Passive Millimeter-Wave Dual-Polarization Imagers

2042 IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 61, NO. 7, JULY 2012

Passive Millimeter-Wave Dual-Polarization ImagersShaolin Liao, Nachappa Gopalsami, Senior Member, IEEE, Thomas W. Elmer, II, Member, IEEE, Eugene R. Koehl,

Alexander Heifetz, Keenan Avers, Eric Dieckman, and Apostolos C. Raptis, Life Member, IEEE

Abstract—We have developed two passive millimeter-wave im-agers for terrestrial remote sensing: one is an integrated imagingand spectroscopy system in the 146–154-GHz range with 16 chan-nels of 500-MHz bandwidth each, and the other is a single-channeldual-polarized imaging radiometer in the 70–100-GHz range. Theimaging in both systems is implemented through translation of a15-cm Gaussian dielectric imaging lens. We compared the outdoorimages of objects such as car, vegetation, sky, and ground by boththe systems under various weather conditions, including clear,cloudy, and rainy times. A ray-tracing simulation with radiativetransfer equation was used to quantify the polarization diversityof the acquired images.

Index Terms—Imaging, millimeter-wave (mmW), polarization,radiometer.

I. INTRODUCTION

PASSIVE millimeter-wave (mmW) imaging has experi-enced dramatic growth in the past decade [1]–[8], due

to its applications in terrestrial remote sensing, radiastronomy,and airport security. The main advantage of passive mmWimaging is that it can provide ground target information underall weather conditions; optical systems (visible and IR), on theother hand, require clear atmospheric conditions for reliableoperation. For example, the atmospheric attenuation in therange of mmW frequencies is 0.07 to 3 dB/km in drizzle andfog conditions, whereas it is one to three orders of magnitudehigher at optical frequencies (exceeding 100 dB/km in foggyconditions) [5], [6]. Image contrast of millimeter waves in out-door environments is increased by cold sky-reflected radiation.For example, the apparent temperature of the sky at 94 GHz is70 K in comparison to 220 K at infrared wavelengths. Becauseof reflectivity variations of common objects for millimeterwaves (metal ∼1, water 0.6, wood 0.6, and concrete 0.2), themmW sky-reflected radiation offers better thermal contrast thanat infrared wavelengths. More importantly, signal “washouts”do not occur since the apparent temperature between the back-ground and the object will rarely be equal.

Manuscript received September 9, 2011; revised November 17, 2011; ac-cepted November 19, 2011. Date of publication February 6, 2012; date ofcurrent version June 8, 2012. This work was supported by the Office of Non-proliferation and Verification Research and Development under the NationalNuclear Security Administration (NNSA) of the U.S. Department of Energyunder Contract No. DE-AC02-06CH11357. The Associate Editor coordinatingthe review process for this paper was Dr. Sergey Kharkovsky.

S. Liao, N. Gopalsami, T. W. Elmer, II, E. R. Koehl, A. Heifetz, andA. C. Raptis are with the Argonne National Laboratory, Argonne, IL 60439USA (e-mail: [email protected]).

K. Avers is with the Department of Physics, University of Illinois at Chicago(UIC), Chicago, IL 60607 USA.

E. Dieckman is with the Department of Physics, Purdue University, WestLafayette, IN 47907 USA.

Color versions of one or more of the figures in this paper are available onlineat http://ieeexplore.ieee.org.

Digital Object Identifier 10.1109/TIM.2012.2183032

Fig. 1. Architecture of a 16-channel 146–154-GHz radiometer with 500-MHzbandwidth for each channel.

In addition, polarization diversity of passive mmW radia-tion may be used for enhanced discrimination of targets andsurfaces. A common example is the change in apparent tem-perature of the sea surface between horizontal and verticalpolarizations, depending on wind speed [4], [7], [8]. In thispaper, we present results of horizontal and vertical polarizationimages from common ground targets such as car, ground,and trees. A ray-tracing simulation is used to interpret thepolarization diversity of the acquired images. We also presentresults of mmW imaging under various weather conditions attwo frequency bands: a 16-channel radiometer in the 146–154-GHz range with 500-MHz bandwidth for each channeland a single-channel dual-polarization radiometer in the 70–100-GHz range.

