Defogging Method For Image Based On Polarization

Image dehazing is a hot topic in the research field of computer image processing,the defogging algorithm has mainly two categories, one is based on the conventionalimage enhancement techniques, the other is image restoration technique based onphysical optics model. This paper analyzes several algorithms of image enhancementand image restoration. In this paper, combined with the polarization and the fogimaging model, we design a demister model based on polarization properties of light.The main task is to calculate the atmospheric optical intensity on the infinitypoint,and to calculate the transmittance of image. In this paper, calculation “A” ofprecise value by He Kaiming dark channel theory and histogram as normaldistribution theory. what’s more, calculation of image transmission diagram rate byHe Kaiming dark channel theory. Finally, using the demister model to defog. Theexperiments prove that this algorithm can effectively on fog degraded imagerestoration, improve the image quality image saturation.Because this method need to use two images of polarization directionperpendicular to each other at the same time and the same scene, in order to reducethe difficulty of acquiring image, using the image registration algorithm to match twoimages of possible rotation or misalignment. In this paper，Use the SIFT-RANSACalgorithm to match two images. The experimental results show that the algorithm canbe very good registration of two images, to provide a guarantee for success for thefinal fusion to fog.Finally, subjective evaluation and objective evaluation methods are used in theexperiments to Compare the original image and the image processed.In thecomparison of this algorithm and histogram equalization algorithm, retinex algorithm,He Kaiming dark channel prior algorithm, our algorithm is better than otheralgorithms in both subjective and objective evaluation.