A nuclide identification method of γ spectrum and model building based on the transformer

分类： 物理学 >> 核物理学 提交时间： 2024-07-30

Fei Li Chu-Yang Luo Ying-Zi Wen Sheng Lv Feng Cheng Guo-Qiang Zeng Jian-Feng Jiang Bing-Hai Li

摘要： In current neural network algorithms for nuclide identification in high-background, poor-resolution detectors,traditional network paradigms including back-propagation networks, convolutional neural networks, recurrentneural networks, etc. have been limited in research on γ spectrum analysis because of their inherent mathemat#2;ical mechanisms. It is difficult to make progress in terms of training data requirements and prediction accuracy.In contrast to traditional network paradigms, network models based on the transformer structure have the charac#2;teristics of parallel computing, position encoding, and deep stacking, which have enabled good performance innatural language processing tasks in recent years. Therefore, in this paper, a transformer-based neural network(TBNN) model is proposed to achieve nuclide identification for the first time. First, the Geant4 program wasused to generate the basic single-nuclide energy spectrum through Monte Carlo simulations. A multi-nuclideenergy spectrum database was established for neural network training using random matrices of γ-ray energy,activity, and noise. Based on the encoder-decoder structure, a network topology based on the transformer wasbuilt, transforming the 1024-channel energy spectrum data into a 32 × 32 energy spectrum sequence as themodel input. Through experiments and adjustments of model parameters, including the learning rate of theTBNN model, number of attention heads, and number of network stacking layers, the overall recognition ratereached 98.7%. Additionally, this database was used for training AI models such as back-propagation networks,convolutional neural networks, residual networks, and long short-term memory neural networks, with overallrecognition rates of 92.8%, 95.3%, 96.3%, and 96.6%, respectively. This indicates that the TBNN model exhib#2;ited better nuclide identification among these AI models, providing an important reference and theoretical basisfor the practical application of transformers in the qualitative and quantitative analysis of the γ spectrum.

通过

 点击量 36  下载量 4  评论 0