This paper is organized as follows. The architectures of thetwo mmW radiometers are presented in Section II, followed bya description of the experimental imaging setup in Section III,presentation of imaging results, analysis and discussion inSection IV, and conclusions in Section V.

II. MMW RADIOMETER ARCHITECTURES

A. 16-Channel 146–154-GHz Radiometer

The architecture of our 16-channel radiometer [9] is shownin Fig. 1. It consists of a circular corrugated horn antenna,followed by a 146–154-GHz bandpass filter (BPF), a mixingstage, a 164-GHz local oscillator obtained by doubling an82-GHz Gunn oscillator signal, an attenuator, a three-stage am-plifier chain, and a 16-channel filter bank. A chopper is used infront of the scalar horn to collect Dicke-switched data betweena reference signal from an absorber on the chopper blades andthe target scene viewed through an opening between chopperblades. The detected signals from 16 channels are collected by

0018-9456/$31.00 © 2012 IEEE

LIAO et al.: PASSIVE MILLIMETER-WAVE DUAL-POLARIZATION IMAGERS 2043

Fig. 2. Imaging and spectroscopy conducted on a gas cell with acetonitrile at different pressures during image formation (vacuum at row 1, 2 torr at rows 2–4,4 torr at rows 5–6, 11 torr at rows 7–8, 740 torr at row 9, and vacuum at row 10).

a data acquisition (DAQ) board interfaced to a laptop. To obtain2-D images, a 15-cm Gaussian lens was raster scanned againstthe radiometer antenna positioned at the imaging plane of thelens. The use of 16 channels allows for detection of spectralsignatures [10] while raster scanning a target for imaging.

Fig. 2 gives, for example, results of imaging and spec-troscopy conducted on a gas cell (7.5-cm ID, 8-cm OD, and1 m long) containing acetonitrile (CH3CN) vapor at differentpressures against a liquid nitrogen background. The solid por-tion of the gas cell has no spectral features, but acetonitrilehas rotational spectra in the 146–154-GHz range [11]. A 10 ×10 image was obtained across 15 spectral channels using 4-sintegration time per pixel. The scene was changed at the endof rows; the sequence of events consisted of i) normalizing all15 channels using cold (absorber immersed in liquid nitrogen)and hot (absorber at ambient temperature) loads, ii) vacuum (nogas) in the first row, iii) pure CH3CN at 2 torr in rows 2–4,iv) pure CH3CN at 4 torr in rows 5 and 6, v) pure CH3CN at11 torr in rows 7 and 8, vi) air added to 11-torr CH3CN tobring the total pressure to 740 torr in row 9, and vii) vacuumin row 10. The composite graph in Fig. 2 consists of individualimages for 15 spectral channels at the top and spectral plotsof acetonitrile at the bottom. The images of the channels arenumbered 1 to 15 starting at the top row of images, whichcorrespond to a center frequency of 146.75 GHz for Ch #1 andincreasing at a 0.5-GHz interval to 153.75 GHz for Ch #15. Theimages show the circular part of the gas cell wall in all channels.Note that the molecular rotational spectrum of CH3CN peaksaround 146 GHz; hence, the image of CH3CN is predominantlyseen in the lower numbered channels depending on the extent ofpressure broadening. The 11-torr image is brighter than that ofthe 4-torr because of increased concentration, and the 740-torrimage is weaker because of dilution with air, and it is spreadinto more channels starting from Ch #1 because of pressurebroadening. The spectral plots include a simulated line at 4 torrfrom the Jet Propulsion Laboratory spectral database [12] andthe measured spectral plots at pressures 4, 11, and 740 torr. Thesimulated and measured data at 4 torr agree very well; the errorbar is the standard deviation of intensity data from the pixels ofsame pressure. The pressure broadening effect is seen from theplots for 11 and 740 torr.

Fig. 3. Architecture of a dual-polarization 70–100-GHz radiometer.

Fig. 4. Experimental setup for mmW imaging of outdoor scenes.

