Depth-Based Underwater Image Restoration Using Blur Map and Channel Priors

Underwater images often suffer from degradation due to light absorption and scattering, resulting in reduced contrast, color distortion, and loss of details. Traditional restoration methods such as Dark Channel Prior (DCP) are designed for atmospheric haze and perform poorly in underwater environments. In this paper, a depth estimation based underwater image restoration method is proposed using blur map analysis, red channel prior, and maximum intensity prior. The blur map captures scattering effects, while channel priors assist in estimating scene depth. The estimated depth is used to compute the transmission map, followed by background light estimation and image recovery to obtain the enhanced image. Experimental results demonstrate improved visual clarity and detail visibility. Performance evaluation using entropy and Underwater Image Quality Measure (UIQM) indicates that the proposed method effectively enhances underwater image quality compared to the existing DCP-based approach.