Abstract:

A super resolution reconstruction method for sonar image based on multi layer sparse representation is proposed. As for the smooth, edge and texture of the sonar image, the discrete wavelet transform, the contourlet wavelet transform and the Gabor wavelet transform are used to establish the over complete dictionary  respectively. Then the multi-layer sparse represented sonar image is reconstructed in super resolution. Experimental results show that the super resolution images can maintain the geometric features and texture features of original high resolution images effectively. Moreover, the proposed method is robust to noises and can achieve a higher peak signal to noise ratio compared with other methods.