M. Soumekh

University at Buffalo, The State University of New York, Buffalo, New York, United States

Are you M. Soumekh?

Claim your profile

Publications (92)96.77 Total impact

  • [Show abstract] [Hide abstract]
    ABSTRACT: We provide an evaluation of spectral features extracted from the signal return of a forward-looking ground penetrating radar to improve the detection performance of buried explosive hazards. The evaluations are performed on data collected at two different lanes at a government test site. The performance of the one-dimensional (1D), two-dimensional (2D) and multiple (ML) spectral features will be contrasted through lane-based cross-validation for training and testing. Additional features to characterize the spectral behaviors of the forward-looking radar return will also be examined.
  • [Show abstract] [Hide abstract]
    ABSTRACT: This paper proposes an effective anomaly detection algorithm for forward-looking ground-penetrating radar (FLGPR). The challenges in detecting explosive hazards with FLGPR are that there are multiple types of targets buried at different depths in a highly-cluttered environment. A wide array of target and clutter signatures exist, which makes classifier design difficult. Recent work in this application has focused on fusing the classifier results from multiple frequency subband images. Each sub-band classifier is trained on suites of image features, such as histogram of oriented gradients (HOG) and local binary patterns (LBP). This prior work fused the sub-band classifiers by, first, choosing the top-ranked feature at each frequency sub-band in the training data and then accumulating the sub-band results in a confidence map. We extend this idea by employing multiple kernel learning (MKL) for feature-level fusion. MKL fuses multiple sources of information and/or kernels by learning the weights of a convex combination of kernel matrices. With this method, we are able to utilize an entire suite of features for anomaly detection, not just the top-ranked feature. Using FLGPR data collected at a US Army test site, we show that classifiers trained using MKL show better explosive hazard detection capabilities than single-kernel methods.
  • [Show abstract] [Hide abstract]
    ABSTRACT: Buried explosives have proven to be a challenging problem for which ground penetrating radar (GPR) has shown to be effective. This paper discusses an explosive hazard detection algorithm for forward looking GPR (FLGPR). The proposed algorithm uses the fast Fourier transform (FFT) to obtain spectral features of anomalies in the FLGPR imagery. Results show that the spectral characteristics of explosive hazards differ from that of background clutter and are useful for rejecting false alarms (FAs). A genetic algorithm (GA) is developed in order to select a subset of spectral features to produce a more generalized classifier. Furthermore, a GA-based K-Nearest Neighbor probability density estimator is employed in which targets and false alarms are used as training data to produce a two-class classifier. The experimental results of this paper use data collected by the US Army and show the effectiveness of spectrum based features in the detection of explosive hazards.
    Proceedings of SPIE - The International Society for Optical Engineering 01/2011; 8017:80171E. DOI:10.1117/12.884685 · 0.20 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper proposes an effective anomaly detection algorithm for a forward-looking ground-penetrating radar(FLGPR). One challenge for threat detection using FLGPR is its high dynamic range in response to different kinds of targets and clutter objects. The application of a fixed threshold for detection in a full-band radar image often yields a large number of false alarms. We propose a method that uses both narrow-band and full-band radar processing, coupled with a classifier that uses complex-valued Gabor filter responses as the features. We then fuse the narrow-band and fullband images into a composite confidence map and detect local maxima in this map to produce candidate alarm locations. Full-band radar images provide a high degree of image resolution, while narrow-band images provide a means to detect targets which have a unique narrow-band signature. Experimental results for our improved detection techniques are demonstrated on data sets collected at a US Army test site.
    Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series; 01/2011
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: In this research, we have developed an algorithm to reduce the residual artifacts of the background clutter (that is, stationary targets) that appear in the MTI imagery that are generated by Global Signal Subspace Difference (GSSD) of the monostatic and bistatic images of an along-track monopulse synthetic aperture radar (SAR) data. We have also established the theoretical foundation for estimating the motion track and parameters of the detected moving targets. We will show the results of these algorithms on measured SAR data.
