arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1806.01161physics.ins-dethep-ex
收藏

基于极大似然法优化散射径迹数据的μ子层析成像改进

Muon Tomography imaging improvement using optimized scattering tracks data based on Maximum Likelihood Method

Xiao-Dong Wang, Kai-Xuan Ye, Yi Wang, Yu-Lei Li, Xie Wei, Ling-Yi Luo, Guo-Xiang Chen

摘要:

基于μ子射线成像的形式化原理,以多次库仑散射(MCS)为信息源,提出了一种基于库仑点近似算法(PoCA)的高Z材料重建图像的方法。PoCA算法中自然μ子的通量和单次散射假设是导致重建图像精度不高的两个因素。本文根据MCS方法得到的μ子径迹的类高斯分布特性,采用极大似然方法预测μ子的最佳出射径迹。采用受试者工作特征(ROC)和局部化ROC(LROC)两种分析方法评价重建图像的质量。从仿真结果来看,极大似然算法可以很好地实现纵向材料的区分,极大似然算法预测的区分率比PoCA算法预测的区分率高15%左右。实验结果表明,极大似然法可以大大提高重建图像的精度。

英文摘要:

Point of colsest Approche algorithm (PoCA) based on the formalism of muon radiogra- phy using Multiple Coulomb scattering (MCS) as information source is previously used to obtain the reconstruction image of high Z material. The low accuracy of reconstruction image is caused by two factors: the flux of natural muon and the assumption of single scattering in PoCA algo- rithm. In this paper, the maximum likelihood method based on the characteristics of Gaussian-like distribution of muon tracks by MCS is used to predict the optimal track of outgoing muon. The receiver operating characteristic (ROC) and the localization ROC (LROC) are used as two analysis methods to evaluate the quality of reconstruction image. From the results of simulation, the perfect discrimination of longitudinal materials could be well achieved by maximum likelihood algorithm and the discriminate ratio that is predicted by the maximum likelihood method is about 15% higher than that of predicted by PoCA algorithm method. It is seen that the maximum likelihood method can greatly improve the accuracy of reconstruction image.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