MHF: A multi-task hybrid fusion method for underwater image enhancement based on biological vision

Abstract

Enhancement of underwater images is a new challenge in image research because low image visibility and contrast due to wavelength attenuation of underwater light and the effect of suspended particles in the water are most obvious. These problems can lead to difficulties in underwater information extraction and affect the development of underwater research, so we propose a multi-task hybrid fusion method (MHF) for underwater image enhancement based on biological vision. In terms of technological innovation, we designed an improved type II fuzzy set computation module based on the foundation of biological vision to improve the visibility of images. Meanwhile, we designed an adjustable contrast stretching module to improve image visibility. In addition, inspired by the fusion approach, we introduce a visual fusion module which fuses the results of the above two modules with a weight ratio. Therefore, this method focusing on multi-task synchronization can overcome the limitations of previous methods and effectively solve the problems of white balance distortion, color shift, low visibility, and low contrast in underwater images, and achieve the best results in the application tests of geometric rotation estimation, feature point matching, and edge detection. The experimental results demonstrate that the application results of this research method on 2 datasets outperform the top 14 existing algorithms. The wide applicability and excellent performance of the method are verified through application tests on various underwater vision tasks. By explicitly addressing the limitations of existing methods, the method becomes an advantageous solution in underwater image processing, providing enhancements in image quality and task-specific applications in a concise and efficient manner.