基于卷积神经网络的放射性核素识别算法
Radionuclide Identification Algorithm Based on Convolutional Neural Network
朱岳武 1梁杰 2董喆 1刘尔聃 1李林珊 2姜麟泉3
作者信息
- 1. 中国中原对外工程有限公司,北京 100044
- 2. 中国兵器装备集团自动化研究所有限公司智能测控事业部,四川 绵阳 621000
- 3. 北京联合大学,北京 100044
- 折叠
摘要
为实现对低计数、多种类的复杂放射性核素的准确识别,引入卷积神经网络(convolutional neural network,CNN)搭建针对低计数、多种类放射性核素识别模型.利用蒙特卡罗仿真创建由 241Am、133Ba、57Co、60Co、137Cs、152Eu 以及 40K 组成的单源、两源以及三源共 63 种不同种类放射性核素能谱数据库.利用仿真训练集和仿真验证集样本完成CNN训练及超参数优化,利用测试集样本验证模型性能.结果表明,该模型对低计数、多种类放射性核素具有良好的识别性能.
Abstract
To realize the accurate identification of low count and multi-class complex radionuclides,the convolutional neural network is introduced to build a recognition model for low count and multi-class complex radionuclides.The radionuclide energy spectrum databaseconsisting of 241Am,133Ba,57Co,60Co,137Cs,152Eu,and 40Kwas established by the Monte Carlo simulation,which contained a total of 63 different radioactive nuclide sources.The simulation training set and simulation verification set samples were used to complete the training and hyperparameter optimization of convolutional neural networks.The test set samples were used to verify the model performance.The results demonstrate that the convolutional neural networks has good recognition performancein the identification of low count and multi-class complex radionuclides.
关键词
放射性核素识别/卷积神经网络/蒙特卡罗仿真Key words
radionuclide recognition/convolutional neural network/Monte Carlo simulation引用本文复制引用
出版年
2025
国家科技期刊平台
NSTL
万方数据