一种基于模糊决策的开关矢量中值滤波方法  

Switching vector median filter based on fuzzy decision

作　　者：张辰锐[1] 樊养余[1] 刘姝[1] 雷涛[1,2] 王剑书[1] 

机构地区：[1]西北工业大学电子信息学院,西安710129 [2]兰州交通大学电子与信息工程学院,兰州730070

出　　处：《计算机应用研究》2013年第12期3870-3873,共4页Application Research of Computers

基　　金：国家自然科学基金资助项目(61202314);中国博士后科学基金资助项目(2012M521801);西北工业大学研究生创业种子基金资助项目(Z2013065)

摘　　要：针对现有的彩色图像脉冲噪声去除方法没有区分滑动窗口中的像素是否为噪声像素而导致滤波效果差的问题,提出一种基于模糊决策的开关矢量中值滤波方法。该方法首先利用开关条件判断像素是否被污染,针对被污染的像素,通过模糊数学理论构造适合脉冲噪声去除的隶属函数;然后计算滑动窗口内所有像素的模糊隶属度,并根据置信区间去除疑似噪声像素以优化滑动窗口的取值空间;最后对优化后的滑动窗口应用矢量中值滤波(VMF)以去除噪声像素。与现有方法相比,新的方法去除了滑动窗口中心像素的邻域疑似噪声,从而有效提升了滤波效果。实验验证了该方法的高鲁棒性和实用性。According to the fact that conventional color image filtering algorithms do not distinguish whether the pixels in the sliding window are noise pixels and thus leading to poor filtering results, this paper proposed a switching vector median filtering method based on fuzzy decision. Firstly, it used the switching conditions to determine whether the center pixel of a neighborhood was noisy. For the corrupted pixels in color image,it constructed a membership function for removal of noise pixels according to fuzzy mathematical theory, estimated all the pixels＇ memberships in the sliding window according to the proposed membership function. And then it removed some pixels with high membership in terms of the confidence interval. Finally, it employed VMF in the optimized sliding window. Compared with the existing method, the novel method could improve the filtering results by removing the suspected noise pixels in the sliding window. Experimental results show that the proposed method performs better performance of robustness and practicality.

关 键 词：彩色图像处理 矢量中值滤波 模糊隶属度 鲁棒性 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]