基于双能X射线的有效原子序数提取算法研究  

An Algorithm for Extracting Effective Atomic Number of Materials Based on Dual-Energy X-Ray System

作　　者：常青青[1] 陈嘉敏[1] 沈天明[1] 李维姣[1] 

机构地区：[1]公安部第三研究所,上海200031

出　　处：《计算机仿真》2014年第2期297-301,共5页Computer Simulation

摘　　要：研究X射线透视安检系统设计问题,实现有机物细分对于违禁物识别具有非常重要的意义。由于传统物质识别方法中存在的无法细分有机物的问题,利用双能X射线的衰减特性,提出一种新的有效原子序数提取算法,进而能够有效识别出多种有机物。改进方法首先将待检物质等效为两种不同厚度的基底材料的叠加,然后利用改进的基底材料分解技术提取出有效原子序数,最后针对液体材料的包装,去除容器外壁的影响,更准确计算出有效原子序数,为违禁物品的识别提供更准确的判别标准。实验结果表明,改进算法能够克服厚度的影响,快速准确地提取出物质的有效原子序数,相对误差在8%以内,能够实现对有机物的细分,满足安检设备的实际需求。In X -ray security inspection systems, the accurate classification of different organic materials is im- portant for contraband identification. In this paper, we proposed a new algorithm to extract effective atomic number from dual - energy transmission images based on the attenuation characteristic of X - ray. Firstly, a basic material decomposition （BMD） technology was employed to decompose the target material as the combination of two basic ma- terials. Then the equivalent atomic number can be computed based on the thickness of basic materials. Finally, the containers for the liquid were identified and the corresponding effective atomic number was refined after the effect of the container was removed. Experimental results show that the algorithm can extract the effective atomic number fast regardless of the thickness of different materials. The relative error is 〈 5% and the algorithm can be used for de- tailed classification of organic materials in real systems.

关 键 词：透视双能图像 有效原子序数 基底材料分解 液体分类 

分 类 号：TN911.72[电子电信—通信与信息系统]