations of Formula 3 the calculated look angleshrank further to 110deg indicating that this photograph was taken at very close toa nadir view Of course this improvement in accuracy could have also led to estimatesof greater obliquity and corresponding larger pixel sizes
We also tested the performance of the scale calculator with auxiliary point byestimating the corner and point locations on the photograph using a 11 000 000Operational Navigational Chart For this test we estimated our ability to readcoordinates from the map as plusmn002degand our error in nding the locations of thecorner points as plusmn015deg (this error varies among photographs depending on thedetail that can be matched between photo and map) For Limmen Bight ( gure 7)the mean diVerence between map estimates and calculator estimates for four pointswas 031deg (SD=018 n=8) For a photograph of San Francisco Bay (STS062-151-291) the mean diVerence between map estimates and calculator estimates forfour points was 0064deg (SD=018 n=8) For a photograph of San Francisco Bay(STS062-151-291) the mean diVerence between map estimates and calculator estim-ates for eight points was 0196deg (SD=0146 n=16) Thus in one case the calculatorestimates were better than our estimate of the error in locating corner points on themap It is reasonable to expect that for nadir-viewing photographs the calculatorused with an auxiliary point can estimate locations of the edges of a photograph towithin plusmn03deg
55 Empirical con rmation of spatial resolution estimatesAs stated previously a challenge to estimating system-AWAR for astronaut
photography of Earth is the lack of suitable targets Small features in an image cansometimes be used as a check on the size of objects that can be successfully resolvedgiving an approximate value for GRD Similarly the number of pixels that make upthose features in the digitized image can be used to make an independent calculationof pixel size We have successfully used features such as roads and airport runwaysto make estimates of spatial scale and resolution (eg Robinson et al 2000c) Whilerecognizing that the use of linear detail in an image is a poor approximation to abar target and that linear objects smaller than the resolving power can often bedetected (Charman 1965) few objects other than roads could be found to make anydirect estimates of GRD Thus roads and runways were used in the images ofHouston (where one can readily conduct ground veri cations and where a numberof higher-contrast concrete roadways were available) to obtain empirical estimatesof GRD and pixel size for comparison with table 5
In the all-digital ESC image of Houston ( gure 3(d )) we examined EllingtonField runway 4-22 (centre left of the image) which is 24384 mtimes457 m This runwayis approximately 6ndash7 pixels in width and 304ndash3094 pixels in length so pixelsrepresent an distance on the ground 7ndash8 m Using a lower contrast measure of astreet length between two intersections (2125 m=211 pixels) a pixel width of 101 mis estimated These results compare favourably with the minimum estimate of 81 mpixels using Formulation 1 (table 5) For an estimate of GRD that would be morecomparable to aerial photography of a line target the smallest street without treecover that could be clearly distinguished on the photograph was 792 m wide Thesmallest non-street object (a gap between stages of the Saturn rocket on display in
J A Robinson et al4430
Tab
le5
App
lica
tio
nof
met
hod
sfo
res
tim
ati
ng
spati
alre
solu
tio
no
fast
ron
aut
ph
oto
gra
ph
sin
clu
din
gF
orm
ula
tion
s12
and
3Im
pro
ved
cen
tre
po
int
esti
mat
esan
dauxilia
ryp
oin
tca
lcu
lati
ons
are
sho
wn
for
gure
7V
aria
ble
sas
use
din
the
text
fle
ns
foca
lle
ngt
hH
al
titu
de
PC
ph
oto
gra
ph
cen
tre
poin
tSN
sp
ace
craft
nadir
po
int
D
dis
tance
cover
edo
nth
egr
oun
d
Pd
ista
nce
on
the
grou
nd
repre
sen
ted
by
apix
el
OVse
td
ista
nce
bet
wee
nP
Cand
SNusi
ng
the
grea
tci
rcle
form
ula
t
look
angle
away
from
SN
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
2L
ake
Eyre
ST
S093
-717
-80
250
276
shy29
0137
0shy
28
413
71
607
11
767
413
7609
ndash64
212
013
760
9ndash64
7121
124
ST
S082
-754
-61
250
617
shy28
5137
0shy
28
113
50
135
726
1201
18
013
8ndash14
827
517
9138ndash15
3277
294
Fig
ure
3H
ou
sto
nS
TS
062
-97-
151
4029
829
5shy
95
530
5shy
95
4410
78
8112
20
5348ndash58
990
1205
351
ndash63
8951
10
36
ST
S094
-737
-28
100
294
295
shy
95
528
3shy
95
5162
31
1133
24
415
8ndash20
334
724
3158ndash20
7351
390
ST
S055
-71-
41250
302
295
shy
95
028
2shy
93
766
412
8192
32
5736
ndash84
715
232
273
9ndash85
7154
185
S73
E51
45300
2685
295
shy
95
024
78
0F
igu
re4
Haw
aii
ST
S069
-701
-65
250
370
195
shy
155
520
4shy
153
881
415
7204
28
9876
ndash98
917
928
688
0ndash10
9181
210
ST
S069
-701
-70
250
370
210
shy
157
519
2shy
150
781
415
7738
63
414
9ndash23
336
760
7148ndash55
6394
10
63
ST
S069
-729
-27
100
398
200
shy
156
620
5shy
148
2219
42
1867
65
432
8ndash13
10
158
62
4323+
311
N
A
Astronaut photography as digital data for remote sensing 4431
Tab
le5
(Cont
inued
)
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
7L
imm
enB
igh
tS
TS
093
-704
-50
250
283
shy15
0135
0shy
15
013
58
623
120
859
16
963
0ndash67
3125
16
863
0ndash67
612
613
2P
CR
e-es
tim
ated
shy14
733
1352
67
64
5128
623
ndash65
5123
128
623
ndash65
712
3127
Auxilia
ryP
oin
t611
062
3ndash65
112
312
5F
igu
re10
N
ear
Au
stin
T
XS
TS
095
-716
-46
250
543
30
5shy
975
28
6shy
1007
120
230
375
34
613
5ndash157
281
34
1136ndash16
128
636
0
1 All
pho
togr
aphs
exce
pt
on
eta
ken
wit
hH
ass
elbla
dca
mer
aso
dis
tan
ceacr
oss
the
image
d=
55m
m
for
S73
E51
45
taken
wit
hel
ectr
onic
still
cam
erad=
276
5m
mho
rizo
nta
l2 P
Cand
SN
are
giv
enin
the
form
(lati
tud
elo
ngit
ude)
deg
rees
w
ith
neg
ativ
en
um
ber
sre
pre
senti
ng
south
and
wes
tA
ccura
cyo
fP
Cis
plusmn0
5and
SN
isplusmn
01
as
reco
rded
inth
eA
stro
naut
Ph
oto
gra
phy
Data
base
ex
cep
tw
her
en
ote
dth
atce
ntr
ep
oin
tw
asre
-est
imate
dw
ith
grea
ter
accu
racy
3 C
alc
ula
ted
usi
ng
the
mea
nofth
em
inim
um
an
dm
axim
um
esti
mate
sof
DF
orm
ula
tion
2co
nsi
der
sD
inth
ed
irec
tion
per
pen
dic
ula
rto
the
Pri
nci
pal
Lin
eo
nly
4 C
alc
ula
ted
inho
rizo
nta
lan
dve
rtic
aldir
ecti
on
sb
ut
still
ass
um
ing
geom
etry
of
gure
6(B
)F
irst
nu
mb
eris
the
mea
no
fth
em
inim
um
Dat
the
bott
om
of
the
ph
oto
and
the
maxi
mum
Dat
the
top
of
the
ph
oto
(range
show
nin
the
tab
le)
and
isco
mpara
ble
toF
orm
ula
tio
n2
Sec
ond
num
ber
isth
em
ean
of
Des
tim
ated
alon
gth
eP
rin
cip
alline
and
alo
ng
the
ou
tsid
eed
ge
(range
not
show
n)
5 Alt
itu
de
isap
pro
xim
ate
for
mis
sion
as
exac
tti
me
pho
togr
ap
hw
asta
ken
and
corr
esp
on
din
gn
adir
data
are
no
tava
ilab
le
Lack
of
nad
irdata
pre
ven
tsap
plica
tion
of
Form
ula
tion
s2
an
d3
6 Au
xilliar
ypo
int
add
edw
as
shy14
80
lati
tud
e13
51
2lo
ngit
ude
shy46
5degro
tati
on
angle
in
stru
ctio