B. 70–100-GHz Dual-Polarization Radiometer

The architecture of our 70–100-GHz dual-polarization ra-diometer (Millitech MCA-10-100512) is shown in Fig. 3. Itconsists of a wire grid polarizer that splits the incoming signalinto vertically and horizontally polarized signal paths. Thesignal in each path is received by a circular corrugated antennaand undergoes an identical signal processing chain comprisedof a 70–100-GHz BPF, a low noise amplifier (LNA), a Schottkybarrier diode detector, and a baseband video amplifier. Thebaseband signals are collected by a DAQ interfaced to a laptop.As in the 16-channel radiometer, a chopper may be used tocollect Dicke-switched data, and the same 15-cm Gaussian lensis raster scanned to obtain 2-D images.

III. EXPERIMENTAL SETUP

The experimental setup for our outdoor imaging is shownin Fig. 4. A LabVIEW program running on a laptop is used


to collect and process the image data. The ray optics of thepassive mmW imaging system was simulated by commercialsoftware Zemax to obtain optimal imaging parameters such asfocal length and the distance between the horn antenna andthe imaging lens, the details of which will be presented in theDiscussion Section, together with determining the point spreadfunction (PSF) for both radiometers.

IV. IMAGING RESULTS, ANALYSIS AND DISCUSSION

A. SNR Performance

For a Dicke-switched radiometer, the minimum de-tectable temperature or noise-equivalent temperature difference(NETD) is given by [13]

NETD = (TA + TR)

√(1

Bτ/2

)2

+(

ΔG

G

)2

∼ 2(TA + TR)/√

Bτ (1)

where TA and TR are the antenna and the receiver temperatures,respectively; B is the bandwidth of the radiometer; τ/2 is theintegration time; and G and ΔG are the system gain and itsvariation. In (1), we have used the approximation ΔG/G � 1,and the factor of 2 comes from the fact that a Dicke-switchedradiometer only acquires signal at half the measurement timeinterval, τ . For our 16-channel 146–154-GHz radiometer [9],[10], TA = 290 K, TR ∼ 4492 K, B = 500 MHz for eachchannel, and if we assume the integration time τ = 1 s, wehave NETD = 0.4 K; for our dual-polarization 70–100-GHzradiometer, TA = 290 K, TR ∼ 3000 K, B = 30 GHz for eachchannel, and if we assume the integration time τ = 1 s, we getNEDT = 0.04 K. Obviously, the dual-polarization radiometeris more sensitive due to two factors: larger bandwidth and theuse of LNA at the front end.

B. Radiative Transfer

The radiative signal from a solid object comes from twosources, namely, the blackbody emission (horizontal/verticalpolarization) of the object itself I

H/Vo and the power reflec-

tion coefficient RH/V (defined as the square of the ampli-tude reflection coefficient) from the sky background emissionRH/V I

H/Vb . Both signals are subjected to the same absorption

e−τo from the object position zo to the radiometer since they arealong an identical path. The expression of the detected signalincluding atmospheric emission is given by [9]

IH/V =(I

H/V0 + RH/V I

H/Vb

)e−τo

+

∞∫0

κa(z)T (z)e−τ(z)dz

τ(z) =

z∫0

κa(z)dz, τo =

zo∫0

κa(z)dz (2)

where qκa is the absorption coefficient, and T (z) is the temper-ature along the wave path at altitude z. Rigorously, all quantitiesin (2) depend on the frequency. However, in the frequency range

Fig. 5. Image of the sum of all 16 channels of a 2001 white MitsubishiMontero Sport car obtained by the 146–154-GHz radiometer (in K).

of interest, the spectral changes are negligible for solid targets.The sky background emission may be slightly polarized, de-pending on the time of day and weather conditions (e.g., sunny,rainy, foggy, etc.). However, the most polarization contrastcomes from the polarization dependent reflection RH/V sincethe electromagnetic wave reflection from the interface betweentwo dielectric materials is very different for horizontal andvertical polarizations. This can be easily understood by notingthat there is no reflection for a vertically polarized wave at itsBrewster angle of incidence, whereas there is no such magicangle for horizontal polarization [14]. Hence, we would expectthere will be polarization contrast when there is a reflectionevent involving a dielectric interface along the optical path ofthe imaging system, e.g., the reflection from the ground, the carwindshield, the clouds in the sky, etc. In principle, the measuredterrestrial image can be simulated using (2), with the help of theray-tracing method.

