Polarization Imaging Descattering Based on Dark Channel Prior Background Light Estimation

Abstract

In turbid water environments, conventional polari-metric imaging descattering methods usually rely on selecting a non-target background region as a substitute for global back-ground light information, leading to image clarity. By integrating polarization imaging with the dark channel prior (DCP) and fully utilizing the polarization information from the Stokes vector and the DCP, the interference of the target information in calculating the background information can be effectively eliminated. This method also enables accurate estimation of the global distribution for the degree of polarization (DoP) and the angle of polarization (AoP) of the backscattered light and eliminates the need for the complex process of background region selection. Experimental results with varying turbid water concentrations demonstrate the proposed method's effectiveness for target recovery in both underwater images with and without background regions. Compared with the conventional active polarization imaging model, our method improves image contrast by 47% in the strongly scattering environment.