An efficient underwater image enhancement framework using optimized color correction and texture restoration

Underwater imaging remains a critical component in applications such as marine exploration, archaeological documentation, and autonomous underwater vehicle (AUV) navigation. However, image quality is significantly compromised due to light absorption and scattering effects that introduce color distortion, reduced visibility, and loss of texture. This study proposes a robust, learning-free enhancement framework specifically engineered to address these challenges through a structured three-stage pipeline. The initial stage introduces a novel Contour Bougie (CB) morphological enhancement technique, utilizing dual-scale ring-shaped structuring elements to effectively restore object boundaries and mitigate edge blurring caused by forward scattering. In the second stage, a high-boost sharpening approach enhances fine textures by amplifying high-frequency components, thereby improving structural clarity. The final stage applies adaptive contrast stretching (ACS), which leverages quantile-based dynamic thresholding to rectify color imbalances and optimize brightness and contrast. Unlike conventional deep learning models, the proposed deterministic framework is computationally efficient and does not require prior scene information, rendering it ideal for real-time deployment on embedded systems. Comprehensive experiments on benchmark underwater image datasets validate the framework's superiority over existing enhancement techniques with notable improvements in structural similarity, color fidelity, and contrast enhancement. The results underscore the framework's practical applicability across a range of underwater conditions.