Abstract:
To address the issue of image degradation caused by backscatter noise in underwater optical imaging, this paper proposes an underwater polarization-correlated imaging method (ABS-UPGI) based on backscatter degradation speckle correction. This method corrects the backscatter degradation of reference speckles to physically weaken the correlation between speckles and background noise, thereby suppressing backscatter. It integrates traditional polarization imaging with computational correlation imaging, incorporating correlation-derived parameters into the polarization imaging model to enhance backscatter suppression at the algorithmic level. Simulation and experimental results show that, in medium-to-low sampling rates and strong scattering conditions, ABS-UPGI improves image reconstruction quality, achieving a peak signal-to-noise ratio (PSNR) increase of ∼4 dB, a structural similarity index (SSIM) improvement of ∼0.15, and an overall image quality enhancement of ∼20%-25% compared to traditional computational correlation imaging (CGI) and polarization-based computational correlation imaging (PCGI). These results validate the method's effectiveness and robustness in underwater backscattering environments, offering a novel approach for suppressing backscatter in turbid water.