C. Imaging Results and Analysis

The imaging scene, as shown in Fig. 4, typically consistedof a parked car at ∼30 feet from the radiometer against thebackground of a dumpster(s), bushes, trees, and cold sky. Wepresent imaging results of the above scene in both clear andrainy conditions and analyze the polarization contrast of theimages.

1) Clear and Cloudy Day Images: Because of longer wave-length than of optics, mmW images of the scene during clear,cloudy, or day or night times showed no observable changes.Fig. 5 shows the image of the scene with the 146–154-GHzradiometer with 2-s integration time (all channels summed up,which is equivalent to 8 GHz BW), and Fig. 6 shows the dual-polarization image with the 70–100-GHz radiometer with 1-sintegration time taken on a fair weather day. The 146–154-GHzradiometer, being at a lower wavelength (2 mm), has a higherresolution than the (3 mm) 70–100-GHz radiometer. We canclearly see a part of the dumpster, the car wheels, and the


Fig. 6. (Left) Horizontal and (right) vertical polarization images of a 2001 red Nissan Altima car obtained by the dual-polarization 70–100-GHz radiometer (in K).

Fig. 7. Image scene analysis using the ray-tracing method.

car windows in Fig. 5, whereas these are not so obviousin Fig. 6. However, due to the narrower bandwidth of the16-channel radiometer (8 GHz compared with 30 GHz of thedual-polarization radiometer), it takes about twice the integra-tion time to obtain a similar SNR.

Typical acquisition time for a 50 × 50 image with 1-sintegration time is approximately 1 h. This is typical of scannedsingle pixel imaging systems, but one may employ compressivesensing to speed up the imaging time by a factor of 10 or more[15]. The advantage of single pixel imaging compared withvideo rate imaging using an array of detectors is that it offers ahigh SNR and allows for the use of one-of-a-kind sophisticateddetectors such as the 146–154-GHz imaging and spectroscopysystem we have built.

The polarization changes in the images may be analyzedusing the ray-tracing method. Fig. 7 classifies the image sceneinto five regions for ray-tracing analysis: #1, the image of thetrees behind the car, which consists of the tree emission andthe reflection of the sky background emission by the tree; #2,

Fig. 8. (Left) Horizontal-polarization image and (right) polarization dif-ference image overlaid on the optical images of the car. The horizontal-polarization image is the same as in Fig. 6 with a different color map, andthe polarization difference image is the subtraction of the vertical-polarizationimage from the horizontal-polarization image in Fig. 6. Also shown are fivedifferent regions that correspond to those analyzed in Fig. 7.

reflection of the sky background emission by the upper partof the car’s front side; #3, reflection of the sky backgroundemission by the ground and then by the lower part of the car’sfront side; #4, reflection of the sky background emission by theupper part of the car’s front side and then by the ground; and#5, reflection of the trees by the ground.


Fig. 9. Measured reflection power coefficients of both (left) vertical and (right) horizontal polarizations.

Fig. 10. Ratio of the vertical power reflection coefficient to the horizontal power reflection coefficient ratio ≡ RV /RH : on the left are the measured values andon the right are the simulated values.

Fig. 8 shows (a) the horizontal polarization image overlaid onthe optical image of the car and (b) the polarization differenceimage. The corresponding five regions of Fig. 7 are marked inFig. 8(a) and (b). The trees behind the car (region #1) emitunpolarized radiation at outdoor temperature, which is muchhigher than the reflection of the sky background emission.Hence, we can consider that the trees image has almost thesame intensity for both polarizations. Because the car frontsurface (region #2) is made of metal, its emission is almostzero and its reflection of the sky background emission is notsensitive to polarization. The sky radiation reflected by theground and bottom part of the car (region #3) shows smallchanges in polarization because of small incidence angles. Onthe other hand, the sky reflection from the ground (region#4) causes strong polarization contrast because of differentreflection coefficients for vertical and horizontal polarizationsat such large incidence angles. The ground-reflected radiationof trees (region #5) offers less polarization contrast due tosmaller temperature contrast.

