Please wait a minute...
 首页  期刊介绍 期刊订阅 联系我们
 
最新录用  |  预出版  |  当期目录  |  过刊浏览  |  阅读排行  |  下载排行  |  引用排行  |  百年期刊
Journal of Tsinghua University(Science and Technology)    2023, Vol. 63 Issue (6) : 980-986     DOI: 10.16511/j.cnki.qhdxxb.2023.22.011
PUBLIC SAFETY |
A gamma radionuclide identification method based on convolutional neural networks
DU Xiaochuang1, LIANG Manchun1, LI Ke1, YU Yancheng1, LIU Xin2, WANG Xiangwei3, WANG Rudong1, ZHANG Guojie1, FU Qi1
1. Department of Engineering Physics, Tsinghua University, Beijing 100084, China;
2. Beijing Yongxin Medical Equipment Co., Ltd., Beijing 102206, China;
3. No. 91515 Unit of PLA, Sanya 572016, China
Download: PDF(4192 KB)  
Export: BibTeX | EndNote | Reference Manager | ProCite | RefWorks    
Abstract  [Objective] Rapid and reliable radionuclide identification can enable rapid monitoring and early warning of radioactive sources, which is essential for safeguarding people from the threat of radioactive materials. However, distinctive peak matching algorithms are not suitable for low gross count gamma-ray spectrum identification, especially when there are overlapped peaks in a spectrum. To improve the identification performance for low gross count gamma-ray spectra, this study creates a radionuclide identification model based on convolutional neural networks that can better identify the spectra obtained at low dose rates.[Methods] Firstly, a gamma-ray spectrum dataset was created. The gamma-ray spectra of 16 radionuclides were obtained at a dose rate of about 0.5μSv/h using a LaBr3 spectrometer with measuring energy ranging from 30 to 3000keV, a resolution of about 5% at 662 keV, and a measured acquisition time about 100s. Secondly, a training dataset was developed. To train the model, a huge number of gamma-ray spectra of 16 radionuclides and their two mixed radionuclides were generated. We created 1100 data points for each type of gamma-ray spectra by varying the gross count and energy drift. Thus, a total of 149 600 gamma-ray spectrum data were generated. Among them, 80% of the data were randomly selected for model training and the remaining 20% for model crossvalidation. Finally, the convolutional neural networks was constructed. The random searching approach was used to search hyperparameters of the model using the Keras-Tuner tool for determining the ideal architecture of convolutional neural networks. The convolutional layer filter numbers were 96, 128, 32, 256, and 256 in order. The activation function for convolutional layers was the rectified linear unit. Furthermore, the neuron number of the hidden layer was 480, and the learning rate was 0.000 029 6. At last, the spectra labels were encoded using the one-hot format, and the softmax function was used as the activation function for the model's output layer. The model parameters were optimized using the Adam optimizer by employing crossentropy as the loss function. We obtained the radionuclide identification model after 100 epochs of training.[Results] To estimate the identification performance of the model under the condition that a dose rate was about 0.5 μSv/h and the measurement acquisition time was up to 120 s, we acquired 1 333 gamma-ray spectra from nine single radionuclides and their two mixed radionuclides using the LaBr3 spectrometer. The nine radionuclides were 241Am, 133Ba, 137Cs, 131I, 226Ra, 232Th, 57Co, 235U, and 60Co. The model was used to identify these spectra and the results showed that the model's accuracy was 90.11% with the acquisition time of 30s, and the accuracy was increased to 92.63% with the acquisition time of 60s.[Conclusions] In this study, we propose a radionuclide identification model based on convolutional neural networks. Analyses show that the model can effectively identify various radionuclides' gamma-ray spectra in a short period of time at a low dose rate.
Keywords radionuclide identification      gamma-ray spectrum      convolutional neural networks      nuclear safety     
Issue Date: 12 May 2023
Service
E-mail this article
E-mail Alert
RSS
Articles by authors
DU Xiaochuang
LIANG Manchun
LI Ke
YU Yancheng
LIU Xin
WANG Xiangwei
WANG Rudong
ZHANG Guojie
FU Qi
Cite this article:   
DU Xiaochuang,LIANG Manchun,LI Ke, et al. A gamma radionuclide identification method based on convolutional neural networks[J]. Journal of Tsinghua University(Science and Technology), 2023, 63(6): 980-986.
URL:  
http://jst.tsinghuajournals.com/EN/10.16511/j.cnki.qhdxxb.2023.22.011     OR     http://jst.tsinghuajournals.com/EN/Y2023/V63/I6/980
  
  
  
  
  
  
  
  
  
