基于边缘增强总变分正则化盲泊松图像反卷积  

Blind Poissonian image deconvolution with edge enhancing total variation regularization

作　　者：时愈[1,2] 洪汉玉[1,2] 华夏[1,2] 

机构地区：[1]武汉工程大学电气信息学院,湖北武汉430205 [2]武汉工程大学图像处理与智能控制实验室,湖北武汉430205

出　　处：《华中科技大学学报（自然科学版）》2016年第8期94-98,共5页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：国家自然科学基金重点资助项目(61433007);湖北省自然科学基金创新群体资助项目(2012FFA046);武汉市国际科技合作计划资助项目(2014030709020310);湖北省教育厅科技研究项目(Q20161504)

摘　　要：针对总变分的方法支持分片常值解,易导致图像内容中的平坦区域出现'阶梯效应'和纹理及部分细节随着噪声消失等问题,提出基于边缘增强总变分正则化的盲泊松图像反卷积方法.受边缘信息对视觉感知的重要性的影响,提出具有空间自适应性的边缘增强指示算子约束总变分正则化项,并采用双边滤波区分图像内容中的边缘和平坦区域.一系列仿真和真实模糊图像实验表明:提出的方法与其他方法相比具有较好的平滑平坦区域和保边缘的特性.The model of total variation method has certain shortcomings,it favors a piecewise constant solution,which leads to staircase effects in flat regions,and some textures and small details are often diminished with noise.In order to solve this problem of image content,blind Poissonian image deconvolution with edge enhancing total variation regularization method was proposed.Motivated by the fact that edge information plays an important role in visual perception,the edge-enhancing indicator with spatial adaptability was proposed to constrain the total variation regularization,and the bilateral filter was introduced to distinguish flat region and edge information.A series of experiments on simulated and real blurring images show that the proposed method can smooth flat region preferably and has the characteristic of edges preserving.

关 键 词：盲反卷积 泊松噪声 边缘增强 总变分 边缘保持 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]