Imaging moving objects in 3D from single aperture synthetic aperture radar
Mark A. Stuff - G, Martin J. Biancalana - General Dynamics AIS, Gregory Arnold - AFRL/SNAT, Joseph Garbarino - General Dynamics AIS
Tue, 27 April 2004, 4:10 PM - 5:30 PM
Abstract- When a moving object is imaged with conventional synthetic aperture radar (SAR) the result is a displaced smear. This is due to the extra information the object?s motion imparts to the radar return. If the motion is rich enough ? and it usually is ? there should be a possibility of forming a 3D image of the object. This involves understanding the way the radar data are arranged in phase space. The data lie on a convoluted surface that occupies three dimensions rather than the two-dimensional plane used to form conventional SAR images. To achieve three-dimensional images the data must be extrapolated from the surface into a volume. In this phase space, there is a great deal of structure and therefore the possibility of extrapolating to a volume of data.
The effort is motivated by the potential value of the three-dimensional image products that become available from the data volume. The volume can produce aspect, aspect, range data that results in views of the moving object not available previously. These include radar images from the point of view of the radar rather than orthogonal to it, range images similar to ladar, orthogonal three-view images similar to mechanical drawings, and other image products.
General Dynamics AIS, supported by the Air Force, has been investigating exploiting moving targets whose returns are captured by conventional SAR systems. The result is a processing system that can extract the detailed three-dimensional motions of a moving object. This system is called Three-Dimensional Motion and Geometric Information (3DMAGI). When the estimated motions are used to motion compensate the radar data to the moving target, the resulting data surface deviates radically from the conventional SAR image plane. No simple function, such as a quadratic, can create this data surface. Therefore no simple focusing scheme can accurately focus this data into a standard 2D SAR image. By using the measured data surface some approximations can be made that will produce conventional-looking radar images. These, however, are likely to include partial side or end views and the azimuth (and / or height) resolution will depend on the shape of the data surface (motion of the object), and may greatly exceed the range resolution defined by the bandwidth of the radar system.
This paper reports on work done with a full volume of data from the National Ground Intelligence Center and vehicle trajectories measured by an inertial system on a moving vehicle. Its goal is to determine how to best use the rich data available from advanced processing to produce images and image products that will simplify the task of exploiting the radar image. The data and sample trajectory are described as well as how they are used to emulate the result of 3DMAGI processing. The work consists of investigations into the methods of creating a 3D data volume that matches the NGIC chamber collection, starting from a small subset defined by the data surface which lies in the full volume. How much extrapolation is needed to get acceptable results is the first question posed. From there, the question of just what methods yield the best results is examined. Limitations of various methods are explained with examples. Comparisons of each method of extrapolation to the original data volume are presented to give an indication of progress toward the goal.
Dr Mark A. Stuff - G
Dr. Mark Stuff received a master?s degree in mathematics from Western Illinois University in 1976, a master?s degree in applied mathematics from the University of Colorado in 1980; and his PhD from the University of Michigan. From 1980 to 1985 he worked as a geophysicist for Shell Oil Company, where he developed new methods for seismic signal processing. From 1985 to the present he has worked for General Dynamics Ann Arbor Research and Development Center. Here, he has developed methods for thermal and electromagnetic simulations and new signal processing methods for radar signals.
Mr. Martin J. Biancalana - General Dynamics AIS
Martin Biancalana received the BS degree from the University of Illinois in 1969 and the MSEE degree from the Air Force Institute of Technology in 1978.
He is Research Program Manager for General Dynamics Ann Arbor Research and Development Center. In the Air Force, he worked on the PAVE MOVER advanced development radar and managed the TACIT BLUE stealth research project. For General Dynamics, he worked on foliage penetrating radar and managed the Gold Pan ?93 project. Currently, he is working on applications of imaging radars to moving targets since 1995. He is an active member of the IEEE Engineering Management Society.
Dr Gregory Arnold - AFRL/SNAT
Dr. Gregory Arnold earned his BSEE at the University of Dayton (Dayton, Ohio), and his Masters and Ph.D. in EE from the University of Virginia (Charlottesville, Virginia).
He is a project engineer for the Air Force Research Laboratory working at Wright-Patterson Air Force Base (Dayton, Ohio) in the Target Recognition and Fusion Algorithms branch of the Sensors Directorate. His research interests include a wide range of theoretical and practical issues in invariance, computer vision, statistics, and signal processing as applied to radar, ladar, video, infrared, and spectral sensors. He is a member of IEEE.
Mr. Joseph Garbarino - General Dynamics AIS
Joseph Garbarino received the B.S. degree in Computer Science and Mathematics from Eastern Michigan University in 1983, and the M.S. degree in Computer Science from University of Michigan in 1988. He is a Lead Engineer for General Dynamics Ann Arbor Research and Development Center. He has worked on various imaging programs, including his most recent work on applications of imaging radars to moving targets. Interests include applications of imaging, the use of 3-dimensional graphics in imaging applications, and the utilization of computer clusters to make CPU intensive algorithms tractable.