The Study On Single Image Defog Based On Atmospheric Scattering Model

At present, a large number of outdoor surveillance equipment are widely used inintelligent traffic monitoring, military reconnaissance, remote sensing and other fields. In theprocess of outdoor image acquisition, image is easily affected by bad weather, leading toserious degradation of obtaining image quality and extracting the characteristics of thejudgment, greatly reduces the image application value.Fog removal for degraded image is afundamental and hot problem in computer vision, which has a wide application prospects.This paper analyzes the physical process of image formation in the foggy weatherconditions. This article describes the atmospheric scattering model. This paper surveyed therecent techniques for image dehazing from the point view of physical model and digital imageprocessing. And two typical defog approach to atmospheric scattering model was introduced.According to the characteristics of atmospheric light, this paper proposes that in YCbCrspace Y channel using quad-tree subdivision searching method and dichotomy method toestimate atmospheric optical value; Combined with the atmosphere scattering model andatmospheric scattering characteristics, this paper puts forward two kinds of methods usingcomposite filter and wavelet transform image fusion method to defog. the composite filterincludes anisotropic diffusion filter and adaptive median filter to defog. It can be proved fromthe experimental results that the proposed approach can achieve better visual results. At thesame time, from the objective angle evaluation, our methods in performance is superior tosome previous methods.