Research On Underwater Image Enhancement Algorithm Based On Adaptive Exposure

In underwater environment,as the light decays rapidly with the water depth increasing and the existence of suspended particles in the water body causes the underwater images to have quality degradation problems such as low contrast,blurring detail,color cast and so on.Correcting the color cast and improving the image contrast to obtain clear underwater images is the key issues in the field of underwater image processing.In this thesis,based on the adaptive exposure method,the problems of underwater image color correction and contrast improvement are studied.An underwater image enhancement method combining color transfer and adaptive exposure is proposed.Firstly,the original image is color-balanced to correct the color cast problem of the image,then the Contrast Limited Adaptive Histogram Equalization(CLAHE)processing is performed on the S and V channels in the HSV space to enrich the image details and improve the contrast of the image.Secondly,the color-balanced image is subjected to color transfer processing to further correct the color of the image.Finally,the CLAHE-processed image is used as a guide image to respectively achieve the adaptive exposure processing of the color-balanced and color-transferred images.The method of linear fusion is used to combine the color and contrast information of the two images.Compared with the typical algorithms,this method can effectively remove the problem of color cast,light over-saturation and image overexposure.An underwater image enhancement method combining Dark Channel Prior(DCP)and adaptive exposure is proposed.Firstly,color balance and Gaussian-Laplacian filter processing are performed on underwater images to obtain more detailed information,and the contrast-enhanced image to be fused is obtained.Secondly,the color-balanced image is subjected to color transfer processing and adaptive exposure processing to be as an input image of the dark channel prior method,thereby obtaining a more accurate global background light estimation value and transmission value,which are used to recover another to-be-fused underwater image.Finally,the four weight coefficients of the two images are extracted separately,and the two images are fused according to the weight coefficient ratio after normalization.This method can effectively avoid the color cast problem after image enhancement,both the contrast and the texture information of the image is improved.