Research On Image Denoising Algorithm Based On Sparse Representation

Images are very susceptible to noise during the process of generation,transmission and acquisition.Image denoising is an essential pre-processing step in image processing and analysis.In recent decades,the research on related issues has been the focus of scholars in this field.Aiming at the problem that the current image denoising algorithm based on sparse representation has low denoising accuracy,this thesis presents two improved image denoising algorithms.The main contents are as follows.1.Aiming at the problem that the current denoising methods based on sparse representation easily lose image details,resulting in low image denoising quality,this thesis proposes an image denoising method based on low rank and sparse representation in a non-local frame.The proposed algorithm is mainly composed of two steps: firstly,matching similar image blocks into groups,establishing a low-rank matrix recovery model,and then using random matrix theory to achieve preliminary denoising;secondly,removing artifacts in the image by non-local sparse representation.Theoretical analysis and experimental results show that the proposed method can filter out noise better,retain image detail information,and obtain better image visual effects compared with the current popular denoising methods of the same kind.2.In view of the problem that the current image denoising algorithm based on K-SVD dictionary learning is not strong enough,this thesis comprehensively considers the characteristics of existing dictionary learning algorithms,and proposes an image denoising optimization algorithm based on sparse representation.The main idea is to use sparse coding and dictionary learning algorithms to represent the main information of the signal with as little atomic information as possible,which can well maintain the structure and texture information of the original image.Specifically,the proposed algorithm uses the AK-SVD algorithm for dictionary atomic update,and at the same time uses the newly given sample data selection method to change the problem of imperfect sample data and greatly improve the performance of dictionary training.Theoretical analysis and simulation experiment results show that the proposed image denoising optimization algorithm based on sparse representation has improved and improved compared to similar algorithms in terms of visual effects or objective indicators such as PSNR.