Research On Image Defogging Algorithm Based On Physical Imaging Model

In foggy weather,because of the scattering effect of tiny particles such as water droplets and particles in the air on atmospheric light,it will cause great interference to the outdoor image information acquisition system.The collected image information lacks integrity,and the quality of the image declines seriously,which causes serious trouble to the detection and identification of useful information of the image.Therefore,it is necessary to improve the image quality and restore the image clarity through technical means.Dark channel prior algorithm is one of the important algorithms for image defogging,but it is easy to cause some problems such as regional distortion,halo effect,dark image after defogging,etc.Aiming at the shortcomings of this algorithm,the following improvement methods are proposed.Firstly,based on analyzing the causes of fog image formation and image quality degradation,the image defogging algorithm based on physical imaging model is studied,and the physical meaning and defogging principle of atmospheric scattering model are clarified.The image quality evaluation system is introduced to provide theoretical basis for image quality evaluation.Secondly,aiming at the ineffectiveness of dark channel prior algorithm in sky area,a defogging algorithm based on sky recognition is proposed.Through the minimum variance mapping combined with region growing segmentation algorithm,the sky region of the image is accurately identified,and the atmospheric light value is optimized in this region;Through the sky area transmittance correction model,combined with Gaussian blur function,the sky area transmittance is optimized,and the global transmittance is obtained by fusing it with the non-sky area transmittance;The atmospheric scattering model is used to restore the image,and the adaptive gamma correction is combined to improve the brightness of the image.Then,in order to further improve the defogging ability of different scene images,a defogging algorithm based on two-channel and image quality evaluation model is proposed.Using two-channel prior theory,the atmospheric light value is accurately optimized;Combining the segmentation theory with morphological operation,the light and dark areas of the image are accurately segmented,and their respective transmittance is obtained by different optimization strategies in these two areas,and the global transmittance is obtained by fusing them;An image quality evaluation model is constructed,and the best fusion coefficient is obtained iteratively to recover the fog-free image with the best quality.Finally,the defogging effect of this algorithm is verified by experiments,and compared with other algorithms in subjective and objective evaluation.The results show that this paper can effectively remove the fog interference in the image,improve the quality and clarity of the image,and has strong robustness and universality,the problems of area distortion,halo effect and dark image after defogging caused by dark channel prior defogging algorithm are solved.There are 41 figures,24 tables and 60 references in this paper.