Rapid, prior-free, single-shot polarization descattering: An ADMM-based dual-layer optimization method for turbid water

Abstract

Accurately and efficiently estimating the degree of polarization (DoP) of both target and background remains a crucial challenge in prior-free underwater polarization descattering imaging. To address this, we propose a novel single-shot underwater polarization image reconstruction algorithm on the basis of the alternating direction method of multipliers. A dual-layer multiobjective optimization framework constitutes the core of the proposed method. The first layer optimizes image contrast and mutual information, while the second layer optimizes information entropy. This hierarchical approach ensures that the reconstructed image achieves high contrast and effective descattering while completely preserving details. Furthermore, the algorithm’s convergence is enhanced by reformulating the constrained optimization problem into an unconstrained one, substantially reducing the processing time and making it suitable for rapid imaging. Experimental results demonstrate that the proposed algorithm outperforms existing methods in key metrics, such as processing speed, contrast, and information entropy, while yielding superior image quality. It is particularly effective for rapid, high-quality imaging in strongly scattering underwater environments.