Underwater image restoration via multiscale optical attenuation compensation and adaptive dark channel dehazing

Underwater images often suffer from color cast and low visibility due to inherent factors such as light absorption, scattering, and turbidity. The quality-degraded underwater images are unfavorable for underwater research and applications.To effectively deal with these quality degradation issues, this paper presents a novel restoration framework tailored specifically for underwater images, aiming to restore their natural clarity and improve their visual quality. Firstly, a multi-scale optical attenuation compensation color correction algorithm is employed to correct the color deviations of underwater images. Subsequently, an adaptive dark channel dehazing algorithm is proposed, including the global background light estimation algorithm based on multiple optical prior properties and a more sensitive segmentation transmission map estimation algorithm. Our approach integrates advanced image restoration techniques with domain-specific optimizations, ensuring robust performance across diverse underwater conditions. We comprehensively evaluate our method on a wide range of underwater image datasets, demonstrating its effectiveness in restoring color fidelity, contrast, and texture details. Furthermore, we analyze the quantitative and qualitative impacts of our framework, showcasing its advantages over existing state-of-the-art methods. Our work not only advances the field of underwater image restoration but also provides valuable insights into designing future restoration algorithms for this domain.