ns
for
esti
mati
ng
the
rota
tio
nan
gle
para
met
erare
conta
ined
as
par
tof
the
scal
eca
lcu
lato
rsp
read
shee
tdo
cum
enta
tio
n
J A Robinson et al4432
a park at Johnson Space Center) that could clearly be distinguished on the photo-graph was 853 m wide
For the photograph of Houston taken with a 250 mm lens ( gure 3(C)) anddigitized from second generation lm at 2400 ppi (106 mm pixel Otilde 1 ) Ellington Fieldrunway 4-22 is 3 pixels in width and 1613 pixels in length so pixels represent151ndash152 m on the ground These results compare favourably with the minimumestimate of 152 m using Formulation 2 and 154ndash185 m pixels using Formulation 3(table 5) For an estimate of GRD using an 8times8 inch print (1369 enlargement) and4times magni cation the smallest street that could clearly be distinguished was 822 mwide the same feature could barely be distinguished on the digitized image
We also made an empirical estimate of spatial resolution for lower contrastvegetation boundaries By clearing forest so that a pattern would be visible to landingaircraft a landowner outside Austin Texas (see also aerial photo in Lisheron 2000)created a target that is also useful for evaluating spatial resolution of astronautphotographs The forest was selectively cleared in order to spell the landownerrsquosname lsquoLUECKErsquo with the remaining trees ( gure 10) According to local surveyorswho planned the clearing the plan was to create letters that were 3100 fttimes1700 ft(9449 mtimes5182 m) Photographed at a high altitude relative to most Shuttle missions(543 km) with a 250 mm lens Formula 3 predicts that each pixel would represent anarea 286 mtimes360 m on the ground (table 5) When original lm was digitized at2400 ppi (106 mm pixel Otilde 1 ) letters correspond to 294times188 pixels for a comparablepixel size of 27ndash32 m
6 Summary and conclusionsAstronaut photographs can be an excellent source of data for remote sensing
applications Best-case resolutions are similar to that for Landsat or SPOT withpixels as small as lt10 m It was estimated that of images taken to date 504 hadlens and obliquity characteristics that make them potential candidates for remotesensing information Digitized astronaut photographs can be overlaid with othersatellite data using GIS (Eckardt et al 2000) or used to ll in gaps in time serieswhen other imagery is not available As a source of public-domain information theycan be very useful for scientists who do not have access to satellite imagery eitherbecause of the expense of image acquisition (to the end user) or the computersystems needed for processing satellite images These diVerences in image acquisitioncosts (to the user) are summarized in table 6
Searching of the complete database of NASA astronaut photography includinglow-spatial-resolution browse images is available via the Web (OYce of EarthSciences 2000) This provides nearly global access for identifying images that willcontribute to a speci c scienti c project For the most detailed studies digitalproducts posted to the web will not be of suYcient quality To date we have madeit a practice of digitizing small numbers of images when requested by scientists atno charge Such requests (including a description of the project involved) can bemade through the authors or using the contact information listed on the websiteInvestigators with needs for larger numbers of digital images have also been servedthrough collaborative agreements
In our experience scientists that nd the data most valuable are those in develop-ing countries those studying areas that have not usually been targets for the majorsatellites those having diYculty in nding low-cloud images those interested inconstructing time series or those interested in using a large number of images
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
Astronaut photography as digital data for remote sensing 4429
plusmn05deg of latitude and longitude When the centre point was re-estimated to plusmn002degwe determined that the photograph was not taken as obliquely as rst thought(change in the estimate of the look angle t from 169 to 128deg table 5) When theauxiliary point was added to the calculations of Formula 3 the calculated look angleshrank further to 110deg indicating that this photograph was taken at very close toa nadir view Of course this improvement in accuracy could have also led to estimatesof greater obliquity and corresponding larger pixel sizes
We also tested the performance of the scale calculator with auxiliary point byestimating the corner and point locations on the photograph using a 11 000 000Operational Navigational Chart For this test we estimated our ability to readcoordinates from the map as plusmn002degand our error in nding the locations of thecorner points as plusmn015deg (this error varies among photographs depending on thedetail that can be matched between photo and map) For Limmen Bight ( gure 7)the mean diVerence between map estimates and calculator estimates for four pointswas 031deg (SD=018 n=8) For a photograph of San Francisco Bay (STS062-151-291) the mean diVerence between map estimates and calculator estimates forfour points was 0064deg (SD=018 n=8) For a photograph of San Francisco Bay(STS062-151-291) the mean diVerence between map estimates and calculator estim-ates for eight points was 0196deg (SD=0146 n=16) Thus in one case the calculatorestimates were better than our estimate of the error in locating corner points on themap It is reasonable to expect that for nadir-viewing photographs the calculatorused with an auxiliary point can estimate locations of the edges of a photograph towithin plusmn03deg
55 Empirical con rmation of spatial resolution estimatesAs stated previously a challenge to estimating system-AWAR for astronaut
photography of Earth is the lack of suitable targets Small features in an image cansometimes be used as a check on the size of objects that can be successfully resolvedgiving an approximate value for GRD Similarly the number of pixels that make upthose features in the digitized image can be used to make an independent calculationof pixel size We have successfully used features such as roads and airport runwaysto make estimates of spatial scale and resolution (eg Robinson et al 2000c) Whilerecognizing that the use of linear detail in an image is a poor approximation to abar target and that linear objects smaller than the resolving power can often bedetected (Charman 1965) few objects other than roads could be found to make anydirect estimates of GRD Thus roads and runways were used in the images ofHouston (where one can readily conduct ground veri cations and where a numberof higher-contrast concrete roadways were available) to obtain empirical estimatesof GRD and pixel size for comparison with table 5
