基于非局部均值去噪算法的PM模型及其Split-Bregman迭代实现  

PM Model Based on Nonlocal Means and Its Implementation via Split-Bregman Iteration

作　　者：胡杰[1] 潘振宽[1] 魏伟波[1] 

机构地区：[1]青岛大学信息工程学院,山东青岛266071

出　　处：《青岛大学学报（自然科学版）》2010年第1期54-57,49,共5页Journal of Qingdao University(Natural Science Edition)

基　　金：新世纪优秀人才支持计划(NCET-05-0601)

摘　　要：基于偏微分方程能量变分模型中规则项的设计,将传统的基于梯度的PM模型与非局部均值算法相结合,对加性噪声进行去噪在最小化非局部PM模型能量函数的过程中,分别运用原始算法和Split-Bregman迭代算法进行了数值运算,结果表明,两种方法均取得到了较好的修复效果。Based on the energy variation model of partial differential equation, a designation to the regula- tion item that combined by traditional gradient-based PM model and nonlocal means algorithm is introduced to the denoising on additive noise. In the processing of minimizing energy function of the nonlocal PM model, the original algorithm and Split-Bregman iterative algorithm was used to numerical computing, respectively. Results of experiments show that both of them can get superior results.

关 键 词：非局部均值 偏微分方程 非局部PM模型 Split—Bregman迭代 图像修复 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]