Research And Implementation Of Denoising Algorithm For Low-illumination Image

In the era of digital information,the images have become an important medium for acquiring and transmitting information.However it is often because of the insufficient exposure,the backlight shooting and the other reasons that make the images captured by appear to be of low brightness,low contrast,narrow dynamic range and lots of noise.In this paper,for the characteristics of low brightness,low contrast and lots of noise existing in the low illumination images,we propose the corresponding algorithms to enhance the low illumination images and remove noise.The main work of this article is as follows:To solve the problem of low enhancement of existing low-light enhancement algorithms,and good enhancement but high algorithm complexity,time consuming problems,a compromised low illumination based on human visual characteristics was proposed from a global and local perspective.The method is based on histogram,optimal threshold segmentation and gray level idea,the gray value negated value of the single channel of the extracted image,the inverse value of the gray level is used as the mixed layer coefficient of the filter and the original image to achieve the purpose of low illumination image enhancement.We analyze the characteristics of the noise under the low illumination condition,a hybrid image denoising method is proposed.The method is based on residual compensation and block matching.It not only has better denoising effect on smooth areas but also does not introduce additional information.It maximizes similarity of the structure and texture information.Finally,the low illumination image enhancement algorithm and denoising algorithm are applied on the VS2010 platform,using the MFC(Microsoft Foundation Classes)framework,introducing OpenCV library and embedding the image defog function,we design a software system for image recovery.The software is able to effectively enhance the image with low illumination level and maximize the imageinformation while filtering noise influence properly so as to generate excellent human eye visual experience.The noise of low-illumination images degrades from Poisson-Gaussian noise to Gaussian noise as the brightness of the image increases.This article make use of this feature,focusing on how to maximize the recovery of contaminated image information in low-light backgrounds,propose to enhance the denoising algorithm,and achieved some research results.