In the all-digital ESC image of Houston ( gure 3(d )) we examined EllingtonField runway 4-22 (centre left of the image) which is 24384 mtimes457 m This runwayis approximately 6ndash7 pixels in width and 304ndash3094 pixels in length so pixelsrepresent an distance on the ground 7ndash8 m Using a lower contrast measure of astreet length between two intersections (2125 m=211 pixels) a pixel width of 101 mis estimated These results compare favourably with the minimum estimate of 81 mpixels using Formulation 1 (table 5) For an estimate of GRD that would be morecomparable to aerial photography of a line target the smallest street without treecover that could be clearly distinguished on the photograph was 792 m wide Thesmallest non-street object (a gap between stages of the Saturn rocket on display in
J A Robinson et al4430
Tab
le5
App
lica
tio
nof
met
hod
sfo
res
tim
ati
ng
spati
alre
solu
tio
no
fast
ron
aut
ph
oto
gra
ph
sin
clu
din
gF
orm
ula
tion
s12
and
3Im
pro
ved
cen
tre
po
int
esti
mat
esan
dauxilia
ryp
oin
tca
lcu
lati
ons
are
sho
wn
for
gure
7V
aria
ble
sas
use
din
the
text
fle
ns
foca
lle
ngt
hH
al
titu
de
PC
ph
oto
gra
ph
cen
tre
poin
tSN
sp
ace
craft
nadir
po
int
D
dis
tance
cover
edo
nth
egr
oun
d
Pd
ista
nce
on
the
grou
nd
repre
sen
ted
by
apix
el
OVse
td
ista
nce
bet
wee
nP
Cand
SNusi
ng
the
grea
tci
rcle
form
ula
t
look
angle
away
from
SN
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
2L
ake
Eyre
ST
S093
-717
-80
250
276
shy29
0137
0shy
28
413
71
607
11
767
413
7609
ndash64
212
013
760
9ndash64
7121
124
ST
S082
-754
-61
250
617
shy28
5137
0shy
28
113
50
135
726
1201
18
013
8ndash14
827
517
9138ndash15
3277
294
Fig
ure
3H
ou
sto
nS
TS
062
-97-
151
4029
829
5shy
95
530
5shy
95
4410
78
8112
20
5348ndash58
990
1205
351
ndash63
8951
10
36
ST
S094
-737
-28
100
294
295
shy
95
528
3shy
95
5162
31
1133
24
415
8ndash20
334
724
3158ndash20
7351
390
ST
S055
-71-
41250
302
295
shy
95
028
2shy
93
766
412
8192
32
5736
ndash84
715
232
273
9ndash85
7154
185
S73
E51
45300
2685
295
shy
95
024
78
0F
igu
re4
Haw
aii
ST
S069
-701
-65
250
370
195
shy
155
520
4shy
153
881
415
7204
28
9876
ndash98
917
928
688
0ndash10
9181
210
ST
S069
-701
-70
250
370
210
shy
157
519
2shy
150
781
415
7738
63
414
9ndash23
336
760
7148ndash55
6394
10
63
ST
S069
-729
-27
100
398
200
shy
156
620
5shy
148
2219
42
1867
65
432
8ndash13
10
158
62
4323+
311
N
A
Astronaut photography as digital data for remote sensing 4431
Tab
le5
(Cont
inued
)
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
7L
imm
enB
igh
tS
TS
093
-704
-50
250
283
shy15
0135
0shy
15
013
58
623
120
859
16
963
0ndash67
3125
16
863
0ndash67
612
613
2P
CR
e-es
tim
ated
shy14
733
1352
67
64
5128
623
ndash65
5123
128
623
ndash65
712
3127
Auxilia
ryP
oin
t611
062
3ndash65
112
312
5F
igu
re10
N
ear
Au
stin
T
XS
TS
095
-716
-46
250
543
30
5shy
975
28
6shy
1007
120
230
375
34
613
5ndash157
281
34
1136ndash16
128
636
0
1 All
pho
togr
aphs
exce
pt
on
eta
ken
wit
hH
ass
elbla
dca
mer
aso
dis
tan
ceacr
oss
the
image
d=
55m
m
for
S73
E51
45
taken
wit
hel
ectr
onic
still
cam
erad=
276
5m
mho
rizo
nta
l2 P
Cand
SN
are
giv
enin
the
form
(lati
tud
elo
ngit
ude)
deg
rees
w
ith
neg
ativ
en
um
ber
sre
pre
senti
ng
south
and
wes
tA
ccura
cyo
fP
Cis
plusmn0
5and
SN
isplusmn
01
as
reco
rded
inth
eA
stro
naut
Ph
oto
gra
phy
Data
base
ex
cep
tw
her
en
ote
dth
atce
ntr
ep
oin
tw
asre
-est
imate
dw
ith
grea
ter
accu
racy
3 C
alc
ula
ted
usi
ng
the
mea
nofth
em
inim
um
an
dm
axim
um
esti
mate
sof
DF
orm
ula
tion
2co
nsi
der
sD
inth
ed
irec
tion
per
pen
dic
ula
rto
the
Pri
nci
pal
Lin
eo
nly
4 C
alc
ula
ted
inho
rizo
nta
lan
dve
rtic
aldir
ecti
on
sb
ut
still
ass
um
ing
geom
etry
of
gure
6(B
)F
irst
nu
mb
eris
the
mea
no
fth
em
inim
um
Dat
the
bott
om
of
the
ph
oto
and
the
maxi
mum
Dat
the
top
of
the
ph
oto
(range
show
nin
the
tab
le)
and
isco
mpara
ble
toF
orm
ula
tio
n2
Sec
ond
num
ber
isth
em
ean
of
Des
tim
ated
alon
gth
eP
rin
cip
alline
and
alo
ng
the
ou
tsid
eed
ge
(range
not
show
n)
5 Alt
itu
de
isap
pro
xim
ate
for
mis
sion
as
exac
tti
me
pho
togr
ap
hw
asta
ken
and
corr
esp
on
din
gn
adir
data
are
no
tava
ilab
le
Lack
of
nad
irdata
pre
ven
tsap
plica
tion
of
Form
ula
tion
s2
an
d3
6 Au
xilliar
ypo
int
add
edw
as
shy14
80
lati
tud
e13
51
2lo
ngit
ude
shy46
5degro
tati
on
angle
in
stru
ctio
ns
for
esti
mati
ng
the
rota
tio
nan
gle
para
met
erare
conta
ined
as
par
tof
the
scal
eca
lcu
lato
rsp
read
shee
tdo
cum
enta
tio
n
J A Robinson et al4432
a park at Johnson Space Center) that could clearly be distinguished on the photo-graph was 853 m wide
For the photograph of Houston taken with a 250 mm lens ( gure 3(C)) anddigitized from second generation lm at 2400 ppi (106 mm pixel Otilde 1 ) Ellington Fieldrunway 4-22 is 3 pixels in width and 1613 pixels in length so pixels represent151ndash152 m on the ground These results compare favourably with the minimumestimate of 152 m using Formulation 2 and 154ndash185 m pixels using Formulation 3(table 5) For an estimate of GRD using an 8times8 inch print (1369 enlargement) and4times magni cation the smallest street that could clearly be distinguished was 822 mwide the same feature could barely be distinguished on the digitized image
We also made an empirical estimate of spatial resolution for lower contrastvegetation boundaries By clearing forest so that a pattern would be visible to landingaircraft a landowner outside Austin Texas (see also aerial photo in Lisheron 2000)created a target that is also useful for evaluating spatial resolution of astronautphotographs The forest was selectively cleared in order to spell the landownerrsquosname lsquoLUECKErsquo with the remaining trees ( gure 10) According to local surveyorswho planned the clearing the plan was to create letters that were 3100 fttimes1700 ft(9449 mtimes5182 m) Photographed at a high altitude relative to most Shuttle missions(543 km) with a 250 mm lens Formula 3 predicts that each pixel would represent anarea 286 mtimes360 m on the ground (table 5) When original lm was digitized at2400 ppi (106 mm pixel Otilde 1 ) letters correspond to 294times188 pixels for a comparablepixel size of 27ndash32 m
6 Summary and conclusionsAstronaut photographs can be an excellent source of data for remote sensing
applications Best-case resolutions are similar to that for Landsat or SPOT withpixels as small as lt10 m It was estimated that of images taken to date 504 hadlens and obliquity characteristics that make them potential candidates for remotesensing information Digitized astronaut photographs can be overlaid with othersatellite data using GIS (Eckardt et al 2000) or used to ll in gaps in time serieswhen other imagery is not available As a source of public-domain information theycan be very useful for scientists who do not have access to satellite imagery eitherbecause of the expense of image acquisition (to the end user) or the computersystems needed for processing satellite images These diVerences in image acquisitioncosts (to the user) are summarized in table 6
