Research On Image Defogging Algorithm Based On Atmospheric Scattering Model

Accurate extraction of image features is a key factor that directly affects the performance of computer vision systems.However,in the hazy weather conditions,the light scattering caused by the atmospheric particles causes the image quality to be seriously degraded,and many features of the blurred image are masked.Therefore,improving the image quality and enhancing the robustness of the dehazing algorithm under hazy weather conditions have important scientific significance and wide application value.In this thesis,the image dehazing algorithm based on the atmospheric scattering physical model is studied,and the improvements are presented from the perspectives of recovering more detailed information and improving the efficiency of the dehazing algorithm.The specific contents are:The algorithm of the dark channel prior defogging is analyzed in depth.Aiming at the problem that existing algorithms can't preserve the details of image,a defogging algorithm based on the dark channel and MRF theory is proposed to effectively remove the fog and improve image visibility.Firstly,this algorithm uses the POSHE algorithm to enhance the foggy image for improving contrast,expanding dynamic range of its histogram,and the transmission map is obtained with the DCP algorithm.Secondly,the constraint characteristics of Markov random field model are adopted to model the transmission map for refining it.Then,based on the salient features of the sky region,the atmospheric light is calculated with the layered search method.Finally,the restored image was obtained according to the atmospheric scattering model.The experimental results show that: the algorithm can get vivid defogging iamge,which enhances the naturalness of iamge,and the visual of image is better.Aiming at the problem that the existing algorithm has low defogging efficiency,an image dehazing algorithm based on linear transformation and bilateral filtering is proposed.Firstly,the algorithm studies the geometric information of the restored image,and the fog image is filtered by the minimum filter.Secondly,analyzing the relationship of the minimum channel map between the fog image and the restored image,the transmission map is quickly obtained by the linear transformation theory.Then,the transmission map of the highlight is modified,and bilateral filtering is used to smooth it,so that the obtained transmission map has more details.Finally,a clear image is restored with the atmospheric scattering model.Experimental results show that the algorithm can effectively remove the fog,and can also restore the image details.