A polarization-guided underwater image descattering method

Due to the scattering and absorption of light caused by the existing particle, the quality of underwater optical imaging is limited, which often restricts tasks such as ocean observation and underwater surveillance. While polarization-based descattering provides an effective way to separate object radiance from scattering components, most existing methods rely on conventional Stokes formulations and heuristic parameter tuning, limiting robustness in complex underwater environments. To address this issue, we propose a polarization-guided underwater image descattering method, which combines physically motivated modeling with automatic parameter optimization. We introduce three global dimensionless parameters to account for polarization-channel imbalance and global airlight scaling, while a degree of linear polarization (DoLP)-gated airlight estimation is designed to suppress scattering leakage. A two-stage optimization strategy is further employed to ensure stable parameter convergence. Experiments show that, compared with traditional methods, this framework can reproduce clearer images with more detailed textures.