现代应用物理2025，Vol.16Issue(3)：19-26,8.DOI:10.12061/j.issn.2095-6223.202501002
人工神经网络在放射性核素能谱分析中的应用进展
Advances in the Application of Artificial Neural Networks to Analyze Radionuclide Energy Spectrum
摘要
Abstract
Artificial neural networks(ANNs)technology has unique advantages in processing nonlinear and multi-variable data sets.In response to the growing demand for analysis of radionuclide energy spectrum in relation to the Comprehensive Nuclear-Test-Ban Treaty(CTBT),this paper provides a detailed overview of the application of ANNs in both γ-ray and β-γ coincidence spectrum analysis.This study presents the training and performance of ANN in the nuclide identification,radioactive nuclide activity calculation,and optimization of experimental conditions,focuses on introducing the key elements of neural network training such as network structure,input parameters,and data sets.Compared with traditional radionuclide energy spectrum analysis method,the method based on ANNs makes full use of spectrum information,which not only improves the accuracy and efficiency of the analysis results,but also helps to improve the automation level of the system and reduce human intervention.The results demonstrate the promising potential of ANNs for resolving different CTBT-related tasks.关键词
CTBT/放射性/人工神经网络/γ能谱/β-γ能谱Key words
CTBT/radionuclide/artificial neural network/γ-ray spectrum/β-γ spectrum分类
信息技术与安全科学引用本文复制引用
郭启隆,牛亚洲,张瑞芹,赵允刚,张新军,秦雨浩,贾昊康,李奇,王世联..人工神经网络在放射性核素能谱分析中的应用进展[J].现代应用物理,2025,16(3):19-26,8.基金项目
国家重点研发计划资助项目(2022YFF0607300) （2022YFF0607300）
