Underwater image enhancement method under dark light condition based on improved Retinex theory

According to the underwater image enhancement method under the dark light condition based on the improved Retinex, the limitation of the traditional Retinex technology under the dark light environment is solved based on the improved Retinex. Firstly, noise suppression and preprocessing are implemented, a unique color correction technology is adopted, and the influence of a water body on a color channel is accurately considered. On the basis of the multi-scale Retinex, a multi-brightness model is combined, so that the illumination variation and reflection characteristics in the image can be separated and analyzed more effectively. And the illumination part is intelligently estimated and dynamically adjusted through a deep learning technology. And finally, color constancy and visual naturalness are ensured through image texture enhancement and color balance adjustment later optimization means. The system provides clear, real and valuable visual data for underwater observation, scientific research and engineerin