突出结构联合L0先验的模糊图像盲复原方法  

BLIND IMAGE RESTORATION METHOD BASED ONPROTRUDING STRUCTURE COMBINED WITH L0 PRIOR

作　　者：高如新[1,2,3] 朱新柳 郭凤云 李雪颖 Gao Ruxin;Zhu Xinliu;Guo Fengyun;Li Xueying(School of Electrical Engineering and Automation,Henan Polytechnic University,Jiaozuo 454000,Henan,China;Henan Key Laboratory of Intelligent Detection and Control of Coal Mine Equipment,Jiaozuo 454000,Henan,China;Henan International Joint Laboratory of Direct Drive and Control of Intelligent Equipment,Jiaozuo 454000,Henan,China;Xinxiang Branch of China United Network Communications,Xinxiang 453000,Henan,China;School of Mechanical and Electrical Engineering,Yellow River Transportation College,Jiaozuo 454950,Henan,China)

机构地区：[1]河南理工大学电气工程与自动化学院,河南焦作454000 [2]河南省煤矿装备智能检测与控制重点实验室,河南焦作454000 [3]河南省智能装备直驱技术与控制国际联合实验室,河南焦作454000 [4]中国联合网络通信有限公司新乡分公司,河南新乡453000 [5]黄河交通学院机电工程学院,河南焦作454950

出　　处：《计算机应用与软件》2025年第3期190-195,共6页Computer Applications and Software

基　　金：国家自然科学基金项目(61903126);河南省科技攻关项目(212102210503)。

摘　　要：为了解决结构类模糊图像在边缘模糊核估计方法中复原效果较差这一问题,提出突出结构联合L0先验的模糊图像盲复原方法。利用模糊图像的突出结构估计图像边缘,根据模糊核估计的数学模型来恢复中间潜像并代替模糊图像作为新的输入,结合L0正则化先验约束模型求解得到模糊核和清晰图像。实验结果表明,提出的方法可较好地复原结构类模糊图像和大尺度模糊核运动模糊图像,抑制振铃效果显著,在复原效果上优于其他方法。In order to solve the problem of poor restoration effect of structural blur image in edge blur kernel estimation method,a blind restoration algorithm ofprotruding structure and L0 prior is proposed in this paper.The edge of the image was estimated by using the prominent structure of the blur image,and the latent image of the middle was recovered according to the mathematical model of the blur kernel estimation,which replaced the blur image as the new input.The blur kernel and the clear image were solved by combining the L0 regularization prior constraint model.The experimental results show that the proposed method can restore the motion blurred image of structural blur image and large-scale blur kernel well,and the ringing suppression effect is remarkable,and the restoration effect is better than other algorithms.

关 键 词：盲复原方法 突出结构 L0先验 数学模型 模糊核 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]