Research On Underwater Image Restoration Via Color Line Theory

Underwater images are an important carrier of marine information and an important source for people to obtain underwater information.They are used widely in the exploration and development of marine resources.However,underwater objects are difficult to identify due to the light absorption and scattering effects which cause underwater images to degenerate,such as serious blurring,low contrast and misleading colors.Therefore,the application of computer vision and image processing technology are becoming important considerations in restoring the true appearance of underwater images.An intensive study of the degraded underwater natural color image based on the color-line(CL)theory is carried out in this paper.Considering that the loss of image color information,noise and other issues will reduce the practical application effect of the color line method in the underwater environment,the main work is as follows:(1)A color compensation method based on color opponent space theory is proposed to address the issue of color information attenuation in the underwater environment.This solution can automatically reduce all types of attenuation while avoiding information interference between components caused by color compensation.This method restores the color information of the image without changing other information compared with some classic color correction algorithms.It lay a solid foundation for the application of color line method in underwater environment.(2)An image denoising algorithm based on color line constraints is established according to the color line distribution characteristics and the dichromatic reflection model(DRM)to remove the noise generated during shooting and transmission of underwater images.This method separates the specular reflection(SR)and diffuse reflection(DR)of pixel and suppresses the noise component in the specular reflection component.The spread of the color clusters is reduced by recombining with the diffuse reflection representing the true color of the object.A better denoising effect can be achieved by this suggestion while retaining the original chromaticity and brightness correlation of the image to the greatest extent.(3)A robust underwater image local transmission optimization model is established according to solve the problem of inconsistent light attenuation rate in the underwater environment.The relationship of three-channel transmission is derived by the association between the global background light(BL)and the inherent optical characteristics,which considering the influence of the attenuation coefficient and object distance on transmission.It allows this model to adapt to change in the underwater environment.In the case of obtaining an accurate transmission estimate,a clear image is derived based on the obtained transmission map(TM)and the estimated background light.The algorithm proposed in this paper has good performance for most underwater images through the comparison and verification of simulation results.