一种基于开关型加权中值滤波的随机脉冲噪声去除方法  

A Random-value Impulse Noise Removal Method Based on Switching Weighted Median Filter

作　　者：史再峰 许泽昊 庞科 曹清洁 姚素英[1] Shi Zaifeng;Xu Zehao;Pang Ke;Cao Qingjie;Yao Suying(School of Microelectronics,Tianjin University,Tianjin 300072,China;School of Computer Science and Technology,Tianjin University,Tianjin 300072,China;School of Mathematical Sciences,Tianjin Normal University,Tianjin 300387,China)

机构地区：[1]天津大学微电子学院,天津300072 [2]天津大学计算机科学与技术学院,天津300072 [3]天津师范大学数学科学学院,天津300387

出　　处：《南开大学学报（自然科学版）》2018年第4期63-67,共5页Acta Scientiarum Naturalium Universitatis Nankaiensis

基　　金：国家高技术研究发展计划(2012AA012705);国家自然科学基金(61674115)

摘　　要：依据图像中随机脉冲噪声像素奇异性的特点,提出了1种新的开关型加权中值滤波算法.首先通过中值滤波的方法对被检测像素邻域进行平滑处理,并定义平均绝对差值进行噪声点的判别;对检测出的噪声像素点采用改进的加权中值滤波进行处理,即仅利用滤波窗口中未受污染的像素进行噪声滤除,加权系数同时包括空间距离与灰度的差异性.实验结果表明,提出的算法既有较高的脉冲噪声检测准确率,又能较好地滤除图像中的随机脉冲噪声.According to the singular characteristics of impulse noise, a new switching median filter was put forward to remove random-value impulse noise. The median filter was first applied for smoothing noise around the central pixel and the mean absolute difference was defined to distinguish whether the detected pixel was noisy. Then an improved weighted median filter which only chose unpolluted pixels was proposed for noise reduction. The spatial distance and gray-value difference were both taken into account when calculating weighted coefficient. The experiment results show that the proposed algorithm not only produces higher detection accuracy but also exhibits better noise removal capability.

关 键 词：随机脉冲噪声 中值滤波 噪声检测 噪声滤除 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]