摘要
利用Geant4模拟软件及MATLAB计算软件,开展多人工神经网络模型定量分析低计数率核素的研究,分析了包含五个不同类型网络的BP神经网络模型对国控环境监测站NaI(Tl)探测器模拟能谱的识别能力。实验结果表明:多网络模型的使用有利于提高低计数核素的识别准确度,能够有效识别未知核素的存在与否。
Geant4 simulation software and MATLAB calculation software were used to carry out the research of quantitative analysis of low count rate nuclides by multi-artificial neural network model.The recognition ability of BP neural network model including five different types of networks for NaI(Tl)detector simulation energy spectrum of state controlled environmental monitoring station was analyzed.The experimental results showed that the use of multi-network model is beneficial to improve the recognition accuracy of low count nuclides,and can effectively identify the presence or absence of unknown nuclides.
作者
朱剑文
帅磊
钮云龙
曹大泉
赵桂芝
梁秀佐
张译文
杨维耿
ZHU Jianwen;SHUAI Lei;NIU Yunlong;CAO Daquan;ZHAO Guizhi;LIANG Xiuzuo;ZHANG Yiwen;YANG Weigeng(School of Nuclear Science and Technology,University of South China,Hengyang,Hunan 421000,China;Beijing Engineering Research Center of Radiography Technology and Equipment,Institute of High Energy Physics Chinese Academy of Sciences,Beijing 100049,China;School of Nuclear Science and Technology,University of Chinese Academy of Sciences,Beijing 100049,China;Radiation Environment Monitoring Technology Center,Ministry of Ecology and Environment,Zhejiang Provincial Department of Ecological Environment,Hangzhou,Zhejiang 310012,China)
出处
《南华大学学报（自然科学版）》
2022年第4期75-81,100,共8页
Journal of University of South China：Science and Technology
基金
中国科学院重点部署项目(ZDRW-CN-2018-101)
中国科学院科研仪器设备研制项目(29201707)
中国科学院青年创新促进会基金项目(292017QCH01)
