Image Denoising And Dehazing Algorithm Based On Sparse Representation

In the process of digital image processing,it is easy to be doped with all kinds of noises,which makes the image blurred and even distorted.especially in recent years,the fog and haze weather appeared frequently,the problem of image distortion is more serious.In the image denoising,sparse representation can be used to express the image signal with elements below the Nyquist sampling rate,and the noise does not contain useful information,so it can achieve the effect of denoising.In this paper,we start from the formation of the noise and haze,around the two aspects that the construction of the atomic library and the sparse decomposition,put forward two kinds of optimization with sparse representation of image denoising algorithm and a dark channel prior model with the sparse representation algorithm to dehaze.The main work and research results are as follows:1.Summarize the domestic and foreign research status of image denoising algorithm and the current mainstream image denoising algorithm,study several representative image denoising algorithms based on sparse representation,analysis the existing problems and the technical problems to solve.2.in order to improve the efficiency of KSVD dictionary and the ability of sparse representation algorithm to improve the KSVD dictionary,the new sparse representation image denoising algorithm based on the optimal dictionary design is proposed.3.Aim at the defects in the initial dictionary design on the KSVD algorithm and the problem of poor denoising effection,propose the image denoising algorithm based on adaptive sparse block,the denoising effection is better than the original KSVD algorithm,and preserve the more image edge information.4.Based on the existing image defogging algorithm,analyze the current problems and the technical problems need to solve of the dark channel algorithm,put forward a kind of image dehazing algorithm that based on sparse respresent and dark channel prior.The new algorithm use dark channel prior model to defog,the sparse representation algorithm is applied to the image to dehaze.Compared to and the dark channel prior model,the new algorithm has a better effect to dehaze and high fidelity.5.This paper proposed sparse representation image denoising algorithm based on the optimal dictionary design,sparse image denoising algorithm based on adaptive block and image dehazing algorithm that based on sparse respresent and dark channel prior,and the simulation results were given.