Research On Underwater Color Image Enhancement Based On Two-Scale Image Decomposition

Oceans,lakes,and other waters are the cradle of life and contain rich product resources.As one of the important means of underwater scene information extraction,underwater image acquisition is very important to obtain high-definition underwater images.However,due to the complex and changeable underwater scenes,the collected underwater images are often of low quality.On the one hand,the water body itself has different absorption capabilities for different colors of light,resulting in color deviation in the underwater color images;on the other hand,due to the underwater suspended media and ambient light scatter each other,resulting in blurred images and reduced contrast.Whether based on traditional methods or deep learning methods,existing underwater image enhancement algorithms cannot take into account the problems of image color deviation,red shadow,and low contrast.To solve the problems of color cast and low contrast in underwater images,we propose an underwater color image enhancement algorithm based on two-scale image decomposition.The research is as follows:(1)Aiming at the serious color deviation problem of underwater color images,a new color enhancement method of underwater color images using two-scale image decomposition is proposed.The method first improves the color deviation problem by contrast stretching method based on mean and variance and uses median filtering to reduce the noise particles introduced by contrast stretching in the red channel.Then,the green channel image is decomposed by a two-scale image to compensate for the details of the red channel image,and finally,the real details and colors of the original red channel image are introduced into the processed red channel image,and the RGB three channels are merged.The underwater images processed by the method can be better improved in color and do not produce red shadows.(2)Aiming at the defects of the local blur,low contrast,and poor details in underwater color images after color correction,a method for enhancing the contrast and details of underwater color images based on guided filtering layering was proposed.The method firstly uses the guided filtering method with a better layering effect to decompose the color-corrected underwater image on two scales and then performs image enhancement processing on the base layer and the detail layer obtained by layering respectively.According to the Pyramid fusion strategy,different methods are used to enhance the brightness channel of the base layer to obtain three initial fusion images.Then,the exposure weight map is used to measure and extract more brightness details of the input image.Resolution reconstruction generates the final enhanced image of the base layer;for the image of the detail layer,an adaptive gamma transform function is used to enhance the details and texture information of the image.Finally,the enhanced image of the base layer and the detail layer is weighted and fused utilizing mean and variance,and then the contrast detail enhanced image is output.The contrast and details of underwater color images processed by the method are significantly improved.The subjective analysis results show that the proposed method can better solve the problem of image color deviation and red shading compared with other methods,and the objective indicators and feature point matching results are better.The two-scale image decomposition method proposed in this paper uses the characteristics of underwater image imaging to solve the problem of image color deviation and low contrast.It has good scene adaptability and improves image quality.It is generally suitable for underwater color image enhancement in complex environments.