A Polarization Image Fusion Approach Using Local Energy and MDLatLRR Algorithm

Polarization is one of the inherent properties of light that human eyes cannot recognize. The polarized image collected by the optical system can be used for target detection, image enhancement, and dehazing. Nevertheless, a single image only represents certain polarization information on the targets. The fusion of images representing different polarization information is more beneficial to improve the image quality. Therefore, in this paper, we utilize a multiscale decomposition method to decompose the polarization degree image and intensity images into a base layer and detail layers. The local energy method is then used to fuse the base layer, and the nuclear norm is utilized to combine the detail layers. The fused base and detail layers are reconstructed to obtain the final fusion image. The proposed method is analyzed and compared to six state-of-the-art algorithms to exactly evaluate how effective it is. Experiments show that our method increases the contrast and clarity of the fusion image and has excellent details and texture. In addition, the outline of the fused image is clear, and the target is prominent, which is conducive to the target recognition.