A Restoration of Underwater Polarized Images Based on DCP

Abstract

To address the problems of the details blurring and color distortion existing in the multi-angle underwater polarized images, the paper proposes a restoration algorithm of underwater polarized images based on DCP(dark channel prior), which combines the characteristics of underwater polarized images with the DCP. Firstly, the transmittance and the object reflected light with polarization characteristics of two Orthogonal polarized images are calculated by the improved DCP. Then, the polarization components existing in the object reflected lights are eliminated by the global polarization angle to obtain the original object radiance. Finally, the color correction is carried out by gray world algorithm to get the final restoration result. The proposed method is mainly designed to solve the problems of the limited utility of DCP in the deblurring of single underwater image and the noise caused by conventional polarized restoration methods. The experimental results show that the proposed method can make the details of images more obvious and significantly improve the clarity, contrast and the color distribution of underwater polarized images.