Underwater Polarization Imaging Method Based on Frequency Domain Processing Combined with Global Estimation

This paper proposes an underwater polarization imaging method that combines frequency domain processing with global estimation to improve scattered light suppression. By preprocessing orthogonal polarization images with histogram stretching and high-pass filtering, the method reduces low-frequency scattered light and accurately estimates the degree of polarization. A low-pass filter and optimal threshold factor in the frequency domain further refine this estimation. Enhanced Metric Evaluation (EME) is used to optimize the de-scattering process, leading to significant improvements in image contrast, sharpness, and clarity compared to traditional methods. Experimental results show that the contrast, sharpness, and EME of the restored images were improved by 80.68%, 30.89%, and 66.22% for one type of target, and by 34.32<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">%</sup>, 34.48%, and 104.76% for another type, respectively, compared to Schechner's method. The data indicate that this method effectively suppresses scattered light in different water quality environments, significantly enhancing the contrast and clarity of the images, with a noticeable improvement over traditional methods.