A method for moving target detection based on airborne multi-aspect SAR system

Abstract

Conventional methods of moving target detection, such as DPCA, GMTI have difficulties in detecting low speed target and showing the movement route. Multi-aspect airborne SAR can illuminate the same scene from different directions to get all-round scattering features of scene. In this paper, a method of moving target detection based on multi-aspect SAR images is proposed. This method can extract the change part from multi-aspect images to estimate the velocity and show the movement route. There is geometric correction, image registration and change detection along the process. As the key part of the detection, an improved image registration method named WT-SIFT is proposed, which combines Wavelet Transform (WT) with Scale-Invariant Feature Transform(SIFT). Finally, the validity of this method of moving target detection is demonstrated by an airborne multi-aspect SAR experiment.