Searching of the complete database of NASA astronaut photography includinglow-spatial-resolution browse images is available via the Web (OYce of EarthSciences 2000) This provides nearly global access for identifying images that willcontribute to a speci c scienti c project For the most detailed studies digitalproducts posted to the web will not be of suYcient quality To date we have madeit a practice of digitizing small numbers of images when requested by scientists atno charge Such requests (including a description of the project involved) can bemade through the authors or using the contact information listed on the websiteInvestigators with needs for larger numbers of digital images have also been servedthrough collaborative agreements
In our experience scientists that nd the data most valuable are those in develop-ing countries those studying areas that have not usually been targets for the majorsatellites those having diYculty in nding low-cloud images those interested inconstructing time series or those interested in using a large number of images
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
J A Robinson et al4430
Tab
le5
App
lica
tio
nof
met
hod
sfo
res
tim
ati
ng
spati
alre
solu
tio
no
fast
ron
aut
ph
oto
gra
ph
sin
clu
din
gF
orm
ula
tion
s12
and
3Im
pro
ved
cen
tre
po
int
esti
mat
esan
dauxilia
ryp
oin
tca
lcu
lati
ons
are
sho
wn
for
gure
7V
aria
ble
sas
use
din
the
text
fle
ns
foca
lle
ngt
hH
al
titu
de
PC
ph
oto
gra
ph
cen
tre
poin
tSN
sp
ace
craft
nadir
po
int
D
dis
tance
cover
edo
nth
egr
oun
d
Pd
ista
nce
on
the
grou
nd
repre
sen
ted
by
apix
el
OVse
td
ista
nce
bet
wee
nP
Cand
SNusi
ng
the
grea
tci
rcle
form
ula
t
look
angle
away
from
SN
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
2L
ake
Eyre
ST
S093
-717
-80
250
276
shy29
0137
0shy
28
413
71
607
11
767
413
7609
ndash64
212
013
760
9ndash64
7121
124
ST
S082
-754
-61
250
617
shy28
5137
0shy
28
113
50
135
726
1201
18
013
8ndash14
827
517
9138ndash15
3277
294
Fig
ure
3H
ou
sto
nS
TS
062
-97-
151
4029
829
5shy
95
530
5shy
95
4410
78
8112
20
5348ndash58
990
1205
351
ndash63
8951
10
36
ST
S094
-737
-28
100
294
295
shy
95
528
3shy
95
5162
31
1133
24
415
8ndash20
334
724
3158ndash20
7351
390
ST
S055
-71-
41250
302
295
shy
95
028
2shy
93
766
412
8192
32
5736
ndash84
715
232
273
9ndash85
7154
185
S73
E51
45300
2685
295
shy
95
024
78
0F
igu
re4
Haw
aii
ST
S069
-701
-65
250
370
195
shy
155
520
4shy
153
881
415
7204
28
9876
ndash98
917
928
688
0ndash10
9181
210
ST
S069
-701
-70
250
370
210
shy
157
519
2shy
150
781
415
7738
63
414
9ndash23
336
760
7148ndash55
6394
10
63
ST
S069
-729
-27
100
398
200
shy
156
620
5shy
148
2219
42
1867
65
432
8ndash13
10
158
62
4323+
311
N
A
Astronaut photography as digital data for remote sensing 4431
Tab
le5
(Cont
inued
)
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
7L
imm
enB
igh
tS
TS
093
-704
-50
250
283
shy15
0135
0shy
15
013
58
623
120
859
16
963
0ndash67
3125
16
863
0ndash67
612
613
2P
CR
e-es
tim
ated
shy14
733
1352
67
64
5128
623
ndash65
5123
128
623
ndash65
712
3127
Auxilia
ryP
oin
t611
062
3ndash65
112
312
5F
igu
re10
N
ear
Au
stin
T
XS
TS
095
-716
-46
250
543
30
5shy
975
28
6shy
1007
120
230
375
34
613
5ndash157
281
34
1136ndash16
128
636
0
1 All
pho
togr
aphs
exce
pt
on
eta
ken
wit
hH
ass
elbla
dca
mer
aso
dis
tan
ceacr
oss
the
image
d=
55m
m
for
S73
E51
45
taken
wit
hel
ectr
onic
still
cam
erad=
276
5m
mho
rizo
nta
l2 P
Cand
SN
are
giv
enin
the
form
(lati
tud
elo
ngit
ude)
deg
rees
w
ith
neg
ativ
en
um
ber
sre
pre
senti
ng
south
and
wes
tA
ccura
cyo
fP
Cis
plusmn0
5and
SN
isplusmn
01
as
reco
rded
inth
eA
stro
naut
Ph
oto
gra
phy
Data
base
ex
cep
tw
her
en
ote
dth
atce
ntr
ep
oin
tw
asre
-est
imate
dw
ith
grea
ter
accu
racy
3 C
alc
ula
ted
usi
ng
the
mea
nofth
em
inim
um
an
dm
axim
um
esti
mate
sof
DF
orm
ula
tion
2co
nsi
der
sD
inth
ed
irec
tion
per
pen
dic
ula
rto
the
Pri
nci
pal
Lin
eo
nly
4 C
alc
ula
ted
inho
rizo
nta
lan
dve
rtic
aldir
ecti
on
sb
ut
still
ass
um
ing
geom
etry
of
gure
6(B
)F
irst
nu
mb
eris
the
mea
no
fth
em
inim
um
Dat
the
bott
om
of
the
ph
oto
and
the
maxi
mum
Dat
the
top
of
the
ph
oto
(range
show
nin
the
tab
le)
and
isco
mpara
ble
toF
orm
ula
tio
n2
Sec
ond
num
ber
isth
em
ean
of
Des
tim
ated
alon
gth
eP
rin
cip
alline
and
alo
ng
the
ou
tsid
eed
ge
(range
not
show
n)
5 Alt
itu
de
isap
pro
xim
ate
for
mis
sion
as
exac
tti
me
pho
togr
ap
hw
asta
ken
and
corr
esp
on
din
gn
adir
data
are
no
tava
ilab
le
Lack
of
nad
irdata
pre
ven
tsap
plica
tion
of
Form
ula
tion
s2
an
d3
6 Au
xilliar
ypo
int
add
edw
as
shy14
80
lati
tud
e13
51
2lo
ngit
ude
shy46
5degro
tati
on
angle
in
stru
ctio
ns
for
esti
mati
ng
the
rota
tio
nan
gle
para
met
erare
conta
ined
as
par
tof
the
scal
eca
lcu
lato
rsp
read
shee
tdo
cum
enta
tio
n
J A Robinson et al4432
a park at Johnson Space Center) that could clearly be distinguished on the photo-graph was 853 m wide
For the photograph of Houston taken with a 250 mm lens ( gure 3(C)) anddigitized from second generation lm at 2400 ppi (106 mm pixel Otilde 1 ) Ellington Fieldrunway 4-22 is 3 pixels in width and 1613 pixels in length so pixels represent151ndash152 m on the ground These results compare favourably with the minimumestimate of 152 m using Formulation 2 and 154ndash185 m pixels using Formulation 3(table 5) For an estimate of GRD using an 8times8 inch print (1369 enlargement) and4times magni cation the smallest street that could clearly be distinguished was 822 mwide the same feature could barely be distinguished on the digitized image
We also made an empirical estimate of spatial resolution for lower contrastvegetation boundaries By clearing forest so that a pattern would be visible to landingaircraft a landowner outside Austin Texas (see also aerial photo in Lisheron 2000)created a target that is also useful for evaluating spatial resolution of astronautphotographs The forest was selectively cleared in order to spell the landownerrsquosname lsquoLUECKErsquo with the remaining trees ( gure 10) According to local surveyorswho planned the clearing the plan was to create letters that were 3100 fttimes1700 ft(9449 mtimes5182 m) Photographed at a high altitude relative to most Shuttle missions(543 km) with a 250 mm lens Formula 3 predicts that each pixel would represent anarea 286 mtimes360 m on the ground (table 5) When original lm was digitized at2400 ppi (106 mm pixel Otilde 1 ) letters correspond to 294times188 pixels for a comparablepixel size of 27ndash32 m
6 Summary and conclusionsAstronaut photographs can be an excellent source of data for remote sensing