    Radar Conference, 2010 IEEE; 06/2010
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Forward-looking ground-penetrating radar (FLGPR) has received a significant amount of attention for use in explosivehazards detection. A drawback to FLGPR is that it results in an excessive number of false detections. This paper presents our analysis of the explosive-hazards detection system tested by the U.S. Army Night Vision and Electronic Sensors Directorate (NVESD). The NVESD system combines an FLGPR with a visible-spectrum color camera. We present a target detection algorithm that uses a locally-adaptive detection scheme with spectrum-based features. The remaining FLGPR detections are then projected into the camera imagery and image-based features are collected. A one-class classifier is then used to reduce the number of false detections. We show that our proposed FLGPR target detection algorithm, coupled with our camera-based false alarm (FA) reduction method, is effective at reducing the number of FAs in test data collected at a US Army test facility.
    Proceedings of SPIE - The International Society for Optical Engineering 04/2010; DOI:10.1117/12.852274 · 0.20 Impact Factor
  • Tuan Ton, David Wong, Mehrdad Soumekh
    [Show abstract] [Hide abstract]
    ABSTRACT: The use of low-frequency, ultra-wideband (UWB) radar technology to help detect concealed or buried targets has been demonstrated in the past, and could provide an important capability for combat systems on the battlefield. As part of Army's mission and technical objective, the Night Vision and Electronic Sensors Directorate, Countermine Division has designed and developed a Forward-Looking Ground-Penetrating Radar (FLGPR) with standoff capability. The forward-looking GPR is an impulse-based radar system with a bandwidth that spans between 300 - 3000 MHz. The innovative design uses commercially available, off-the shelf (COTS) components to create an effective sampling scheme of approximately 8 GHz. The design is modular, and can be scaled to provide a means for future improvement of electronic components to eventually meet the requirements of various combat systems. Early results from recent tests suggest that metallic targets buried near the surface can be detected with this radar system.
  • [Show abstract] [Hide abstract]
    ABSTRACT: This paper is concerned with imaging and moving target detection using a synthetic aperture radar (SAR) platform that is known as Gotcha. The SAR platform can interrogate a scene using an imperfect circular trajectory; we refer to this as nonlinear SAR data collection. This collection can make monostatic and quasi-monostatic measurements in the along-track domain. We present subaperture-based wavefront reconstruction algorithms for motion compensation and imaging from this nonlinear SAR database. We also discuss adaptive filtering algorithms to construct MTI imagery from the two receiver channels of the system. Results will be provided.
    Radar Conference, 2009 IEEE; 06/2009
  • Mehrdad Soumekh, Tuan Ton, Pete Howard
    [Show abstract] [Hide abstract]
    ABSTRACT: The U.S. Department of Defense Humanitarian Demining (HD) Research and Development Program focuses on developing, testing, demonstrating, and validating new technology for immediate use in humanitarian demining operations around the globe. Beginning in the late 1990's, the U.S. Army Countermine Division funded the development of the NIITEK ground penetrating radar (GPR) for detection of anti-tank (AT) landmines. This work is concerned with signal processing algorithms to suppress sources of artifacts in the NIITEK GPR, and formation of three-dimensional (3D) imagery from the resultant data. We first show that the NIITEK GPR data correspond to a 3D Synthetic Aperture Radar (SAR) database. An adaptive filtering method is utilized to suppress ground return and self-induced resonance (SIR) signals that are generated by the interaction of the radar-carrying platform and the transmitted radar signal. We examine signal processing methods to improve the fidelity of imagery for this 3D SAR system using pre-processing methods that suppress Doppler aliasing as well as other side lobe leakage artifacts that are introduced by the radar radiation pattern. The algorithm, known as digital spotlighting, imposes a filtering scheme on the azimuth-compressed SAR data, and manipulates the resultant spectral data to achieve a higher PRF to suppress the Doppler aliasing. We also present the 3D version of the Fourier-based wavefront reconstruction, a computationally-efficient and approximation-free SAR imaging method, for image formation with the NIITEK 3D SAR database.