From the geometry of the scene, we can simulate the sceneimage and compare it with the measured image. We calculate

the reflections for both polarizations from the measurementdata and compare them with theoretical calculation. Fig. 9shows the measured reflections for both polarizations, whichare calculated for the car reflection by the ground accord-ing to (2), ignoring the relatively small absorption coefficientκa, i.e.,

RH/V = IH/V (#4) − IGround/IH/V (#2) (3)

where IGround is the ground emission. We can clearly seethe contour of the car and detailed features such as the carwindows with the help of nonlinear operation of (3), whereasthey were not so obvious in the raw images in Fig. 6. In addi-tion, Fig. 10 shows the ratio of vertical to horizontal reflectionratio ≡ RV /RH and its comparison to the theoretical value[14]. For theoretical calculation, we have used the dry groundpermittivity εGround = 3.5 and the distance from the antenna tothe Gaussian lens di = 7.5′′. The incident angle of region #4 inFig. 6 is given by

θi(#4) = arctan[√

x2(#4) + y2(#4)/di

]. (4)


Fig. 11. Dual-polarization images (top) during rainfall and (bottom) after rainfall show different polarization contrast for the 2001 Nissan Altima car used inFig. 6.

Fig. 12. Power reflection ratio between vertical and horizontal polarizations: (left) during rainfall and (right) after rainfall.

Now, the horizontal and vertical reflection coefficients can beobtained as [14]

RH(#4) = |η2 cos θi − η1 cos θt/η2 cos θi + η1 cos θt|2RV (#4) = |η2 cos θt − η1 cos θi/η2 cos θt + η1 cos θi|2

sin(θt) = γAir/γGround sin(θi), γ ≡ jω√

μ0ε0(ε′ − jε′′)(5)

where η1 and η2 are the wave impedance values of air andground, respectively. In addition, the Brewster angle of theground is θB = arctan[

√εGround] ∼ 62◦ for the vertical polar-

ization, which is smaller than the incident angle θi(#4) ∼ 80◦,as calculated from (4). Note that only the metallic part of the caris shown in Figs. 9 and 10 since nonmetallic dielectric materialssuch as car windows differently behave, i.e., instead of com-pletely reflecting all polarizations, dielectric materials reflectpolarized wave differently and is thus polarization sensitive.

Fig. 13. Zemax simulation of a 15-cm imaging lens shows the focal length,the optical depth, and the PSF at 100 GHz.


Fig. 14. Simulated PSF using Zemax at 100 GHz (the dual-polarization radiometer) and at 150 GHz (the 16-channel radiometer).

2) Rainy Day Images: The top and bottom images in Fig. 11give the dual-polarization images during moderate rain andafter the rain stopped, respectively. Clearly, the temperaturesof the trees and ground are higher during rainfall. This isbecause the mmW emission of rain droplets is much higher thanthat of the sky. Furthermore, in Fig. 12, we give the imagesof the power reflection ratio between vertical and horizontalpolarizations according to (3). One can see that the polarizationcontrast is higher during rainfall (which is RV /RH ∼ 0.35)compared with that after the rain stopped (which is RV /RH ∼0.2). This is because the rain changes the dielectric constant ofthe ground and thus the power reflection according to (5).

V. DISCUSSION

From the images obtained with both radiometers, it is clearthat the 146–154-GHz radiometer provides higher resolutionthan that of the dual-polarization 70–100-GHz radiometer be-cause of its shorter wavelength. This may be quantitativelyverified using the ray-tracing software Zemax. Fig. 13 showsthe ray-tracing simulation of a 15-cm lens at 100 GHz. Thefocal length determined from the simulation is: f ∼ 16.5 cmat 100 GHz and f ∼ 16.76 cm at 150 GHz. The distancefrom the detecting antenna to the imaging lens, di, followsthe standard imaging formula 1/f = 1/di + 1/do, with do