applications Best-case resolutions are similar to that for Landsat or SPOT withpixels as small as lt10 m It was estimated that of images taken to date 504 hadlens and obliquity characteristics that make them potential candidates for remotesensing information Digitized astronaut photographs can be overlaid with othersatellite data using GIS (Eckardt et al 2000) or used to ll in gaps in time serieswhen other imagery is not available As a source of public-domain information theycan be very useful for scientists who do not have access to satellite imagery eitherbecause of the expense of image acquisition (to the end user) or the computersystems needed for processing satellite images These diVerences in image acquisitioncosts (to the user) are summarized in table 6
Searching of the complete database of NASA astronaut photography includinglow-spatial-resolution browse images is available via the Web (OYce of EarthSciences 2000) This provides nearly global access for identifying images that willcontribute to a speci c scienti c project For the most detailed studies digitalproducts posted to the web will not be of suYcient quality To date we have madeit a practice of digitizing small numbers of images when requested by scientists atno charge Such requests (including a description of the project involved) can bemade through the authors or using the contact information listed on the websiteInvestigators with needs for larger numbers of digital images have also been servedthrough collaborative agreements
In our experience scientists that nd the data most valuable are those in develop-ing countries those studying areas that have not usually been targets for the majorsatellites those having diYculty in nding low-cloud images those interested inconstructing time series or those interested in using a large number of images
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
Astronaut photography as digital data for remote sensing 4431
Tab
le5
(Cont
inued
)
Form
ula
tion
1F
orm
ula
tio
n2
Form
ula
tio
n3
fH
DP
OV
set
tR
ange
DP
3t
Ran
ge
DP
4P
ho
togr
aph1
(mm
)(k
m)
PC
2S
N(k
m)
(m)
(km
)(deg
)(k
m)
(m)
(deg)
(km
)(m
)
Fig
ure
7L
imm
enB
igh
tS
TS
093
-704
-50
250
283
shy15
0135
0shy
15
013
58
623
120
859
16
963
0ndash67
3125
16
863
0ndash67
612
613
2P
CR
e-es
tim
ated
shy14
733
1352
67
64
5128
623
ndash65
5123
128
623
ndash65
712
3127
Auxilia
ryP
oin
t611
062
3ndash65
112
312
5F
igu
re10
N
ear
Au
stin
T
XS
TS
095
-716
-46
250
543
30
5shy
975
28
6shy
1007
120
230
375
34
613
5ndash157
281
34
1136ndash16
128
636
0
1 All
pho
togr
aphs
exce
pt
on
eta
ken
wit
hH
ass
elbla
dca
mer
aso
dis
tan
ceacr
oss
the
image
d=
55m
m
for
S73
E51
45
taken
wit
hel
ectr
onic
still
cam
erad=
276
5m
mho
rizo
nta
l2 P
Cand
SN
are
giv
enin
the
form
(lati
tud
elo
ngit
ude)
deg
rees
w
ith
neg
ativ
en
um
ber
sre
pre
senti
ng
south
and
wes
tA
ccura
cyo
fP
Cis
plusmn0
5and
SN
isplusmn
01
as
reco
rded
inth
eA
stro
naut
Ph
oto
gra
phy
Data
base
ex
cep
tw
her
en
ote
dth
atce
ntr
ep
oin
tw
asre
-est
imate
dw
ith
grea
ter
accu
racy
3 C
alc
ula
ted
usi
ng
the
mea
nofth
em
inim
um
an
dm
axim
um
esti
mate
sof
DF
orm
ula
tion
2co
nsi
der
sD
inth
ed
irec
tion
per
pen
dic
ula
rto
the
Pri
nci
pal
Lin
eo
nly
4 C
alc
ula
ted
inho
rizo
nta
lan
dve
rtic
aldir
ecti
on
sb
ut
still
ass
um
ing
geom
etry
of
gure
6(B
)F
irst
nu
mb
eris
the
mea
no
fth
em
inim
um
Dat
the
bott
om
of
the
ph
oto
and
the
maxi
mum
Dat
the
top
of
the
ph
oto
(range
show
nin
the
tab
le)
and
isco
mpara
ble
toF
orm
ula
tio
n2
Sec
ond
num
ber
isth
em
ean
of
Des
tim
ated
alon
gth
eP
rin
cip
alline
and
alo
ng
the
ou
tsid
eed
ge
(range
not
show
n)
5 Alt
itu
de
isap
pro
xim
ate
for
mis
sion
as
exac
tti
me
pho
togr
ap
hw
asta
ken
and
corr
esp
on
din
gn
adir
data
are
no
tava
ilab
le
Lack
of
nad
irdata
pre
ven
tsap
plica
tion
of
Form
ula
tion
s2
an
d3
6 Au
xilliar
ypo
int
add
edw
as
shy14
80
lati
tud
e13
51
2lo
ngit
ude
shy46
5degro
tati
on
angle
in
stru
ctio
ns
for
esti
mati
ng
the
rota
tio
nan
gle
para
met
erare
conta
ined
as
par
tof
the
scal
eca
lcu
lato
rsp
read
shee
tdo
cum
enta
tio
n
J A Robinson et al4432
a park at Johnson Space Center) that could clearly be distinguished on the photo-graph was 853 m wide
For the photograph of Houston taken with a 250 mm lens ( gure 3(C)) anddigitized from second generation lm at 2400 ppi (106 mm pixel Otilde 1 ) Ellington Fieldrunway 4-22 is 3 pixels in width and 1613 pixels in length so pixels represent151ndash152 m on the ground These results compare favourably with the minimumestimate of 152 m using Formulation 2 and 154ndash185 m pixels using Formulation 3(table 5) For an estimate of GRD using an 8times8 inch print (1369 enlargement) and4times magni cation the smallest street that could clearly be distinguished was 822 mwide the same feature could barely be distinguished on the digitized image
We also made an empirical estimate of spatial resolution for lower contrastvegetation boundaries By clearing forest so that a pattern would be visible to landingaircraft a landowner outside Austin Texas (see also aerial photo in Lisheron 2000)created a target that is also useful for evaluating spatial resolution of astronautphotographs The forest was selectively cleared in order to spell the landownerrsquosname lsquoLUECKErsquo with the remaining trees ( gure 10) According to local surveyorswho planned the clearing the plan was to create letters that were 3100 fttimes1700 ft(9449 mtimes5182 m) Photographed at a high altitude relative to most Shuttle missions(543 km) with a 250 mm lens Formula 3 predicts that each pixel would represent anarea 286 mtimes360 m on the ground (table 5) When original lm was digitized at2400 ppi (106 mm pixel Otilde 1 ) letters correspond to 294times188 pixels for a comparablepixel size of 27ndash32 m
6 Summary and conclusionsAstronaut photographs can be an excellent source of data for remote sensing
applications Best-case resolutions are similar to that for Landsat or SPOT withpixels as small as lt10 m It was estimated that of images taken to date 504 hadlens and obliquity characteristics that make them potential candidates for remotesensing information Digitized astronaut photographs can be overlaid with othersatellite data using GIS (Eckardt et al 2000) or used to ll in gaps in time serieswhen other imagery is not available As a source of public-domain information theycan be very useful for scientists who do not have access to satellite imagery eitherbecause of the expense of image acquisition (to the end user) or the computersystems needed for processing satellite images These diVerences in image acquisitioncosts (to the user) are summarized in table 6
Searching of the complete database of NASA astronaut photography includinglow-spatial-resolution browse images is available via the Web (OYce of EarthSciences 2000) This provides nearly global access for identifying images that willcontribute to a speci c scienti c project For the most detailed studies digitalproducts posted to the web will not be of suYcient quality To date we have madeit a practice of digitizing small numbers of images when requested by scientists atno charge Such requests (including a description of the project involved) can bemade through the authors or using the contact information listed on the websiteInvestigators with needs for larger numbers of digital images have also been servedthrough collaborative agreements
