Image Deconvolution Model And Algor Ithm Base On Non-convex L_p Norm And G-Norm

With the development of science and technology, people can use all kinds of ways to save the image. And people also like to communicate by image, t hus image processing catches people attention more and more. However the image we get is usually blurred because of the movement of the object or filming equipment, which affects the subsequent application of the image. To eliminating the image blurring, image deblurring attracts the researchers' attention of home and abroad.In this paper, we propose a novel model that integrates non-convex l_p(0?p<1) norm and Meyer's G-norm. Specifically, the regularization term based on l_p(0?p<1) can describe the sparse feature of image, and the G-norm can effectively suppress noise and keep small feature of image in deblurring. In addition, we design a simple algorithm based on the alternating minimization. We can use Split Bregman iteration to solve the problem about G- norm and use GST algorithm to solve the problem about non-convex l_p(0?p<1) norm.At last, we compare the result of coupling algorithm with the result that we can get by using Split Bregman iteration and GST algorithm,the experimental result show that the new model is practicable.