Research Of Image Denoising Using Low-Rank Matrix Recovery

In recent years, image denoising model using low-rank matrix recovery gradually becomes a hot research topic in the field of image processing. In the model, Natural image matrix with no noise is regarded as low rank, and the noise matrix of the image is regarded as sparse, so the image information and noise can be separated by decomposing the image matrix into a low rank matrix and a noise matrix. In this paper, the image denoising algorithm using low rank matrix recovery is carried on the thorough research.When removing the salt and pepper noise, median filter can lead to edge shift and blocky, texture detail is not clear, and when the salt and pepper noise is too large, a lot of salt and pepper noise will remain in the image. To the salt-and-pepper noise denoising algorithm using low-rank matrix recovery, it will produce stripes distortion, and texture details have some fuzzy, to solve these problems, presents a new robust image denoising model, salt and pepper noise denoising algorithm using total variation and low rank matrix recovery. Experiments demonstrate that the algorithm can remove salt and pepper noise efficiently.In order to increase the performance of salt and pepper noise denoising algorithm using total variation and low rank matrix recovery, overcome its shortage, such as the clarity of texture details is not enough, when noise increased, the stamp of the salt and pepper noise will remain in the image, it is presented that salt and pepper noise denoising algorithm using fractional total variation and reweighted low rank matrix recovery, which add fractional total variation constraint to the salt and pepper noise denoising algorithm using total variation and low rank matrix recovery, what’s more, the reweighted low rank matrix and the reweighted sparse matrix are introduced. Experiments prove that the algorithm effectively overcomes the shortage of salt and pepper noise denoising algorithm using total variation and low rank matrix recovery, and the performance is greatly improved.When removing the mixture noise, the image denoising algorithm based on low-rank matrix recovery can lead to stripes distortion, if the image mixed with Gaussian noise and salt and pepper noise, for the algorithm only has the constraint of salt and pepper noise, Gaussian noise will still have a part in the image. In addition, when noise increases, the denoising effect has fallen sharply. To solve these problems, mixture noise image denoising using reweighted low-rank matrix recovery is presented. Experiments prove that the algorithm can remove mixed noise efficiently.