    Proceedings of SPIE - The International Society for Optical Engineering 05/2009; DOI:10.1117/12.821390 · 0.20 Impact Factor
  • Kenneth Ranney, Anthony Martone, Mehrdad Soumekh
    [Show abstract] [Hide abstract]
    ABSTRACT: Radar systems have long been recognized as an effective tool for detecting moving targets--a problem commonly referred to as moving target indication (MTI). Recent advances, including Space Time Adaptive Processing (STAP), allow for even more precise determination of a target's location relative to the radar. Still, most of these methods approach MTI from the point of view of parameter estimation, and this sort of an approach can become problematic when the target speed is low and its associated Doppler frequency is near zero. In such cases the target signature is masked by the stationary, background clutter. Another potential drawback to STAP techniques arises from the fact that they require a relatively large number of receive channels, adding additional complexity to the radar system hardware. In this paper we present a moving-target-indication (MTI) technique that is based on a change detection paradigm. That is, rather than estimating the Doppler frequency associated with a target's motion, we propose to detect subtle differences between simultaneously collected, complex SAR images. We use simulated data to illustrate the feasibility of the approach under several different operating scenarios.
    Proceedings of SPIE - The International Society for Optical Engineering 05/2007; DOI:10.1117/12.720743 · 0.20 Impact Factor
  • B. Himed, M. Soumekh
    [Show abstract] [Hide abstract]
    ABSTRACT: Coherent signal processing methods for combining the data that are collected via a multi-channel airborne radar system for moving target detection and image formation, are examined. Methods that convert multi-channel radar data into dual along-track monopulse synthetic aperture radar (SAR) signals of the radiated scene, are studied. A two-dimensional adaptive filtering method that projects the data in one synthesised SAR channel into the signal subspace of the other, is used for blind calibration of the monopulse SAR signals and generation of the moving target indication statistic. The merits of these algorithms are studied using the data from the multi-channel airborne radar measurement system that has been developed by the Air Force Research Laboratory at Rome, New York
    IEE Proceedings - Radar Sonar and Navigation 01/2007; DOI:10.1049/ip-rsn:20050128 · 1.12 Impact Factor
  • K.I. Ranney, Mehrdad Soumekh
    [Show abstract] [Hide abstract]
    ABSTRACT: This letter describes the extension of signal subspace processing (SSP) to the arena of anomaly detection. In particular, we develop an SSP-based, local anomaly detector that exploits the rich information available in the multiple bands of a hyperspectral (HS) image. This SSP approach is based on signal processing considerations, and its entire formulation reduces to a straightforward (and intuitively pleasing) geometric and algebraic development. We extend the basic SSP concepts to the HS anomaly detection problem, develop an SSP HS anomaly detector, and evaluate this algorithm using multiple HS data files
    IEEE Geoscience and Remote Sensing Letters 08/2006; 3(3-3):312 - 316. DOI:10.1109/LGRS.2006.870833 · 1.81 Impact Factor
  • Kenneth Ranney, Heesung Kwon, Mehrdad Soumekh
    [Show abstract] [Hide abstract]
    ABSTRACT: In the past, many researchers have approached the "Hyperspectral-imagery-anomaly-detection" problem from the point of view of classical detection theory. This perspective has resulted in the development of algorithms like RX (Reed-Xiaoli) and the application of processing techniques like PCA (Principal Component Analysis) and ICA (Independent Component Analysis--algorithms and techniques that are based primarily on statistical and probabilistic considerations. In this paper we describe a new anomaly detection paradigm based on an adaptive filtering strategy known as "signal subspace processing". The signal-subspace-processing (SSP) techniques on which our algorithm is based have yielded solutions to a wide range of problems in the past (e.g. sensor calibration, target detection, and change detection). These earlier applications, however, utilized SSP to relate reference and test signals that were collected at different times. For our current application, we formulate an approach that relates signals from one spatial region in a hyperspectral image to those from a nearby spatial region in the same image. The motivation and development of the technique are described in detail throughout the course of the paper. We begin by developing the signal subspace processing anomaly detector (SSPAD) and proceed to illustrate how it arises naturally from the adaptive filtering formulation. We then compare the algorithm with existing anomaly-detection schemes, noting similarities and differences. Finally, we apply both the SSPAD and various existing anomaly detectors to a hyperspectral data set and compare the results via receiver operating characteristic (ROC) curves.
