Image Restoration Parameters Adaptive Selection Algorithm Basing on Sparse Representation Model

The SL0 algorithm is a sort of sparse reconstruction algorithm approximate to l0 norm, which has significant applications in the field of deblurring. In the SL0 algorithm, usually a number of important parameters need to be set to obtain deblurred images. This paper first introduces the basic of the SL0 algorithm, then it analyzes the operator, which has higher dipartite degree for blurred images in edge extraction algorithm, and choose the Roberts operator as the standard for judging parameter optimization. Finally, an algorithm for image restoration parameter adaptive selection is designed, and experiments are conducted. The experimental results show that comparing with the traditional SL0 algorithm, the algorithm in this paper has a great improvement in terms of repairing quality. The repairing effect of the algorithm in this paper is more natural, and the PSNR of images can be increased about 1.5dB.