In our experience scientists that nd the data most valuable are those in develop-ing countries those studying areas that have not usually been targets for the majorsatellites those having diYculty in nding low-cloud images those interested inconstructing time series or those interested in using a large number of images
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
J A Robinson et al4432
a park at Johnson Space Center) that could clearly be distinguished on the photo-graph was 853 m wide
For the photograph of Houston taken with a 250 mm lens ( gure 3(C)) anddigitized from second generation lm at 2400 ppi (106 mm pixel Otilde 1 ) Ellington Fieldrunway 4-22 is 3 pixels in width and 1613 pixels in length so pixels represent151ndash152 m on the ground These results compare favourably with the minimumestimate of 152 m using Formulation 2 and 154ndash185 m pixels using Formulation 3(table 5) For an estimate of GRD using an 8times8 inch print (1369 enlargement) and4times magni cation the smallest street that could clearly be distinguished was 822 mwide the same feature could barely be distinguished on the digitized image
We also made an empirical estimate of spatial resolution for lower contrastvegetation boundaries By clearing forest so that a pattern would be visible to landingaircraft a landowner outside Austin Texas (see also aerial photo in Lisheron 2000)created a target that is also useful for evaluating spatial resolution of astronautphotographs The forest was selectively cleared in order to spell the landownerrsquosname lsquoLUECKErsquo with the remaining trees ( gure 10) According to local surveyorswho planned the clearing the plan was to create letters that were 3100 fttimes1700 ft(9449 mtimes5182 m) Photographed at a high altitude relative to most Shuttle missions(543 km) with a 250 mm lens Formula 3 predicts that each pixel would represent anarea 286 mtimes360 m on the ground (table 5) When original lm was digitized at2400 ppi (106 mm pixel Otilde 1 ) letters correspond to 294times188 pixels for a comparablepixel size of 27ndash32 m
6 Summary and conclusionsAstronaut photographs can be an excellent source of data for remote sensing
applications Best-case resolutions are similar to that for Landsat or SPOT withpixels as small as lt10 m It was estimated that of images taken to date 504 hadlens and obliquity characteristics that make them potential candidates for remotesensing information Digitized astronaut photographs can be overlaid with othersatellite data using GIS (Eckardt et al 2000) or used to ll in gaps in time serieswhen other imagery is not available As a source of public-domain information theycan be very useful for scientists who do not have access to satellite imagery eitherbecause of the expense of image acquisition (to the end user) or the computersystems needed for processing satellite images These diVerences in image acquisitioncosts (to the user) are summarized in table 6
Searching of the complete database of NASA astronaut photography includinglow-spatial-resolution browse images is available via the Web (OYce of EarthSciences 2000) This provides nearly global access for identifying images that willcontribute to a speci c scienti c project For the most detailed studies digitalproducts posted to the web will not be of suYcient quality To date we have madeit a practice of digitizing small numbers of images when requested by scientists atno charge Such requests (including a description of the project involved) can bemade through the authors or using the contact information listed on the websiteInvestigators with needs for larger numbers of digital images have also been servedthrough collaborative agreements
In our experience scientists that nd the data most valuable are those in develop-ing countries those studying areas that have not usually been targets for the majorsatellites those having diYculty in nding low-cloud images those interested inconstructing time series or those interested in using a large number of images
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
Astronaut photography as digital data for remote sensing 4433
Figure 10 Patterns of forest clearing outside Austin Texas serve as a test pattern forestimating spatial resolution (a) Complete eld of view for STS095-716-46 taken witha 250 mm lens from 543 km altitude (2 November 1998) (b) Detail of a portion of theframe (c) Aerial photograph taken with an ESC (Kodak DCS620C) from an unknownaltitude with zoom lens (2 March 2000 B K Diggs Austin-American Statesmanused with permission) (d ) Detail showing the limits of spatial resolution when the lmwas digitized at 2400 ppi (106 mm pixelOtilde 1 ) The legend in the right-hand corner showsthe eld measurements for the letters (P Tovar Jr personal communication) and thecorresponding pixel size
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
J A Robinson et al4434
Table
6C
om
pari
sons
of
chara
cter
isti
csof
ast
ron
aut-
dir
ecte
dp
ho
togr
aphy
of
Ear
thw
ith
auto
mati
csa
tellit
ese
nso
rs1
Info
rmat
ion
com
piled
fro
mw
ebpage
sand
pre
ssre
lease
sp
rovi
ded
by
the
age
nt
list
edun
der
lsquoRef
eren
cesrsquo
Land
sat
Ast
ronaut-
acq
uir
edSP
OT
orb
italph
oto
gra
ph
yM
SS
TM
ET
M+
XS
AV
HR
RC
ZC
SSea
WiF
S
Len
gth
of
reco
rd19
65ndash
1972
ndash19
82ndash
1999ndash
198
6ndash
1979
ndash19
78ndash1986
1997
ndashSp
atia
lre
solu
tio
n2
m9ndash65
79
30
30
20110
0times40
00825
1100
ndash4400
Alt
itu
de
km
222
ndash611
907ndash91
5705
705
822
830
ndash870
955
705
Incl
inati
on
285
ndash51
699
2deg982
deg982
deg98
deg81deg
99
1deg
983
degSce
ne
size
km
Vari
able
185times
185
185times
170
185times
170
60times
60
239
9sw
ath
1566
swath
2800
swath
Co
vera
gefr
equ
ency
Inte
rmit
tent
18
days
16
days
16
day
s26
day
s2times
per
day
6d
ays
1day
Su
n-s
ynch
rono
us
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
orb
it3
Vie
wan
gles
Nad
irO
bliqu
eN
adir
Nadir
Nadir
Nadir
O
bli
qu
eN
adir
Nadir
plusmn
40deg
Nadir
plusmn
20deg
Est
imat
edpro
duct
$0ndash25
$20
0$425
$600
$155
0ndash230
00
00
cost
p
erpre
vio
usl
yacq
uir
edsc
ene
(basi
c)R
efer
ence
sN
AS
AE
RO
SE
RO
SE
RO
SSP
OT
NO
AA
NA
SA
NA
SA
NA
SA
1 Ab
bre
viat
ion
sfo
rse
nso
rsand
gover
nm
ent
age
nci
esare
as
follow
sM
SS
Mult
i-sp
ectr
al
Sca
nner
T
MT
hem
atic
Map
per
E
TM
+E
nhan
ced
Th
emat
icM
app
erP
lus
AV
HR
RA
dvance
dV
ery-h
igh
Res
olu
tio
nR
adio
met
erC
ZC
SC
oast
alZ
on
eC
olo
rS
cann
erS
eaW
iFS
Sea
-vie
win
gW
ide
Fie
ld-
of-
view
Sen
sor
SP
OT
Sate
llit
epo
ur
lrsquoOb
serv
ati
on
de
laT
erre
X
SM
ult
isp
ectr
alm
ode
ER
OS
Eart
hR
esou
rces
Obse
rvat
ion
Syst
ems
Data
Cen
ter
Un
ited
Sta
tes
Geo
logi
cal
Surv
ey
Sio
ux
Falls
So
uth
Dako
ta
NO
AA
Nati
onal
Oce
an
ogra
ph
icand
Atm
osp
her
icA
dm
inis
trati
on
NA
SA
Nat
ion
alA
eron
auti
csan
dSp
ace
Adm
inis
