基于改进Richardson-Lucy算法的中子照相图像几何不锐度修正方法研究  

A geometric unsharpness correction method for neutron photographic image based on improved Richardson-Lucy algorithm

作　　者：伍华杰 甘佺 吕永强 汪启宏 伍一凡 宋婧 凤麟核团队 WU Huajie;GAN Quan;LYU Yongqiang;WANG Qihong;WU Yifan;SONG Jing;FDS Team(Hefei Institutes of Physical Science,Chinese Academy of Sciences,Hefei 230031,China;University of Science and Technology of China,Hefei 230026,China;SuperRay Technology Co.,Ltd.,Qingdao 266041,China;International Academy of Neutron Science,Qingdao 266199,China;Shandong Key Laboratory of Neutron Science and Technology,Qingdao 266199,China;不详)

机构地区：[1]中国科学院合肥物质科学研究院,合肥230031 [2]中国科学技术大学,合肥230026 [3]中科超睿技术有限公司,青岛266041 [4]中子科学国际研究院,青岛266199 [5]山东省中子科学技术重点实验室,青岛266199 [6]不详

出　　处：《核技术》2023年第12期30-35,共6页Nuclear Techniques

基　　金：青岛市科技计划园区培育计划项目(No.22-3-5-yqpy-3-qy);泰山产业领军人才工程资助。

摘　　要：中子照相系统因其特定的成像结构,使得成像后的中子照相图像不可避免地存在几何不锐度。为提升中子照相图像的质量,本文提出一种基于改进Richardson-Lucy算法的中子照相图像几何不锐度修正方法,该方法首先根据几何不锐度的形成原理构建点扩散函数,然后通过拉普拉斯算子和中值滤波方法去除图像中的γ白斑噪声,最后采用Richardson-Lucy算法对图像进行复原。采用线对模板样品进行测试,并与4种现有的中子照相图像几何不锐度修正算法进行对比,验证了本文算法可将衡量中子照相图像几何不锐度修正效果的平均梯度、空间频率指标分别提升60.23%和29.90%,能够实现几何不锐度的修正,同时解决图像复原过程中γ白斑噪声的放大问题,为实现高分辨率的中子照相提供重要技术支撑。[Background]Due to the specific imaging structure of neutron imaging system,the geometric unsharpness is unavoidable in neutron photographic images after imaging.[Purpose]This study aims to design a geometric unsharpness correction method for neutron photographic image based on improved Richardson-Lucy(RL)algorithm.[Methods]First,according to the geometric unsharpness of the neutron photographic image,the mathematical model of the point spread function(PSF)was established.Then,the Laplace operator and median filter were used to remove the PSF related gamma spots noise(GSN)in the image,and the neutron photographic image was restored by RL algorithm.Finally,the restoration quality of neutron photographic images of the line-pair sample was evaluated using average gradient(AG)and spatial frequency(SF).[Results]The results of the line-pair sample demonstrate that compared with four existing correction algorithms of geometric unsharpness,the proposed method improves AG and SF by 60.23% and 29.90%,respectively.[Conclusions]The proposed method of this study can effectively correct the geometric unsharpness caused by the amplification of gamma spots noise in the process of neutron photographic image restoration,providing an important technical support for performing high-resolution neutron radiography.

关 键 词：中子照相图像 图像复原 几何不锐度 Richardson-Lucy算法 高分辨率 

分 类 号：TL81[核科学技术—核技术及应用]