Research Of Single Image Haze Removal Algorithm Based On Histogram Equalization And Dark Channel Prior

In bad weather conditions such as fog and haze,a large number of small water droplets and dust particles suspended in the air will scatter and refract the airlight,resulting in reduced outdoor visibility,and the images obtained by the imaging equipment will be seriously polluted.The images affected by the bad weather has reduced contrast,color distortion,or missing scene information.It greatly influences and limits the utility of visual systems considering the poor visibility of foggy images and the difficulty in analyzing and extracting image features.Therefore,restoring degraded image under hazy weather conditions has important theoretical significance and application value.In this thesis,we deeply study the dark channel prior theory,atmospheric scattering model and guided filtering algorithm,and combine the histogram equalization algorithm to propose a new defogging algorithm that refines the transmission map.The article first obtains the minimum color component image of foggy images,which contains a great deal of the depth of field and detail information in foggy images;the minimum color component image is processed by the histogram equalization algorithm to improve the contrast,and at the same time,the compensation of the local illumination unevenness is provided in the minimum color component image;Considering the partial texture effect in the minimal color component image processed by the histogram equalization,using the bilateral filtering algorithm which has the feature of edge-preserving and denoising smooth the image;Contrast refined transmission map which obtained by the single color channel image and the RGB color channel image being respectively the transmission map refined by the guide image,and the single color channel image is selected as the guide image.After processing by the above-mentioned histogram equalization algorithm and bilateral filter algorithm,the smallest color component image is used as a guide map and substituted into the guide filter algorithm,and finally the refined transmission map is obtained.The top 0.1% of the pixel value is selected in the dark channel image,and corresponds to the original fog image.The value of the 0.1% pixel is averaged as an estimate of the atmospheric light value;Finally,the refined transmission map and the re-evaluated atmospheric light value is brought into the atmospheric scattering model,and the defogging image is recovered.The experimental results show that the fog-free image recovered by this algorithm has better definition,color saturation,and rich local details.The color at the sky area has more realistically restored,and at the same time,the operation speed of the defogging algorithm is improved,and it is possible to process the fog image in real time.