trat
ion
2 A
ddit
ion
aldet
ail
isin
table
2B
est-
case
nad
irp
ixel
size
for
ara
nge
of
alti
tudes
isin
clud
edfo
ro
rbit
al
pho
togr
aphs
3 Det
erm
ines
deg
ree
of
var
iab
ilit
yo
flighti
ng
con
dit
ion
sam
on
gim
ages
taken
of
the
sam
epla
ceat
diV
eren
tti
mes
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
Astronaut photography as digital data for remote sensing 4435
Because use of astronaut photography data is unfamiliar to most scientists andremote sensing experts we have tried to provide a general synopsis of its majorcharacteristics One common source of confusion about the data arises from itsvariable spatial resolution The issue of spatial resolution of astronaut photographshas been treated to a level of detail that has not been previously published Inaddition a primer of equations has been provided that can be used for calculatingspatial resolution of a given photograph and user-friendly methods have beenincorporated for non-specialists to estimate spatial resolution of speci c photographs(using tools on our Web interface or by downloading a spreadsheet) Although morevariable than other types of satellite data the information in the images can beextracted using familiar remote sensing techniques such as georeferencing and imageclassi cation This makes the data source valuable for remote sensing applicationsin ecology and conservation biology geography geology and other related elds
AcknowledgmentsWe thank the astronauts cataloguers and scientists whose eVorts over the last
25 years have developed and preserved the remote sensing resource described in thispaper Barbara Boland served as a primary beta-tester for the Photo FootprintCalculator and helped to improve its user interface James Heydorn searched data-bases to help in compilation of summary statistics and gure 5 he also did theprogramming for our web display of photo footprints Joe Caruana researched older lm types James C Maida helped to produce the three-dimensional visualizationin gure 9 Karen Scott provided comments on this manuscript and input into thesection on window eVects Serge Andrefouet Kamlesh P Lulla Edward L Webband anonymous reviewers made constructive comments that greatly improved themanuscript
ReferencesAmsbury D L 1989 United States manned observations of Earth before the Space Shuttle
Geocarto International 4 (1) 7ndash14Arnold R H 1997 Interpretation of Airphotos and Remotely Sensed Imagery (Upper Saddle
River NJ Prentice Hall )Baltsavias E P 25 April 1996 Desktop publishing scanners OEEPE Workshop on the
Application of Digital Photogrammetric Workstations 4ndash6 March 1996 L ausannehttpdgrwwwep chPHOTworkshopwks96art_1_4html
Campbell J B 1996 Introduction to Remote Sensing 2nd edn (New York Guilford Press)Chamberlain R G 1996 What is the best way to calculate the distance between two points
GIS FAQ Question 51 httpwwwcensusgovcgi-bingeogisfaqQ51 (revised inNovember 1997 and February 1998 11 April 2000)
Charman W N 1965 Resolving power and the detection of linear detail T he CanadianSurveyor 19 190ndash205
Colwell R N 1971 Monitoring Earth Resources from Aircraft and Spacecraft NASASP-275 (Washington DC National Aeronautics and Space Administration)
Drury S A 1993 Image Interpretation in Geology 2nd edn (London Chapman and Hall )Eastman Kodak Company 1998 Kodak Aerochrome II Inf rared lm 2443 Kodak Aerochrome
II Infrared NP lm SO-134 Aerial Systems Data Sheet TI2161 Revised 3-98 Alsoavailable at httpwwwkodakcomUSengovernmentaerialtechnicalPubstiDocsti2161ti2161shtml (29 November 1999)
Eckardt F D Wilkinson M J and Lulla K P 2000 Using digitized handheld SpaceShuttle photography for terrain visualization International Journal of Remote Sensing21 1ndash5
El-Baz F 1977 Astronaut Observations from the Apollo-Soyuz Mission Smithsonian Studiesin Air and Space Vol 1 (Washington DC Smithsonian Institution Press)
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
J A Robinson et al4436
El-Baz F and Warner D M 1979 Apollo-Soyuz T est Project Vol II Earth Observationsand Photography NASA SP-412 (Houston Texas National Aeronautics and SpaceAdministration Lyndon B Johnson Space Center)
Eppler D Amsbury D and Evans C 1996 Interest sought for research aboard newwindow on the world the International Space Station Eos T ransactions AmericanGeophysical Union 77 121127
Estes J E and Simonett D S 1975 Fundamentals of image interpretation In Manual ofRemote Sensing edited by R G Reeves (Falls Church VA American Society ofPhotogrammetry) pp 869ndash1076
Evans C A Lulla K P Dessinov L V Glazovskiy N F Kasimov N S andKnizhnikov Yu F 2000 Shuttle-Mir Earth science investigations studying dynamicEarth environments from the Mir space station In Dynamic Earth EnvironmentsRemote Sensing Observations from Shuttle-Mir Missions edited by K P Lulla andL V Dessinov John (New York John Wiley amp Sons) pp 1ndash14
Helfert M R and Wood C A 1989 The NASA Space Shuttle Earth Observations OYceGeocarto International 4 (1) 15ndash23
Helfert M R Mohler R R J and Giardino J R 1990 Measurement of areal uctu-ations of Great Salt Lake Utah using recti ed space photography Houston GeologicalSociety Bulletin 32 16ndash19
Jensen J R 1983 Urbansuburban land use analysis In Manual of Remote Sensing 2nd ednVol II Interpretation and Applications edited by J E Estes (Falls Church VAAmerican Society of Photogrammetry) pp 1571ndash1666
Jones T D Godwin L M Wisoff P J Amsbury D L and Evans C A 1996 AstronautEarth observations during the Space Radar Laboratory missions Proceedings of theIEEE Aerospace Applications Conference 3ndash10 February 1996 Snowmass at AspenColorado (Piscataway NJ Institute of Electrical and Electronics Engineers) Vol 2pp 29ndash46
Light D L 1980 Satellite photogrammetry In Manual of Photogrammetry 4th edn editedby C C Slama (Falls Church VA American Society of Photogrammetry) pp 883ndash977
Light D L 1993 The National Aerial Photography Program as a geographic informationsystem resource Photogrammetric Engineering and Remote Sensing 59 61ndash65
Light D L 1996 Film cameras or digital sensors The challenge for aerial imagingPhotogrammetric Engineering and Remote Sensing 62 285ndash291
Lisheron M 6 March 2000 Jimmy Luecke thinks big Austin American-Statesman E1 E6Lowman P D Jr 1980 The evolution of geological space photography In Remote Sensing
in Geology edited by B S Siegal and A R Gillespie (New York Wiley and Sons)pp 90ndash115
Lowman P D Jr 1985 Geology from space A brief history of geological remote sensingIn Geologists and Ideas A History of North American Geology edited by E T Drakeand W M Jordan (Boulder CO Geological Society of America) pp 481ndash519
Lowman P D Jr 1999 Landsat and Apollo the forgotten legacy PhotogrammetricEngineering and Remote Sensing 65 1143ndash1147
Lowman P D Jr and Tiedemann H A 1971 T errain Photography from Gemini SpacecraftFinal Geologic Report Report X-644-71-15 (Greenbelt MD National Aeronauticsand Space Administration Goddard Space Flight Center)
