A Review of Research Progress on Image Dehazing Algorithms
Sheng Zhong, Junling Zhao, Tian Mao, Meimei Song, Xiaochun Zhang, Feiran Xu · 2024
Image dehazing stands as a pivotal and formidable subject within the realm of computer vision. This paper delves into recent advancements in image dehazing techniques, with a specific focus on the burgeoning trend of diversified fusion in this field. Drawing upon four foundational mathematical methodologies routinely employed in image dehazing—namely linear computation, enhanced filtering, variational methods, and function approximation—the paper systematically scrutinizes and consolidates approaches within image dehazing that leverage multiple-image information, pixel features, energy functionals, and deep learning. Furthermore, to dissect the nuances of dehazing methods, this paper delineates seven quantitative evaluation criteria from diverse perspectives of image representation.t.