Oceanic Radiance and Chromatic Adaptation: A Three-Stage Framework for Underwater Image Enhancement

Light absorption and backscattering in underwater environments cause severe chromatic aberration and visibility loss. To address this, we propose a modular framework for holistic restoration termed Oceanic Radiance and Chromatic Adaptation (ORCA). Contrary to conventional end-to-end deep learning architectures that behave like "black boxes," our proposed system divides the problem into three physically informed blocks. The Radiance Restoration module estimates back-scattering and transmission, the Chromatic Adaptation Network recovers the attenuated red channel, and the Adaptive Illumination block applies multi-scale contrast enhancement. By physically grounding these adaptive layers, ORCA preserves color constancy while emphasizing high frequency details. Benchmarking on U45 and UIEB datasets (N = 935 images) achieved a UIQM of 3.48 and MUSIQ of 72.7, outperforming current state-of-the-art methods. Statistical validation via Paired t-tests (p < 0.005) confirms ORCA as a reliable vision tool for autonomous underwater vehicles (AUVs).