Research On Image Super-resolution Reconstruction Algorithm Based On Sparse And Low Rank Theory

Image super-resolution reconstruction technology is to reconstruct high-resolution images in the current environment by using image priori information and combining current single or sequence low-resolution images without changing the level of imaging hardware.This technology not only improves the image quality,saves the economic cost,but also facilitates further research and processing.With the development of science and technology and the improvement of people's living standards,image super-resolution reconstruction technology has been widely used in medical,military and computer vision fields.In this paper,sparse representation and low rank matrix restoration theory are studied.Based on these theories,two novel single image super-resolution reconstruction algorithms are proposed.Experiments are carried out in noisy or non-noisy environments to verify the effectiveness of these algorithms.Firstly,in view of the insufficiency of high-frequency information reconstruction in super-resolution technology,this paper introduces the residual supplementary high-frequency information of the image to enhance the details of the reconstructed image.After the test image is down-sampled,it is reconstructed to the same size as the original test image.The difference between the two images is called as the supplementary high-frequency information.Then the supplementary high-frequency information and the reconstructed image of test image are superimposed together.Experiments show that the performance of the algorithm has been improved.Secondly,in view of the inefficiency of image super-resolution reconstruction algorithm,this paper decomposes the test image by low-rank decomposition,and then reconstructs each part of the image according to its content characteristics.The low-rank part of the image contains most of the information of the original image,so the low-rank part is reconstructed by the sparse theory,and the sparse part is interpolated to reduce the impact of image noise on the sparse learning process,and then the reconstruction results are superimposed together.The experimental results show that both performance and efficiency of the proposed algorithm are improved.Finally,in order to improve the robustness of the reconstruction algorithm to noise,the weighted nuclear norm minimization is introduced to suppress the noise of the sparse part of the image.The weighted nuclear norm minimization makes full use of the singular value information of the image.The experimental results show that the algorithm has good robustness to noise.