改进型NL-means算法滤除快中子图像噪声  

Improved NL-means Algorithm Filters Fast Neutron Image Noise

作　　者：刘斌[1] 刘耀光[1] 尹伟[1] 王胜[1] 霍合勇[1] 吴洋[1] LIU Bin;LIU Yao-guang;YI Wei;WANG Sheng;HUO He-yong;WU Yang(Institute of Nuclear Physics and Chemistry,CAEP,Mianyang,Sichuan 621000,China)

机构地区：[1]中国工程物理研究院核物理与化学研究所,四川绵阳621000

出　　处：《核电子学与探测技术》2020年第2期298-302,共5页Nuclear Electronics & Detection Technology

基　　金：国防科工局基础科研(项目编号Jcky2017212b002)资助。

摘　　要：分析了快中子照相图像的主要噪声特点和传统NL-means滤波算法的不足,提出了一种改进的适用于快中子照相特点的Nl-means图像滤波算法。实验结果表明,改进之后的算法不仅可以有效滤除快中子照相图像中的大量白疵点,而且有效的保护了图像中的边缘与细节信息,可为快中子照相图像的进一步处理提供参考。This paper analyzes the main noise characteristics of fast neutron radiography images and the shortcomings of the traditional NL means filtering algorithm,and proposes an improved NL-means filtering algorithm suitable for fast neutron radiography.Experimental results show that the improved algorithm can not only effectively filter out a large number of white defects in fast neutron radiography images,but also effectively protect the edges and details of the images,which can provide reference for further processing of fast neutron radiography images.

关 键 词：快中子照相 噪声 白疵点 NL-means 

分 类 号：TL99[核科学技术—核技术及应用]