    Proceedings of SPIE - The International Society for Optical Engineering 06/2006; DOI:10.1117/12.665967 · 0.20 Impact Factor
  • Lam Nguyen, Mehrdad Soumekh
    [Show abstract] [Hide abstract]
    ABSTRACT: In support of the U.S. Army vision for increased mobility, survivability, and lethality, the Army Research Laboratory is currently developing a new version of the low frequency ultra-wideband (UWB) synthetic aperture radar (SAR) to support forward imaging. One of the goals in the development of this version of the radar is to make it affordable. This paper presents a study of various forward imaging radar configurations that could be employed in a forward imaging radar system to achieve good imaging resolution with a reasonable number of transmitters/receivers. This study provided us with insights to efficiently configure our transmitter/receiver array. In this study, we examined various radar configurations such as monostatic and some variations of bistatic cases. We provide the analysis of the synthetic aperture radar (SAR) image resolution for these configurations and show the effectiveness of the bistatic configuration with only two transmitters at the ends of the physical array. In addition to the analysis, we also provide simulation results to demonstrate the expected imaging resolutions with respect to the radar configuration and the imaging geometry. Finally, we also consider the use of two squinted transmitters at the two ends and exploit the forward motion of the vehicle to form image on the two sides.
    Proceedings of SPIE - The International Society for Optical Engineering 06/2006; DOI:10.1117/12.666111 · 0.20 Impact Factor
  • Source
    M. Soumekh
    [Show abstract] [Hide abstract]
    ABSTRACT: This work is concerned with developing radar signaling strategies and algorithms that enhance the performance of a synthetic aperture radar (SAR) to image targets in the presence of digital radio frequency memory (DRFM) repeat jammers. The approach is based on a manipulation/alteration of the transmitted signal and/or its parameters in the fast-time domain at every position of the synthetic aperture (pulse repetition interval, PRI); this scheme is known as pulse diversity. In particular, we utilize a modulated form of the transmitted chirp signal and/or vary the rate of the chirp at each PRI; the modulation and/or the rate is varied from one PRI to another, and these variations are only known to the user. The merits of these pulse diversity methods are studied using the autocorrelation and cross-correlation of these signals. We also present a signal processing method for penalizing a DRFM jammer that tries to repeat the radar signal.
    IEEE Transactions on Aerospace and Electronic Systems 02/2006; 42(1):191- 205. DOI:10.1109/TAES.2006.1603414 · 1.39 Impact Factor
  • Source
    Mehrdad Soumekh
    [Show abstract] [Hide abstract]
    ABSTRACT: The main objectives of this effort were to develop a theoretical foundation for Time Domain Non-Linear SAR Processing, and corresponding DSP algorithms to efficiently implement the process on existing computer architectures. We formulated the equations to convert a flight path OPS/INS data into ECEF data that were suitable for mapping into a desired slant imaging plane. The GPS data of an existing SAR platform were used for testing this mapping. Codes were developed for simulating the non-linear SAR signature of targets for a given set of flight path GPS data. We established the mathematical foundation and MATLAB codes for backprojection and wavefiont image formation algorithms for on a non-linear SAR trajectory using multi-processor computers. We conducted parallel computing for the proposed reconstruction algorithms on a shared memory SGI computer and a distributed memory Dell computer using MatlabMPI and C. We also converted the algorithms into parallel Matlab code and created graphical user interfaces for both programs. Parallel algorithms for a SAR-MTI problem were also developed. The two imaging algorithms were studied and tested using both actual and simulated SAR data. These algorithms have also been chosen and implemented for a US Army platform under the WAAMD (Wide Area Airborne Mine Detection) Program.
  • [Show abstract] [Hide abstract]
