Abstract:

To improve the synthetic aperture radar (SAR)target variant recognition performance, on the basis of the discriminative dictionary learning and joint dynamic sparse representation model, a new SAR target recognition method is proposed based on the multi-information dictionary learning and sparse representation. In the training stage, the discriminative dictionary learning method label consistent KSVD (LC-KSVD) is used to learn dictionaries for both the image domain amplitude information and the frequency domain amplitude information of the targets. In the test stage, based on the learned dictionaries for the two kinds of information, the test target representation coefficients for the two kinds of information are computed using the joint dynamic sparse representation model. Finally, the test target can be classified according to the representation residual for the two kinds of information. The MSTAR dataset is used to verify the effectiveness of the proposed method. Experimental results show that the proposed method has better recognition performance than some existed methods.