HFM: A hybrid fusion method for underwater image enhancement
Shunmin An,Lihong Xu,IEEE Senior Member,Zhichao Deng,Huapeng Zhang
DOI: https://doi.org/10.1016/j.engappai.2023.107219
IF: 8
2023-10-06
Engineering Applications of Artificial Intelligence
Abstract:Different underwater images captured by different devices may exhibit varying degrees of nonlinear rendering , leading to white balance distortion. Meanwhile, wavelength attenuation and light scattering associated with distance can cause color shifts, reduced visibility, and decreased contrast in underwater images. This will result in difficult access to underwater information, which in turn will affect the application of advanced vision. Considering these degradation issues, we propose a hybrid fusion method for underwater image enhancement, called HFM. In terms of technical contributions, we introduce a color and white balance correction module that addresses color and white balance distortion in underwater images using the gray world principle and a nonlinear color mapping function. We design a visibility recovery module based on type-II fuzzy sets and a contrast enhancement module using curve transformation. Besides, inspired by image fusion methods, we propose an underwater image perception fusion module that focuses on two different tasks simultaneously, fusing underwater images of visibility and contrast. Therefore, the proposed method can effectively solve the problems of white balance distortion, color shift, low visibility and low contrast in underwater images, and achieves optimal results in application tests of geometric rotation estimation, feature point matching and edge detection. Through comparative experiments analyzed on four real scene datasets, the proposed method achieves superior results compared to 14 state-of-the-art underwater image enhancement methods. The code is publicly available at: https://github.com/An-Shunmin/HFM .
automation & control systems,computer science, artificial intelligence,engineering, electrical & electronic, multidisciplinary