    ABSTRACT: Ground Penetrating Radar (GPR) has been applied for several years to the problem of detecting both antipersonnel and anti-tank landmines. RDECOM CERDEC NVESD is developing an airborne wideband GPR sensor for the detection of minefields including surface and buried mines. In this paper, we describe the as-built system, data and image processing techniques to generate imagery, and current issues with this type of radar. Further, we will display images from a recent field test.
    Proceedings of SPIE - The International Society for Optical Engineering 01/2006; DOI:10.1117/12.673479 · 0.20 Impact Factor
  • Kenneth I. Ranney, Mehrdad Soumekh
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper addresses change detection in averaged multilook synthetic aperture radar (SAR) imagery. Averaged multilook SAR images are preferable to full-aperture SAR reconstructions when the imaging algorithm is approximation-based (e.g., polar format processing) or when motion data are not accurate over a long full aperture. We examine the application of a SAR change-detection method, known as signal subspace processing, which is based on the principles of two-dimensional adaptive filtering, and we use it to recognize the addition of surface landmines to a particular area under surveillance. We describe the change-detection problem as a trinary hypothesis testing problem, and define a change signal and its normalized version to determine whether: 1) there is no change in the imaged scene; 2) a target has entered the imaged scene; or 3) a target has exited the imaged scene. A statistical analysis of the error signal is provided to show its properties and merits. Results are presented for averaged noncoherent multilook and coherent single-look X-band SAR imagery.
    IEEE Transactions on Geoscience and Remote Sensing 01/2006; 44:201-213. DOI:10.1109/TGRS.2005.859956 · 2.93 Impact Factor
  • Source
    Braham Himed, Mehrdad Soumekh
    [Show abstract] [Hide abstract]
    ABSTRACT: This report examines coherent signal processing methods for combining the data that are collected via a multichannel airborne radar system for moving target detection and image formation. We study methods that convert multi-channel radar data into dual along-track monopulse Synthetic Aperture Radar (SAR) signals of the radiated scene. A two dimensional (2D) adaptive filtering method, that projects the data in one synthesized SAR channel into the signal subspace of the other, is used for blind calibration of the monopulse SAR signals and generation of the Moving Target Indication (MTI) statistic. The merits of these algorithms are studied using the data from the Multi-Channel Airborne Radar Measurement (MCARM) system that has been developed by the Air Force Research Laboratory at Rome, New York.
  • Source
    K. Ranney, M. Soumekh
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper is concerned with change detection in averaged multi-look SAR imagery. Averaged multi-look SAR images are preferable to full aperture SAR reconstructions when the imaging algorithm is approximation based (e.g., polar format processing), or motion data are not accurate over a long full aperture. We study the application of a SAR change detection method, known as signal subspace processing, that is based on the principles of 2D adaptive filtering (M. Soumekh, January 1999) and we use it to recognize the addition of surface landmines to a particular area under surveillance. We identify the change detection problem as a trinary hypotheses testing problem, and identify a change signal and its normalized version to determine whether there is i) no change in the imaged scene; ii) a target has entered the imaged scene; or iii) a target has exited the imaged scene. A statistical analysis of the error signal is provided to show its properties and merits. Results are provided with a realistic X band SAR platform using averaged noncoherent multi-look and coherent single-look SAR imagery.
    Radar Conference, 2005 IEEE International; 06/2005

Publication Stats

1k Citations
96.77 Total Impact Points

Institutions

  • 1987–2012
    • University at Buffalo, The State University of New York
      • Department of Electrical Engineering
      Buffalo, New York, United States
  • 2009
    • Wright-Patterson Air Force Base
      Dayton, Ohio, United States
  • 2004–2007
    • Army Research Laboratory
      • Sensors and Electron Devices Directorate (SEDD)
      Aberdeen Proving Ground, Maryland, United States
  • 2001
    • Naval Undersea Warfare Center
      Newport, Rhode Island, United States
  • 1997
    • Massachusetts Institute of Technology
      Cambridge, Massachusetts, United States
  • 1988–1994
    • State University of New York
      New York City, New York, United States
  • 1984–1986
    • Worcester Polytechnic Institute
      Worcester, Massachusetts, United States
  • 1983
    • University of Minnesota Duluth
      Duluth, Minnesota, United States