Underwater image restoration via adaptive color correction and dehazing.

Abstract

Due to the absorbing and scattering effects, underwater images are often degraded by low contrast, color cast, and haze, which have limited their further applications to underwater vision task. To address this issue, we propose a hybrid framework by adaptive color correction and dehazing for underwater image restoration. Specifically, according to the color attenuation principle, we first design an adaptive color compensation strategy to correct the color cast of the underwater image. In addition, based on an underwater image formation model, we develop a robust dehazing algorithm, in which a new scoring formula across three indicators (i.e.,darkness, distance, and blurriness) and a double-max DCP method are proposed to estimate the background light and transmission map, respectively. The experimental results validate that the proposed method is effective in color correction and dehazing. Both qualitative and quantitative comparisons further demonstrate that the proposed method outperforms several state-of-the-art methods.