being the distance from the imaging lens to the scene. Sincethe resolution of an image is the convolution of the scenewith the PSF, we can obtain the PSFs for both cases. Fig. 14shows the optical transfer function (OTF) obtained at 100 GHz(or 3-mm wavelength) for the dual-polarization radiometerand at 150 GHz (or 2-mm wavelength) for the 16-channelradiometer. The full width at half maximum (FWHM) for thetwo frequencies are: FWHM ∼4 mm at 100 GHz and FWHM∼3 mm at 150 GHz. The PSF is closely related to the OTF;the PSF and the modulation transfer function, which is themagnitude of the OTF, are a Fourier transform pair, PSF =IFT{MFT} = IFT{|OTF|} [16]. The knowledge of the PSF for

the imagers will be useful in the image analysis and recognitionof targets in cluttered environment.

VI. CONCLUSION

We have developed two radiometers for terrestrial imagingand spectroscopic application, namely, the 16-channel 146–154-GHz radiometer and the dual-polarization 70–100-GHzradiometer. Zemax optics software was used to optimize thepassive mmW imaging optics and to determine the PSF of theimagers. Images of an outdoor scene consisting of a car, dump-sters, trees, and the sky were taken under different weatherconditions to compare their performance. It has been shownthat the 16-channel radiometer has higher resolution due to itsshorter wavelength. However, its SNR is lower than that of thedual-polarization radiometer due to the narrower bandwidth;hence, longer integration time is required to obtain images ofthe same quality. The polarization contrast is mainly due tothe reflection from the interface of two dielectric materialssuch as the ground, the car windshield, and the clouds in theatmosphere. The polarization changes from the vertical and hor-izontal reflections of the sky radiation by the car and the groundwere analyzed using a combination of ray-tracing and radiativetransfer simulation. The shadow of the car by the ground, whichwas not obvious in the raw image, became evident with theabove analysis. We also extended the analysis to determinethe polarization changes of the scene under rainy conditions.These results show that the ray-tracing method in combinationwith radiative transfer equations is a powerful tool for scenesimulation and image analysis for passive mmW systems.

REFERENCES

[1] M. A. Janssen, Ed., Atmospheric Remote Sensing by Microwave Radiom-etry. New York: Wiley, 1993.

[2] J. J. Lynch, P. A. Macdonald, H. P. Moyer, and R. G. Nagele, “Passivemillimeter wave imaging sensors for commercial markets,” Appl. Opt.,vol. 49, no. 19, pp. E7–E12, Jul. 2010.


[3] M. R. Fetterman, J. Grata, J. G. Jubic, W. L. Kiser, Jr, and A. Visnansky,“Simulation, acquisition, and analysis of passive millimeter-wave imagesin remote sensing applications,” Opt. Exp., vol. 16, no. 25, pp. 20 503–20 515, Dec. 2008.

[4] M. Peichl, S. Dill, M. Jirousek, J. W. Anthony, and H. Sub, “Fully-polarimetric passive mmW imaging systems for security applications,”in Proc. SPIE, 2010, vol. 7837, p. 783 70C.

[5] L. Yujiri, M. Shoucri, and P. Moffa, “Passive millimeter-wave imaging,”IEEE Microw. Mag., vol. 4, no. 3, pp. 39–50, Sep. 2003.

[6] R. Appleby and R. N. Anderson, “Millimeter-wave and submillimeter-wave imaging for security and surveillance,” Proc. IEEE, vol. 95, no. 8,pp. 1683–1690, Aug. 2007.

[7] N. Skou and D. M. Le Vine, Microwave Radiometer Systems: Design andAnalysis. Norwood, MA: Artech House, 2006.

[8] C. W. O’Dell, D. S. Swetz, and P. T. Timbie, “Calibration of millimeterwave polarimeters using a thin dielectric sheet,” IEEE Trans. Microw.Theory Tech., vol. 50, no. 9, pp. 2135–2141, Sep. 2002.

[9] N. Gopalsami, S. Bakhtiari, T. W. Elmer, and A. C. Raptis, “Application ofmillimeter-wave radiometry for remote chemical detection,” IEEE Trans.Microw. Theory Tech., vol. 56, no. 3, pp. 700–709, Mar. 2008.

