能量权重图法快速核素识别

Fast Nuclide Identification Based on Energy Weight Graph

在序贯贝叶斯算法相关研究的基础上,提出了能量权重图法快速核素识别方法。基于探测器的实测能谱响应,通过并道、平滑等能谱预处理生成能量权重图,从而搭建Co处理器和Cs处理器以并行判别60Co和137Cs的放射源。针对能谱抽样生成的模拟数据进行处理器的参数调整,并在一套1 Gs/s(1 Gs=10-4 T)的数字化LaBr3探测系统中开展了初步的实验验证。处理器平均使用不到100个光子就能实现核素识别,且虚警率低于2%,漏检率低于1%。

Based on the research of sequential bayesian algorithm,a new method based on energy weight graph is proposed.The energy weight graph is generated by the energy spectrum response of the detector and the curve is smoothed by combined calculation.The Co processor and the Cs processor are made to determine if 60Co and 137Cs exist in parallel.The simulation data generated by the energy spectrum sampling is used to adjust the processor parameters.The measured data of the 1 Gs/s digital LaBr3 detector preliminarily verifies the processor performance.The processor uses less than 100 photons on average to make a discriminant,and the false alarm rate is less than 2%,and the missed detection rate is less than 1%.