SAR target configuration recognition based on the Dempster–Shafer theory and sparse representation using a new classification criterion

Abstract

ABSTRACTSynthetic aperture radar (SAR) images are sensitive to target aspect angles. To weaken the influences of target aspect angle sensitivity on recognition, a new classification criterion is proposed for sparse representation (SR) based target configuration recognition in this paper. Different from the existing SR-based algorithms which utilize the reconstruction error of each class to identify the targets, the proposed algorithm establishes a supportive degree function to realize recognition. The supportive degree function can enhance the impacts of the samples with small reconstruction errors. Moreover, to further improve the performance of the proposed algorithm, the Dempster–Shafer theory (DST) is used to fuse the information of several similar samples. Experiments on the moving and stationary target acquisition and recognition (MSTAR) database verify the advantage of the proposed algorithm.