Research On Underwater Image Restoration Algorithm Based On Transmission Optimization

Underwater image,as an important source of information for humans exploration of the marine world,has been extensively applied in the domain of ocean engineering.Due to the complexity of the underwater environment,Underwater images are influenced by the absorption and scattering of light from water bodies and suspended solids in water,resulting in problems such as low contrast,serious color casts,and unclear edge details in underwater images,which hinder the recognition and processing of underwater images.Therefore,how to improve the imaging quality of underwater images is a sticking point issue.After analyzing the reasons for the degradation of underwater image,based on the underwater optical imaging model,this paper analyzes and optimizes the transmission,which is the main factor affecting underwater image restoration,and puts forward two underwater image restoration algorithms.The main work is as follows:1)Aiming at the problems of blur and distortion in underwater images,a transmission-optimized underwater image restoration algorithm based on super-pixel prior is proposed.First,transmission is roughly obtained by use of the super-pixel prior method and refined by combining with the improved guided filtering method;Then,the quadtree hierarchical search method is used to estimate the water background light;Finally,combined with the optimized transmission and background light,the imaging model is inversely solved to restore the underwater image.And the Rayleigh stretching method is introduced to enhance the detail of the image.The experimental results show that the restoration algorithm proposed in this paper can effectively remove the blur of the image,improve the contrast of the image,and have good visual effects.2)Aiming at the problems of serious color cast,poor contrast in underwater images,a transmission-optimized underwater image restoration algorithm based on multi-scale fusion is proposed.First,a two-stage transmission fusion optimization method is proposed to estimate and optimize the transmission,and the effective background light is estimated using a super-pixel segmentation method;Then,the original underwater image is initially recovered based on the underwater optical imaging model,combining the optimized transmission and background light.Finally,a multi-scale fusion framework based on the Laplace pyramid is used to fuse the resultant image processed by the improved white balance algorithm and the image initially recovered by the imaging model.Experimentally verified,the underwater image after fusion processing is rich in detail information,improves the image detail contrast,and the recovered image color is realistic and natural,proving the superiority of this algorithm in improving contrast and enhancing image detail.