结合双边滤波与暗通道的图像保边去雾算法-中国图象图形学报.DOCVIP

双边滤波与暗通道结合的图像保边去雾算法 曾接贤，余永龙 南昌航空大学计算机视觉研究所，南昌330063现有主流基于模型的有雾图像复原算法基本都是对介质透射率进行直接求解，先再细化抠图，运算量利用双边滤波方式进行介质透射率的求解，用以简化去雾算法时间复杂度。方法该算法对有雾图像进行全局清晰化处理的同时恢复有雾图像的局部细节结果 双边滤波方式对介质透射率的求解，是先估算出较精确的大气散射函数及大气光值，然后间接求出，提高算法的时效性中图法分类号及文献标识码: TP391.7，A Image Defogging and edge preserving Algorithm based on dark channel prior and bilateral filtering Zeng Jiexian, Yu Yonglong Institute of Computer Vision, Nanchang Hangkong University, Nanchang, 330063, China Abstract: Objective: Researchers have previously focused on edge detail preservation and restoration in image defogging. An edge-preserving image defogging algorithm based on bilateral filtering and dark channel prior is proposed in this study. This algorithm can defog the image globally and restore the local detail of the fog image. Method: A classic model-based defogging algorithm was solved directly utilizing atmospheric transmittance. The atmospheric transmittance is estimated first, and then the intermediate processed images are refined through soft matting. However, soft matting has high computation complexity. The atmospheric transmittance is solved indirectly utilizing bilateral filtering to decrease the time complexity of the defogging algorithm. Result: The process of solving the atmospheric transmittance by bilateral filtering has two steps. First, the atmospheric scattering function and atmospheric light value are estimated. Transmittance is then solved indirectly. This scheme avoids the refining process of transmittance with soft matting and improves the computational efficiency. Conclusion: Experiment results show that the proposed algorithm can protect the edge details and improves time efficiency Key words: image defogging , dark channel, bilateral filtering , atmospheric scattering, edge preserving 0 引 言( 1 暗通道去雾 1.1 雾天图像退化模型 根据大气散射物理模型[8]可知，摄像机或是接收到的光线强度由两部分组成目标图像在传播过程中衰减 （）表示采集到的雾天图像表示要得到的无雾的清晰图像表示大气的透射率，光线的传播率表示大气光，一般用来表示无穷远处光照强度表示图像上点的位置坐标、、变量均与有关。 从研究无雾图像的统计特征出发，发现了一个在绝大多数无雾图像中广泛存在的一条规律——暗原色先验，也称