[10] N. Gopalsami, S. Liao, E. R. Koehl, T. W. Elmer, A. Heifetz,H. T. Chien, and A. C. Raptis, “Passive millimeter wave imaging spec-troscopy system for terrestrial remote sensing,” in Proc. SPIE, 2010,vol. 7670, pp. 767 003-1–767 003-7.

[11] N. Gopalsami and A. C. Raptis, “Millimeter-wave radar sensing of air-borne chemicals,” IEEE Trans. Microw. Theory Tech., vol. 49, no. 4,pp. 646–653, Apr. 2001.

[12] R. L. Poynter and H. M. Picket, Submillimeter, Millimeter, and MicrowaveSpectral Line Catalog. Pasadena, CA: Jet Propulsion Lab., 1983.

[13] M. Sato, T. Hirose, and K. Mizuno, “Advanced MMIC receiver for94-GHz band passive millimeter-wave imager,” IEICE Trans. Electron.,vol. E92. C, no. 9, pp. 1124–1129, 2009.

[14] C. A. Balanis, Advanced Engineering Electromagnetics. New York:Wiley, 1998.

[15] N. Gopalsami, T. W. Elmer, S. Liao, R. Ahern, A. Heifetz, A. C. Raptis,M. Luessi, D. Babacan, and A. K. Katsaggelos, “Compressive samplingin passive millimeter-wave imaging,” in Proc. SPIE, 2011, vol. 8022,pp. 802 201-1–802 201-6.

[16] J. W. Goodman, Introduction to Fourier Optics. New York:McGraw-Hill, 1996.

Shaolin Liao received the B.S. degree in materialsscience and engineering from Tsinghua University,Beijing, China, in July 2000 and the M.S. degree inmaterial science, the M.S. degree in electrical andcomputer engineering, and the Ph.D. degree in elec-trical and computer engineering from the Universityof Wisconsin at Madison, Madison, in August 2003,December 2005, and May 2008, respectively.

He held the Research Fellow position in thePhysics Department, Queens College, City Univer-sity of New York, New York, from May 2008 to

January 2010. He is currently with Argonne National Laboratory, Argonne,IL, doing microwave/millimeter-wave/THz-wave/laser research. He has beeninterested in both active and passive millimeter-wave applications, such asremote sensing, Nondestructive Examination and national security, in bothexperiment and theory; high-power microwave transmission and conversion;gyrotron component design; microwave beam shaping technique; theory andalgorithm development for high-performance simulation of electromagneticwave propagation and scattering; antenna design and simulation; as well asmicrowave and light multiple scattering and Anderson localization insiderandom media.

Nachappa “Sami” Gopalsami (SM’95) receivedthe B.E. and M.S. degrees in electrical engineer-ing from the University of Madras, Chennai, India,and the Ph.D. degree in electrical engineering andcomputer science from the University of Illinois,Chicago.

He joined Argonne National Laboratory, Argonne,IL, in 1980, where he is currently a Senior ElectricalEngineer in the Sensors and Instrumentation sectionof the Nuclear Engineering Division. His currentresearch interests include development of radio fre-

quency, microwave, millimeter-wave, and terahertz sensors and imaging sys-tems for national security, biosensing, environmental monitoring, and materialsapplications. He has published over 150 technical papers in the area of sensorsand Nondestructive Evaluation (NDE) and has seven U.S. patents to his credit.

Dr. Gopalsami has received two R&D 100 awards from the R&D Magazine,both on millimeter-wave sensors in 1986 and 2007; an outstanding paper awardfrom the American Society of Nondestructive Testing; and an OutstandingMentor award from the Office of Science Undergraduate Research Programs.He is a member of Sigma Xi and Society of Photo-optical InstrumentationEngineers (SPIE).

Thomas W. Elmer, II (M’07) received the B.S.degree in physics with minors in math and computerscience from La Sierra University, Riverside, CA,in 1998 and the M.S. degree in physics from theUniversity of Illinois, Chicago, in 2004.