  
[1] 周平坤.核辐射对人体的生物学危害及医学防护基本原则[J].首都医科大学学报, 2011, 32(2):171-176. ZHOU P K. The detriment of nuclear radiation on human and the protection principal[J]. Journal of Capital Medical University, 2011, 32(2):171-176.(in Chinese)
[2] LI K, CHEN W H, LIANG M C, et al. A simple data assimilation method to improve atmospheric dispersion based on Lagrangian puff model[J]. Nuclear Engineering and Technology, 2021, 53(7):2377-2386.
[3] LIKAR A, VIDMAR T. A peak-search method based on spectrum convolution[J]. Journal of Physics D:Applied Physics, 2003, 36(15):1903-1909.
[4] 任俊松,张江梅,王坤朋.基于SVD和SVM的核素识别算法[J].兵工自动化, 2017, 36(5):50-53. REN J S, ZHANG J M, WANG K P. Radioactive radionuclide identification method based on SVD and SVM[J]. Ordnance Industry Automation, 2017, 36(5):50-53.(in Chinese)
[5] CHEN L, WEI Y X. Nuclide identification algorithm based on K-L transform and neural networks[J]. Nuclear Instruments and Methods in Physics Research Section A:Accelerators, Spectrometers, Detectors and Associated Equipment, 2009, 598(2):450-453.
[6] 刘灏霖,姚远程,张江梅,等.基于K-SVD的γ能谱特征提取及核素识别[J].电脑知识与技术, 2018, 14(2):177-179. LIU H L, YAO Y C, ZHANG J M, et al. γ Spectrum feature extraction and radionuclide identification by K-SVD[J]. Computer Knowledge and Technology, 2018, 14(2):177-179.(in Chinese)
[7] HE J P, TANG X B, GONG P, et al. Rapid radionuclide identification algorithm based on the discrete cosine transform and BP neural network[J]. Annals of Nuclear Energy, 2018, 112:1-8.
[8] 王瑶,刘志明,万亚平,等.基于长短时记忆神经网络的能谱核素识别方法[J].强激光与粒子束, 2020, 32(10):106001. WANG Y, LIU Z M, WAN Y P, et al. Energy spectrum radionuclide recognition method based on long short-term memory neural network[J]. High Power Laser and Particle Beams, 2020, 32(10):106001.(in Chinese)
[9] 胡浩行,张江梅,王坤朋,等.卷积神经网络在复杂核素识别中的应用[J].传感器与微系统, 2019, 38(10):154-156, 160. HU H H, ZHANG J M, WANG K P, et al. Application of convolutional neural networks in identification of complex radionuclides[J]. Transducer and Microsystem Technologies, 2019, 38(10):154-156, 160.(in Chinese)
[10] LIANG D J, GONG P, TANG X B, et al. Rapid radionuclide identification algorithm based on convolutional neural network[J]. Annals of Nuclear Energy, 2019, 133:483-490.
[11] IAEA. Identification of radioactive sources and devices:Reference manual[M]. Vienna:IAEA, 2011.
[12] 盖荣丽,蔡建荣,王诗宇,等.卷积神经网络在图像识别中的应用研究综述[J].小型微型计算机系统, 2021, 42(9):1980-1984. GAI R L, CAI J R, WANG S Y, et al. Research review on image recognition based on deep learning[J]. Journal of Chinese Computer Systems, 2021, 42(9):1980-1984.(in Chinese)
[13] 杜晓闯,涂红兵,黎岢,等.基于径向基神经网络仿真γ能谱模板库的核素识别方法[J].清华大学学报(自然科学版), 2021, 61(11):1308-1315. DU X C, TU H B, LI K, et al. Radionuclide identification method based on a gamma-spectra template library simulated by radial basis function neural networks[J]. Journal of Tsinghua University (Science&Technology), 2021, 61(11):1308-1315.(in Chinese)
[1] DU Xiaochuang, TU Hongbing, LI Ke, ZHANG Jie, WANG Kang, LIU Hemin, LIANG Manchun, WANG Xiangwei. Radionuclide identification method based on a gamma-spectra template library simulated by radial basis function neural networks[J]. Journal of Tsinghua University(Science and Technology), 2021, 61(11): 1308-1315.
[2] HAN Kun, PAN Haiwei, ZHANG Wei, BIAN Xiaofei, CHEN Chunling, HE Shuning. Alzheimer's disease classification method based on multi-modal medical images[J]. Journal of Tsinghua University(Science and Technology), 2020, 60(8): 664-671,682.
[3] LIN Peng, WEI Pengcheng, FAN Qixiang, CHEN Wenqi. CNN model for mining safety hazard data from a construction site[J]. Journal of Tsinghua University(Science and Technology), 2019, 59(8): 628-634.
[4] LIU Qiong, LI Zongxian, SUN Fuchun, TIAN Yonghong, ZENG Wei. Image recognition and classification by deep belief-convolutional neural networks[J]. Journal of Tsinghua University(Science and Technology), 2018, 58(9): 781-787.
[5] LU Xiaofeng, ZHANG Shengfei, YI Shengwei. Free-text keystroke continuous authentication using CNN and RNN[J]. Journal of Tsinghua University(Science and Technology), 2018, 58(12): 1072-1078.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
Copyright © Journal of Tsinghua University(Science and Technology), All Rights Reserved.
Powered by Beijing Magtech Co. Ltd