Shadow-Tracking Algorithm For Moving Target Detection
Walter G. Carrara, Chris Roussi, Susan A. S. Werness · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 1989
An algorithm to detect the presence of ground-moving targets in synthetic aperture radar (SAR) images has been investigated. Its performance is independent of the direction of the target velocity. Called the notch algorithm for its method of processing, it makes use of the shadow characteristics of a target moving over a clutter background in a unique way. While the image of a moving target may be displaced from its true position in a SAR image, the shadow cast by that target will not be shifted, although it will be de-focused and difficult to detect with conventional SAR processing. The notch algorithm attempts to find the target's ground-track and estimate its velocity vector using a single, coherent aperture of radar data. This information is critical to finding the moving target in the image of the scene and to indicating phase compensations required to begin to re-focus the target. Ground-track information can be obtained by applying the notch algorithm to the same data set several times while varying the algorithm parameters, followed by a comparison of the resulting images with the conventionally processed image. In this paper we describe the basis for the algorithm and present simulation results showing the performance over a range of target velocities. We also suggest several methods to improve detection.