Abstract: In the process of recognizing and restoring image data,data’s sparsity usually occurs due to the similarity of images.In the process of compressed sensing image restoration,the lack of prior information on statistical data generally brings about higher computational complexity and lower restoring accuracy.This paper introduced a refined algorithm of compressed sensing to restore the images,and it defined similarity distances of the matrix as well.The similarity distances and similarity of the matrix are defined by the similarity of the image.Based on the present definition,the application of principal component analysis mapping and Bayesian prior information will enhance images’ iterative recovery algorithm.Experimental results show that the proposed method is more accurate than other restoration algorithms and the images restored are of better definition.Comparatively speaking,the proposed algorithm has a lower computational complexity and consumes a shorter period of computational time.

Key words: Compressed sensing,Principal component mapping,Image restoration,Computational complexity