Underwater image restoration by turbulence model based on image gradient distribution

Abstract

For the underwater degenerative image caused by obscure and noise, atmospheric turbulence model is utilized to construct the restoration model. By computing the underwater image gradient distribution which is used to estimate the adaptive degradation model parameter, underwater image is restored through Wiener filters. The experiments show that the proposed method exerts powerful effort on restoration of the obscure underwater image caused by scattering. The proposed method has better improvement on contrast than other common preprocessing methods and is especially suitable for further degradation model analysis and objective detection.