Lulla K Evans C Amsbury D Wilkinson J Willis K Caruana J OrsquoNeill CRunco S McLaughlin D Gaunce M McKay M F and Trenchard M1996 The NASA Space Shuttle Earth Observations Photography Database an under-utilized resource for global environmental geosciences Environmental Geosciences3 40ndash44
Lulla K P and Helfert M R 1989 Analysis of seasonal characteristics of Sambhar SaltLake India from digitized Space Shuttle photography Geocarto International 4(1)69ndash74
Lulla K Helfert M Evans C Wilkinson M J Pitts D and Amsbury D 1993Global geologic applications of the Space Shuttle Earth Observations Photographydatabase Photogrammetric Engineering and Remote Sensing 59 1225ndash1231
Lulla K Helfert M Amsbury D Whitehead V S Evans C A Wilkinson M JRichards R N Cabana R D Shepherd W M Akers T D and Melnick
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
Astronaut photography as digital data for remote sensing 4437
B E 1991 Earth observations during Shuttle ight STS-41 Discoveryrsquos mission toplanet Earth 6ndash10 October 1990 Geocarto International 6 (1) 69ndash80
Lulla K Helfert M and Holland D 1994 The NASA Space Shuttle Earth Observationsdatabase for global change science In Remote Sensing and Global Change Scienceedited by R A Vaughan and A P Cracknell NATO ASI Series Vol I24 (BerlinSpringer-Verlag) pp 355ndash365
Lulla K and Holland S D 1993 NASA Electronic Still Camera (ESC) system used toimage the Kamchatka Volcanoes from the Space Shuttle International Journal ofRemote Sensing 14 2745ndash2746
Luman D E Stohr C and Hunt L 1997 Digital reproduction of historical aerialphotographic prints for preserving a deteriorating archive PhotogrammetricEngineering and Remote Sensing 63 1171ndash1179
McRay B Scott J and Robinson J A 2000 Technical applications of astronautorbital photography how to rectify multiple image datasets using GIS EarthObservations and Imaging Newsletter OYce of Earth Sciences Johnson SpaceCenter httpeoljscnasagovnewslettergis
Merkel R F Salmon J G and Sims B A 1985 Mission Ephemeris for the Maiden Voyageof the L arge Format Camera (L FC) aboard the Space T ransportation System (ST S)Mission 41G Job Order 84-427 JSC-20383 (Houston TX Lyndon B JohnsonSpace Center)
Mohler R R J Helfert M R and Giardino J R 1989 The decrease of Lake Chad asdocumented during twenty years of manned space ight Geocarto International4 (1) 75ndash79
Moik J P 1980 Digital Processing of Remotely Sensed Images NASA SP-431 (WashingtonDC National Aeronautics and Space Administration)
National Aeronautics and Space Administration 1974 Skylab Earth Resources DataCatalog JSC-09016 (Houston TX Lyndon B Johnson Space Center)
Office of Earth Sciences NASA-Johnson Space Center 14 Mar 2000 The Gateway toAstronaut Photography of Earth httpeoljscnasagovsseopdefaulthtm
Rasher M E and Weaver W 1990 Basic Photo Interpretation A Comprehensive Approachto Interpretation of Vertical Aerial Photography for Natural Resource Applications (FortWorth TX United States Department of Agriculture Soil Conservation ServiceNational Cartographic Center)
Ring N and Eyre L A 1983 Manned spacecraft imagery In Introduction to RemoteSensing of the Environment 2nd edn edited by B F Richason Jr (Dubuque IowaKendallHunt Publishing Company) pp 112ndash129
Robinson J A Feldman G C Kuring N Franz B Green E Noordeloos M andStumpf R P 2000a Data fusion in coral reef mapping working at multiple scaleswith SeaWiFS and astronaut photography Proceedings of the 6th InternationalConference on Remote Sensing for Marine and Coastal Environments Vol 2pp 473ndash483
Robinson J A Holland S D Runco S K Pitts D E Whitehead V S andAndrersquofouet S 2000b High De nition Television (HDTV) Images for EarthObservations and Earth Science Applications NASA TP-2000-210189 (HoustonTX Lyndon B Johnson Space Center) Also available at httpeoljscnasagovnewsletterHDTV
Robinson J A McRay B and Lulla K P 2000c Twenty-eight years of urban growthin North America quanti ed by analysis of photographs from Apollo Skylab andShuttle-Mir In Dynamic Earth Environments Remote Sensing Observations fromShuttle-Mir Missions edited by K P Lulla and L V Dessinov (New York JohnWiley amp Sons) pp 25ndash42 262 269ndash270
Robinson J A Lulla K P Kashiwagi M Suzuki M Nellis M D Bussing C ELee Long W J and McKenzie L J In press Astronaut-acquired orbital photo-graphy as a data source for conservation applications across multiple geographicscales Conservation Biology
Scott K P 2000 International Space Station L aboratory Science W indow Optical Propertiesand Wavefront Veri cation T est Results ATR-2000 (2112)-1 (Los Angeles CAAerospace Corporation)
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101
Astronaut photography as digital data for remote sensing4438
Simonett D S 1983 The development and principles of remote sensing In Manual of RemoteSensing 2nd edn Vol I Theory Instruments and Techniques edited by D S Simonett(Falls Church VA American Society of Photogrammetry) pp 1ndash35
Sinnott R W 1984 Virtues of the haversine Sky and T elescope 68 159Slater P N 1980 Remote Sensing Optics and Optical Systems (Reading MA Addison-
Wesley)Slater P N Doyle F J Fritz N L and Welch R 1983 Photographic systems for
remote sensing In Manual of Remote Sensing 2nd edn Vol I Theory Instrumentsand Techniques edited by D S Simonett (Falls Church VA American Society ofPhotogrammetry) p 231ndash291
Slater R 1996 USRussian Joint Film T est NASA Technical Memorandum 104817(Houston TX Lyndon B Johnson Space Center)
Smith J T and Anson A editors 1968 Manual of Color Aerial Photography (Falls ChurchVA American Society of Photogrammetry)
Snyder J P 1987 Map ProjectionsmdashA Working Manual US Geological Survey ProfessionalPaper 1395 (Washington DC US Government Printing OYce)
Townshend J R G 1980 T he Spatial Resolving Power of Earth Resources Satellites AReview NASA Technical Memorandum 82020 (Greenbelt Maryland Goddard SpaceFlight Center)
Underwood R W 1967 Space photography In Gemini Summary Conference NASA SP-138(Houston TX Manned Spacecraft Center National Aeronautics and SpaceAdministration) pp 231ndash290
Webb E L Evangelista Ma A and Robinson J A 2000 Digital land use classi cationusing Space Shuttle acquired orbital photographs a quantitative comparisonwith Landsat TM imagery of a coastal environment Chanthaburi ThailandPhotogrammetric Engineering and Remote Sensing 66 1439ndash1449
Welch R 1982 Spatial resolution requirements for urban studies International Journal ofRemote Sensing 3 139ndash146
Wilmarth V R Kaltenbach J L and Lenoir W B editors 1977 Skylab Exploresthe Earth NASA SP-380 (Washington DC National Aeronautics and SpaceAdministration)
Wong K W 1980 Basic mathematics of photogrammetry In Manual of Photogrammetry4th edn edited by C C Slama (Falls Church VA American Society ofPhotogrammetry) pp 37ndash101