Underwater Image Restoration via Information Distribution and Light Scattering Prior. (May 2022)

• Record Type: Journal Article
• Authors: Zhou, Jingchun; Wang, Yanyun; Zhang, Weishi
• Is Part Of: Computers & electrical engineering. Volume 100(2022)
• Publication Date: 2022-05
• Subjects: Underwater image restoration -- information distribution -- depth map -- transmission map; Computer engineering -- Periodicals; Electrical engineering -- Periodicals; Electrical engineering -- Data processing -- Periodicals; Ordinateurs -- Conception et construction -- Périodiques; Électrotechnique -- Périodiques; Électrotechnique -- Informatique -- Périodiques; Computer engineering; Electrical engineering; Electrical engineering -- Data processing; Periodicals; Electronic journals; 621.302854
• DOI: 10.1016/j.compeleceng.2022.107908
• Languages: English
• ISSNs: 0045-7906

Abstract

Highlights: A novel model based on information distribution and light scattering is designed to adaptively estimate the background light. The method only uses a single image as input. Estimate scene depth with feature priors. Perform extensive experiments to validate the proposed method. Abstract: Underwater images suffer from blurred details, color cast, and contrast reduction, caused by absorption and scattering of light. To address these degradation issues, we propose a novel underwater image restoration method based on information distribution and light scattering prior. Concretely, we first estimate the background light according to the information distribution and light scattering characteristics of the background light area. Then, the scene depth map and transmission map are obtained by the relationship between the brightness information and color attenuation. Finally, the underwater image formation model is inversed to obtain the restored image. Extensive experiments indicate that our method can effectively remove the color cast and increase contrast. Furthermore, compared with several state-of-the-art methods, the proposed method yields better results. The qualitative and quantitative results demonstrate that our approach is superior to other comparison methods. Graphical abstract: Image, graphical abstract