彩色图像的模糊测度快速对等组滤波方法  

Fast Peer Group Filter Based on Fuzzy Metrics in Color Images

作　　者：钟灵[1] 章云[1] 

机构地区：[1]广东工业大学自动化学院,广州510006

出　　处：《小型微型计算机系统》2013年第1期181-184,共4页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(U0735003)资助;教育部高等学校博士学科点专项科研基金项目(20070562005)资助;广东省自然科学基金项目(07117423)资助;广州市科技计划项目(12C22111580)资助

摘　　要：为消除彩色图像中的脉冲噪声并提高矢量中值滤波方法的计算效率,提出一种新的模糊测度下的快速滤波方法.该方法是一种矢量中值滤波和等值操作之间开关型滤波算法.在模糊测度下新方法扩展原对等组的概念,采用了滤波窗口内中心像素邻近的对等组元素个数来判断噪声的可能.通过设计的对等组新计算方式,该方法能够有效减少对等组内的矢量距离计算次数.仿真实验表明新方法相较原快速对等组滤波方法能减少近50%的计算量,并能够在噪声抑制和细节保持做到较好的平衡.To improve the computation efficiency of vector median filter for impulsive noise removal in color images,a new fast peer group filter based on fuzzy metric was introduced.The new approach is a switching vector filter between the classical vector median filter and identity filter.The concept of peer group based on fuzzy metric is employed to judge the noise probability from the cardinality of the fuzzy peer group nearby the central pixel in filtering window.The new method can also effectively reduce vector distance computation compared with the original fast peer group filter（FPGF）.Simulation experiment results indicate that the new proposed filter can reduce about 50% computation compared with the original fast peer group filter and also achieves better trade-off between noise suppression and signal-detail preservation.

关 键 词：矢量中值滤波 开关型滤波 脉冲噪声 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]