A novel approach to moving targets shadow detection in VideoSAR imagery sequence

Abstract

The Doppler shift effect results in some targets shadows in theirs actual position, and a strong correlation exists between adjacent frames of Video Synthetic Aperture Radar (VideoSAR) imagery. Based on the above rationale, a novel approach to moving targets shadow detection for high-frame-rate VideoSAR imagery sequence is presented. First, a fast preprocessing stage is essential in real applications, where the SIFT with RANSAC registration algorithm is employed to compensate for the changing background, and the CattePM model is used to suppress the speckle noise. Then, in order to separate the targets and the background automatically, a threshold segmentation algorithm, called maximizing the Tsallis entropy, is applied. Finally, background difference with three frame difference method implements the precise moving targets extraction. Experimental results utilizing VideoSAR imaging fragment show that multiple moving vehicles are detected effectively and, hence, the validity has been demonstrated.