Underwater Vision Enhancement: Adaptive White Balancing Combined with Nature-Inspired Algorithm Mechanisms

Underwater reconstructed images often suffer from color cast, low contrast, and reduced brightness due to absorption and scattering effects. Research has shown that underwater environments exhibit low-pass characteristics, leading to the loss of crucial image information and degradation in visual quality. Image enhancement techniques can help recover this lost information. Various underwater image enhancement methods have been proposed; however, existing approaches often introduce artefacts and produce unrealistic results. To address these challenges, we propose a novel adaptive white-balancing method specifically designed for underwater images. This method primarily focuses on correcting a color cast in underwater scenes. Following white balancing, both local and global contrast enhancements are applied to further improve image quality. The artificial bee colony algorithm is employed to significantly enhance contrast, while a Laplacian filter is used to refine local contrast. The proposed method effectively improves contrast and eliminates color cast, resulting in more natural-looking underwater images. Additionally, the integration of different algorithms enhances the quality of degraded images. Experimental results demonstrate that our method outperforms existing techniques, achieving the highest UIQM score (41.78 on RUIE, 44.82 on EUVP datasets) and the lowest BRISQUE score (2.8 on RUIE, 0.8 on EUVP), indicating superior contrast, color correction, and image naturalness.