Abstract:

A new approach is developed for synthetic aperture radar (SAR) automatic target recognition based on Bayesian compressive sensing (BCS). Firstly SAR images are segmented into image data of target zones by constant false alarm rate. Then based on the BCS model, the sensing matrix is constructed by all training sets. The sparse coefficient vectors corresponding to the test samples are solved. Recognition is performed according to the L2 norm corresponding to each of training types of samples in the sensing matrix. Experimental results with the moving and stationary target acquisition and recognition public dataset show that the proposed approach has good recognition effects.