摘要
提出了基于人工神经网络的γ能谱定量分析方法。通过对标准源γ能谱的线性组合,建立了γ能谱分析网络训练样本集,并对RBF人工神经网络进行了训练。对已知标准点源不同测量时间的组合γ能谱和标准环境样品γ能谱进行了仿真分析,分析结果的相对误差分别小于3%和小于8%。研究表明,基于RBF人工神经网络的γ能谱分析方法是有效可行的。
A method of analyzing γ- spectra based on RBF ANN was introduced. The training sample sets were established by linearly combining the spectra of the standard radioactive sources that contain a single nuclide composition respectively,and the RBF ANN was trained with these sample sets. The combination spectra by different collecting time of the standard point sources and those of a standard environmental sample were analyzed.The deviation was less than 3% and 8% respectively. The results show that the method of analyzing γ- spectra based on RBF ANN is effective and feasible.
出处
《核电子学与探测技术》
CAS
北大核心
2016年第1期56-59,共4页
Nuclear Electronics & Detection Technology
基金
辽宁省教育厅科学技术研究项目(No.L2013410)资助
关键词
人工神经网络
RBF网络
γ能谱分析
线性时不变系统
线性叠加原理
artificial neural network
γ spectra
RBF network
linear time invariant system
principle of linear superposition
