Research of polarized image defogging technique based on color space conversion

The traditional image defogging algorithm based on polarization characteristic is mostly optimized for polarization degree, atmospheric light intensity and depth information, but the traditional defogging algorithm based on polarization characteristics in heavy fog is not satisfactory. In this paper, a new color space conversion algorithm using dark channel prior for polarization image defogging was proposed. This method can effectively remove the haze in heavy fog, we analyze the characteristics of image de-fog restoration under low visibility, and propose a special feature of the algorithm to enhance the image from the polarization information to the color space transformation. In order to combine the polarization information and defogging model, we firstly converted the polarization information into HIS color space and then mapped the HIS color space to the RGB space. Secondly we obtained the image dark channel with the combination of the atmospheric scattering model in haze weather. Finally we elaborated the atmospheric transmission rate using a Guide filter algorithm based on sparse prior of the image. Compared with the traditional image restoration results based on polarization characteristics, the algorithm we proposed is effective, the details of the reconstructed image are clearer which can effectively identify the target of distant scene in low visibility scene.