While at La Sierra, he worked for the PhysicsDepartment, writing and maintaining programs torun laboratory experiments. He has also lectured onAstronomy and Gravitational Physics for the depart-ment. In 1999, he joined Argonne National Labo-ratory, Argonne, IL, as a Student Intern, eventually

staying as a Software Engineering Associate for the System Technologies& Diagnostics Department of the Nuclear Engineering Division. He writesprogramming for modeling, motion control, data acquisition, and data analysisin the microwave, millimeter-wave, and terahertz sensors labs.

Mr. Elmer received an R&D 100 award from the R&D Magazine in 2007.

Eugene R. Koehl received the B.A. degree inphysics from Lewis University, Romeoville, IL, in1969 and the B.S. degree in electrical engineeringfrom Midwest College of Engineering (Illinois Insti-tute of Technology), Chicago, IL, in 1981.

He is an Electrical Engineer at Argonne NationalLaboratory, Argonne, IL, with 37 years of experiencein the design, fabrication, instrumentation, control,and automated data acquisition of experiments andfacilities for testing the design and characteristics ofsensors, power systems, and Liquid Metal Reactor

(LMR) components. His activities cover a broad range of physical sciences,including acoustic signal transport, X-ray inspection and detection, high andprocess vacuum, alkali metal chemistry and heat transport, and satellite trackingand millimeter-wave systems. He has been involved with the precision manipu-lation and nondestructive testing of engineered materials used in heavy vehicle,stationary power, and aerospace systems for the past 16 years.

Alexander Heifetz received the B.S. degree (SummaCum Laude) in applied mathematics, the M.S. degreein physics, and the Ph.D. degree in electrical engi-neering from Northwestern University, Evanston, IL.

He is an Electrical Engineer with the Nuclear En-gineering Division at Argonne National Laboratory,Argonne, IL. He came to Argonne as a Director’sPostdoctoral Fellow with the Nuclear EngineeringDivision. His research interests are in simulationsand modeling for electromagnetic engineering, nu-clear engineering, and signal processing and con-

trols. He has published over 20 peer-reviewed papers in major scientific journalsand has one provisional U.S. patent.


Keenan Avers is an Undergraduate Student at theUniversity of Illinois at Chicago (UIC), Chicago,currently working on the B.A. degree in engineeringphysics.

He joined Argonne National Laboratory, Argonne,IL, as a Student Intern, in the summer of 2010and again in the summer of 2011. He assisted theMillimeter Wave group in assembling and operatingthe 70–100-GHz dual-polarization radiometer. He iscurrently working for the Nanoscale Physics groupin the UIC Physics Department, where he assists in

the preparation of thin-film oxide samples.

Eric Dieckman is an Undergraduate Student atPurdue University, West Lafayette, IN, majoring inphysics.

He joined Argonne National Laboratory, Argonne,IL, in the summer of 2011, as a Student Intern, wherehe worked on the Zemax simulation software formillimeter-wave optics.

Apostolos C. Raptis (S’66–M’67–LM’06) receivedthe Ph.D. degree in electrical engineering from theUniversity of Akron, Akron, OH, in 1973.

He is a Senior Electrical Engineer at ArgonneNational Laboratory (ANL), Argonne, IL, with35 years of experience in teaching and in researchand development. His areas of expertise includesensors, instrumentation and controls, nondestructiveevaluation, data processing, electromagnetics, plas-mas, optics, acoustics, and geophysical exploration.He is presently the Department Manager for Systems

Technologies and Diagnostics in the Nuclear Engineering Division. At ANL,he initiated and helped develop ANL’s I&NDE (instrumentation and NDE)programs for nuclear and fossil energy, conservation and arms controls. Heis responsible for work on national security for remote detection of chemical,biological, and nuclear agents and explosives. In addition, he has been a strongparticipant in collaborative work with the University of Chicago, Chicago, IL,in the development of the Bioengineering initiative. He is the author of morethan 250 publications and holder of 17 patents.

Mr. Raptis was the recipient of the 1994, 1996, 2007, and 2011 R&D100 awards and the 1994 American Society of Nondestructive Testing (ASNT)best paper award.

Passive Millimeter-Wave Dual-Polarization Imagers

Documents

Transcript of Passive Millimeter-Wave Dual-Polarization Imagers