滤除椒盐噪声的改进中值滤波算法比较  

Comparison of Improved Median Filters for Removing the Salt-and-Pepper Noise in Digital Images

作　　者：王波[1] WANG Bo(School of Mathematics and Information Science,Guiyang University,Guiyang 550005,China)

机构地区：[1]贵阳学院数学与信息科学学院

出　　处：《贵阳学院学报（自然科学版）》2019年第3期83-87,共5页Journal of Guiyang University：Natural Sciences

基　　金：贵阳市科学技术局贵阳学院科技专项资金(项目编号:GYU–KYZ(2019-2020)PT04-09)资助项目

摘　　要：用标准中值滤波去除数字图像中的椒盐噪声时,会丢失图像的边缘细节,且噪声密度较大时会失效。基于此,大量的改进方法被提出。对几种典型的改进中值滤波方法进行了仿真实验,并完成了定量分析。实验结果表明,在滤除椒盐噪声的问题上,基于决策的算法滤波效果优于开关中值滤波;在噪声密度较大时,自适应中值滤波优于基于决策的算法,但时间复杂度较高。认为在基于决策的算法中引入自适应窗口的方法,能在不明显提高时间复杂度的前提下有效地提高降噪效果,对进一步改进中值滤波具有一定的指导意义。When standard median filter is employed to remove salt-and-pepper noise,the edge details of the image will be lost,and when used to deal with a high density noisy image,it will even be invalid.Therefore,lots of improved median filters had been proposed.In this paper,some typical improved median filters are implemented,and the experimental results are analyzed quantitatively.The results show that on the problem of eliminating salt-and-pepper noise,the DBAs have better filtering effect than the SMFs,while when the noise density is high,the AMFs perform better than the DBAs,though the computing time complexity is high.So,we believe that the method of introducing adaptive window into the DBAs can effectively improve the noise reduction effect without obviously increasing the time complexity,and it has certain guiding significance for further improvement of median filtering.

关 键 词：噪声滤除 椒盐噪声 开关中值滤波 基于决策的算法 自适应中值滤波 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]