Research On Retinex Image Enhancement Algorithm Based On Sparse Representation

Image enhancement is an important part of the image processing, the aim of image processing, according to the specific requirements to improve the image visual effects or enhance the kind of information, highlight the information of interest or restrain not interest in information,enhanced into a better image.Among them,the degraded image enhancement is most widely used,due to the bad environment lead to reduced image quality because of the complexity of its causes, particularity, the current research about the degraded image enhancement has certain limitation,needs to be improved and strengthened.Therefore,this article in view of the bad environment, such as weather, illumination and noise mechanism of causes of reduced image quality, based on the Retinex algorithm variational framework,advance the research of degraded image enhancement.This article first introduces the principle of image enhancement, and the method of image enhancement based on the Retinex algorithm and conducted a comparative analysis of their advantages and disadvantages.The research based on the Retinex algorithm based on variational framework.For its lack of prior knowledge of reflection components, puts forward the improved method using sparse representation theory. Sparse representation theory of K-SVD algorithm has a good reduction in piecewise constant properties , but due to the large dictionary update computation result in slow speed, so we use the batch orthogonal matching algorithm, in the process of dictionary update only for update column calculation,greatly reduce the computational complexity and improve the operation speed.And we apply the sparse representation theory to the Retinex algorithm based on variational framework, in the degraded image enhancement.Based on the theory of the appeal, this algorithm can use for different types of degraded image caused by bad environment enhanced, by compared with other algorithms, proves that the algorithm in this paper is better in the brightness promotion, contrast enhancement, good effect on reduction of structure and detail.