[1]邱晓晖,赵 阳.一种亮暗比例悬殊的脉冲噪声重污染图像去噪方法(英文)[J].南京师范大学学报(工程技术版),2008,08(03):056-60.
　Qiu Xiaohui,Zhao Yang.A Switching Median Filter for Extremely Corrupted Images With Great Disparity Between Low and High Intensity Impulse Noise[J].Journal of Nanjing Normal University(Engineering and Technology),2008,08(03):056-60.
点击复制

一种亮暗比例悬殊的脉冲噪声重污染图像去噪方法(英文)
分享到：

南京师范大学学报（工程技术版）[ISSN:1006-6977/CN:61-1281/TN]

卷:
08卷
期数:
2008年03期
页码:
056-60
栏目:
出版日期:
2008-09-30

文章信息/Info

Title:
A Switching Median Filter for Extremely Corrupted Images With Great Disparity Between Low and High Intensity Impulse Noise
作者:
邱晓晖1;  2;  赵 阳3
1. 中国矿业大学信息与电气工程学院, 徐州221008; 2. 南京邮电大学通信与信息工程学院, 江苏南京210003;
3. 南京师范大学电气与自动化工程学院, 江苏南京210042
Author(s):
Qiu Xiaohui1; 2; Zhao Yang 3
1.College of Electrical Engineering,China University of Mining and Technology,Xuzhou 221008,China;2.College of Communication Engineering,Nanjing University of Post and Telecom,Nanjing 210003,China;3.School of Electrical and Automatization Engineering,Nanj
关键词:
图像处理;  图像去噪;  中值滤波;  脉冲噪声
Keywords:
im age processing;  im age deno ising;  med ian filter;  impu lse no ise
分类号:
TP391.41
摘要:
虽然目前边界分离噪声检测(BDND)算法被普遍认为是一种较好的去除图像脉冲噪声方法,然而研究发现当"盐"或"胡椒"噪声污染比例之一超过50%,其性能却急剧下降,几乎无法去噪.为此在BDND算法基础上提出了改进的噪声检测方法(Mod ified BDND),对每个子图,新方法选取与"128"最接近的灰度值作为分界点,从而避免原方法因选取中值为分界点而可能落入污染区域的问题.该方法既可保留BDND算法优点,而且还能明显提高此时图象的重建效果.通过对Lena图象仿真处理说明了本算法的有效性和优越性.
Abstract:
Though boundary d iscrim inative no ise detection ( BDND) m ethod is frequently considered as an effic ient m ethod for image deno ise, how eve rwhen co rruption density o f salt or pepper no ise of im age corrupted by im pu lse no ise is larg er than 50%, research shows tha t th ism ethod is no long er va lid because its perform ance becom es ser iously deter iorated. Thus am od ified BDND m ethod is proposed in this case. In th is m ethod the nearest intensity va lue to 128 is se lected as the separating po int in each sub image instead o f them ed ian be ing separating po int in or ig inal BDND m ethod, so as to avo id separating po int selected in co rrupted area. The new m ethod can obv iously im prove PSNR va lue o f reconstruction im ages as we ll as rem a in the advantages o f the or ig ina l BDND w ith a lm ost no add ition o f com plex ity.

参考文献/References:

[ 1] Eng H L, M a K K. Noise adaptive so ft- sw itching m ed ian filter[ J] . IEEE Trans on Im ag ing Processing, 2001, 10( 2): 242-251.
[ 2] A sto la J, Kuosmanen P. Fundam enta ls o f Non linear D ig ita l Filtering[M ]. Roca Raton FL: CRC, 1997.
[ 3] Lee K C, Jong H, Sohn K H. Detection) estim ation based approach for im pu lsive no ise rem oval[ J] . E lectron ics Letters,1998, 34( 6): 449-450.
[ 4] Lin T C, Yu P T. Thresho ld no ise- free ordered m ean filter based on dempster- sha fer theory fo r im ag e restoration[ J]. IEEE T rans on C ircuit and System s, 2006, 53( 5): 1 057-1 064.
[ 5] Chang C C, h isao J Y, H sieh C P. A fast no ise reduction m e thod based on hum an v isua l system [ C ] / / IC ICS- PCM. S ingapore,2003( 15-18) : 1879-1882.
[ 6] Ng P E, M a K K. A sw itch ing m edian filter w ith boundary discr im ina tive no ise detection fo r ex trem e ly co rrupted im ages[ J]. IEEE Trans on Im ag ing Processing, 2006, 15( 6) : 1 506-1 516.

相似文献/References:

[1]王亮,黄飞,冯少彤,等.一种改进的二值图像像素标记算法[J].南京师范大学学报(工程技术版),2006,06(02):023.
　WANG Liang,HUANG Fei,FENG Shaotong,et al.An Improved Pixel Labeled Algorithm in Binary Image[J].Journal of Nanjing Normal University(Engineering and Technology),2006,06(03):023.
[2]陈 扬,曾 诚,程 成,等.一种基于CNN的足迹图像检索与匹配方法[J].南京师范大学学报(工程技术版),2018,18(03):039.[doi:10.3969/j.issn.1672-1292.2018.03.006]
　Chen Yang,Zeng Cheng,Cheng Cheng,et al.A CNN-based Approach to Footprint Image Retrieval and Matching[J].Journal of Nanjing Normal University(Engineering and Technology),2018,18(03):039.[doi:10.3969/j.issn.1672-1292.2018.03.006]
[3]陈宇航,卢仕兵,胡军军,等.循环流化床提升管中团聚物的结构及发展过程[J].南京师范大学学报(工程技术版),2020,20(03):015.[doi:10.3969/j.issn.1672-1292.2020.03.003]
　Chen Yuhang,Lu Shibing,Hu Junjun,et al.Structure and Development of Clusters in the Circulating Fluidized Bed Riser[J].Journal of Nanjing Normal University(Engineering and Technology),2020,20(03):015.[doi:10.3969/j.issn.1672-1292.2020.03.003]

备注/Memo

备注/Memo:
Foundation item: Supported by PostdoctoralFund of J iangsu Prov ince( 2007- 337 ) .
C orresponding auther: Q iu X iaohu,i Ph. D, associate professor, m ajored in d ig ital signal p rocess ing. E-m ail: q iuxh@ n jupt. edu. cn
更新日期/Last Update: 2013-04-24