基于边缘检测暗通道先验的全变差图像去雾算法  

A total variation image dehazing algorithm based on dark channel prior with edge indicator

作　　者：马悦 金正猛[1] 冯灿 MA Yue;JIN Zhengmeng;FENG Can(College of Science,Nanjing University of Posts and Telecommunications,Nanjing 210023,China;North Information Control Research Academy Group Co.Ltd.,Nanjing 211153,China)

机构地区：[1]南京邮电大学理学院,江苏南京210023 [2]北方信息控制研究院集团有限公司,江苏南京211153

出　　处：《南京邮电大学学报（自然科学版）》2023年第4期47-57,共11页Journal of Nanjing University of Posts and Telecommunications：Natural Science Edition

基　　金：国家自然科学基金(12271262);南京邮电大学校级自然科学基金(NY221097)资助项目。

摘　　要：为了恢复有雾图像中更多如边缘、结构等有价值的图像细节信息,文中引入带边缘检测的暗通道先验来估计初始场景深度,提出基于边缘增强的全变差图像去雾模型,证明该模型极小化问题解的存在性和唯一性。进一步,结合原始⁃对偶方法,设计该模型的快速数值求解算法,并给出了该算法的收敛性结果。最后,数值实验结果验证所提算法的可行性与有效性。In order to recover more valuable image details such as edges and structures in hazy images,this paper introduces a dark channel prior with the edge indicator function to estimate the initial scene depth,and proposes a total variation image dehazing model based on edge enhancement.Then,the existence and uniqueness of the solutions to the minimization problem of this model are proved.Further,we design a primal⁃dual algorithm to solve the model and provide its convergence result.Finally,numerical results verify the feasibility and effectiveness of the proposed algorithm.

关 键 词：图像去雾 暗通道先验 边缘检测函数 原始⁃对偶算法 

分 类 号：TN911.73[电子电信—通信与信息系统]