Research On Color Low Illumination Image Enhancement Algorithm

In modern social life,image has become an important carrier of information dissemination.The imaging quality and human visual effect of image will play a decisive role in the display and transmission of information.However,due to the influence of weather,light,imaging equipment and other external factors,the real information of the original image is submerged,which brings inconvenience to the subsequent image recognition,detection and segmentation operations,Therefore,the research on color image enhancement algorithm is particularly important.This thesis takes the color low illumination image as the research object,which aims to improve the overall quality of the low illumination image and highlight and show the important information areas in the image.The specific work of this thesis is as follows:(1)Firstly,the research status of low-illumination image enhancement algorithms at home and abroad is introduced,and the steps,processes,advantages and disadvantages of low-illumination image enhancement algorithms based on different methods are briefly described.Then,the basic principles of image enhancement algorithms used for reference in this thesis are introduced in detail,the influences of various parameters on image enhancement effects are analyzed,and the realization process of image enhancement algorithms is deeply analyzed.(2)Aiming at the defects of low illumination image,such as dark overall,low contrast and color saturation,and poor detail display,a low illumination image enhancement algorithm based on illumination reflection model is proposed.Firstly,the original image RGB color space is converted to HSV color space,and the V component is guided and filtered to estimate the illumination component in the image.A joint gray histogram adaptive gamma correction algorithm is proposed to process the illumination component,and the illumination intensity of the image is improved adaptively.The reflection component is obtained according to the illumination reflection model,and the reflection component is processed by multi-scale unsharp masking strategy to improve the essential information and edge details of the image.Finally,the enhanced illumination and reflection components are fused to obtain the final enhanced image.The experimental results show that this algorithm can effectively improve the contrast,clarity and detail display ability of the image,while maintaining the color without distortion.It is also superior to other low-illumination image enhancement algorithms based on Retinex theory in the improvement of standard deviation,average gradient and information entropy.(3)To improve the overall quality of low-illumination images,it is easy to lose local colors or details,and the reality of images is destroyed,a low-illumination image enhancement algorithm based on illumination component correction and compensation is proposed.Firstly,the original image is converted into HSV color space,and an adaptive nonlinear function is constructed to stretch the saturation component S,Combining multi-scale weighted average strategy with gradient domain guided filtering,the illumination component of V component can be accurately extracted,and the illumination component can be processed by two-dimensional adaptive gamma correction to solve the problem of uneven illumination of low-illumination images.Finally,it is converted to RGB color space,and an improved homomorphic filtering algorithm is proposed to compensate the illumination of images in frequency domain,so as to improve the visual effect of images.Experimental results show that this algorithm can improve the contrast and overall quality of the image,keep the original naturalness of the image,and the overall structure of the image is not distorted.It is also superior to other low-illumination image enhancement algorithms based on different methods in terms of peak signal-to-noise ratio and structural similarity.