Research On Low Illumination Image Enhancement Algorithm

In social life,image processing technology is used in many fields,and it is widely used and has important value in medical care,monitoring,and security.However,in real life,there is a kind of images of special scenes.The acquisition of such images is obtained by the hardware equipment of image acquisition under the condition of insufficient lighting.The process of image acquisition is often accompanied by the phenomenon of low image contrast,color distortion,blurred contours and loss of details,which brings great trouble to the subsequent processing and application.Therefore,it is of great significance to enhance the study of low-illuminance images.The thesis takes low-illuminance images as the research object and makes the following work:(1)Analyze the formation mechanism of low-illuminance images and the characteristics of the images.By studying the histogram equalization algorithm and the contrast-limiting histogram equalization algorithm,perform simulation experiments on MATLAB experimental software to verify the advantages and disadvantages of the algorithm for low-illuminance image enhancement Sex.At the same time,the SSR(Single-Scale Retinex),MSR(Multi-Scale Retinex)and MSRCR(Multi-Scale Retinex with Color Restoration)algorithms in the Retinex algorithm are analyzed and experimentally simulated.(2)In view of the problems of color distortion and "whiteness" of the lowillumination image enhanced by the Retinex algorithm,this paper proposes an improved Retinex algorithm.In order to distinguish the edge details and noise in lowilluminance images,this paper replaces the Gaussian filter of the traditional Retinex algorithm with a guided filter,which better preserves the edge information of the image;in order to prevent color distortion of the image,this paper will RGB color space is converted to YIQ color space,and the luminance component Y is extracted separately;in order to ensure the enhancement of the image contrast,this article sets the image contrast enhancement factor.The experimental results show that compared with the SSR,MSR,and MSRCR in the Retinex algorithm,the improved Retinex algorithm enhances the low-illuminance image,the image is more natural and real,and solves the problem of excessive enhancement and color distortion in the Retinex algorithm.(3)In view of the problems of low contrast,local edge blurring,uneven illumination of the image,and whitening of the picture in low-illuminance images,this paper first studies the concentric receptive field model of the biological characteristics of the human eye,and the biological characteristics of the enhancement of lowilluminance images;second research Image morphology top-hat transform and bottom-hat transform processing algorithms,by introducing Gaussian noise to Lena images,and using a variety of denoising algorithms to compare and verify the denoising effect;finally,this paper proposes a combination of top-hat transform and bottom-hat transform Bionic image enhancement algorithm.The implementation steps of the algorithm are: first,convert the color space of low-illuminance image: RGB color space into HSV color space;secondly,extract the luminance component V in HSV space separately,in order to make the brightness of the low-illuminance image reach the visible observation range of human eyes,Global logarithmic transformation of the luminance component V;in order to enhance the local details of the image,this article uses the retinal neuron receptive field three Gaussian model to process the image;finally,according to the morphological top-hat transformation and bottom-hat transformation of the image,the low-illuminance image is illuminated Average and background extraction.The experimental results show that the proposed algorithm has a significant enhancement effect on low-illuminance images,colorful images and clear contours.It also solves the problems of uneven distribution of image illumination and image depth of field obtained by image acquisition hardware,and has a good visual experience.Finally,through subjective evaluation and objective evaluation,the images improved by the two improved algorithms and the traditional algorithm are evaluated.The experimental results show that the improved algorithm enhances the best lowilluminance images.