Research On Underwater Image Restoration Based On Variational Method

The underwater captured images often suffer from haze,noise and low contrast due to the absorption and scattering of the water and suspended particles,which make it difficult for analysis and understanding.Therefore,it is significant to improve the quality of degraded image and recover a clear image for image understanding and recognition.However,the existing restoration methods do not fully consider the effect of noise in the degradation process,resulting in an unsatisfactory result.Aiming at solving the problems of underwater image defogging and denoising,the mainly contribution of our work are as follows:(1)First,a variational method based on curvature term for underwater image restoration is proposed to solve the problem of foggy and fuzzy.The data and rule terms of the variational model are designed according to the underwater optical imaging model.Then,the underwater dark channel prior is used to estimation the global background light and the transmittance map.In order to further improve the computational efficiency,a fast alternating direction multiplier method(ADMM)is introduced to solve the energy function.Experimental results show that the proposed algorithm can effectively remove water mist,suppress underwater image noise,and improve image contrast and visibilty.(2)Moreover a fast variational approach based on Laplace operator prior term is proposed to dehaze and denoise simultaneously.Based on underwater optical image formation model,the data item and the regular item of the unified variational model are designed,wherein the Laplacian operator prior is adopted as the regular term.The improved dark channel(DCP)and underwater red channel prior are applied to estimate the global background light and the transmission map,respectively.Likewise,the fast alternating direction multiplier method(ADMM)toaccelerate the whole progress.A considerable of qualitative and quantitative experiments are conducted on different challenge scenes.Compared with other state-of-the-art methods,our proposed method has a better performance on contrast enhancement,visibility improvement and color correction.The restored resultsare more consistent with the human visual system(HVS).