Single Underwater Image Restoration Using Variational Framework Guided by Imaging Model with Noise

– Underwater images often suffer from severe degradation due to light absorption, scattering, and noise, resulting in low contrast, color distortion, and reduced visibility. To address these challenges, this paper presents a robust image restoration framework based on a variational model guided by the underwater imaging process. The method begins with accurate background light estimation using dark channel prior techniques, followed by transmission map computation and refinement to correct illumination and haze effects. Further enhancement is achieved by integrating Bidimensional Empirical Mode Decomposition (BEMD) and unsharp masking to improve structural details and edge sharpness while reducing noise. The variational framework formulates image restoration as an optimization problem that balances data fidelity and regularization, ensuring both clarity and smoothness in the output image. Experimental results demonstrate significant improvement in Peak Signal-to-Noise Ratio (PSNR) and visual quality, proving the effectiveness of the proposed approach for underwater image enhancement.