A Shadow-based Method for Moving Target Detection in Video SAR

Abstract:
In a video synthetic aperture radar (video SAR), a ground moving target often appears as a dynamic shadow at its true position in radar imagery because its own image is shifted away and smeared. The shadow-based moving target detection in video SAR does not rely on the radar cross section (RCS) of target, providing an attractive approach for the SAR ground moving target indication (SAR-GMTI). This paper first describes the models of the moving target shadow. Then, a novel method is proposed for the detection of the dynamic shadow of the moving target through a shadow-scoring operation for each pixel in the SAR imagery, followed by a connected domain screening in segmented images and an inter-frame association. Simulation results validates that the proposed method can detect ground moving targets effectively in video SAR.