Radionuclide Identification Method of Two-Dimensional Convolutional Neural Network Based on Gamma Pulse Peak Sequence

[Background]: Rapid radionuclide identification is a crucial step in preventing the loss, smuggling, terrorist attacks, and radioactive contamination involving hazardous materials. Most current identification methods rely on gamma spectra as the primary analytical tool. However, due to limitations in spectral statistics, these methods require extended processing times to achieve results, making them slow, less accurate, and poorly generalized for low-count-rate applications. Emerging radionuclide identification methods now utilize nuclear pulse peak sequence for analysis. However, these methods often fail to fully capture the features of nuclear pulse peak sequence, which limits the identification accuracy. [Purpose]: To overcome the limitations of spectral statistics and enhance the speed and performance of radionuclide identification. [Methods]: This paper introduced a radionuclide identification method that employs a Two-Dimensional Convolutional Neural Network (2D-CNN) utilizing nuclear pulse peak sequences. First, we used four sources—137Cs, 60Co, 155Eu, and 22Na—as research subjects and collected low-count-rate nuclear pulse peak sequences for single sources, mixed sources, and environmental backgrounds at varying source distances using a NaI(Tl) detector in the lab. Then, we preprocessed the collected sequences through fixed-length segmentation, min-max normalization, and two-dimensional matrix mapping to generate multiple nuclear pulse peak sequence datasets with different matrix sizes. Next, we developed a 2D-CNN model and optimized the convolution kernel size and padding method using 10-fold hierarchical cross-validation to enhance feature extraction from nuclear pulse peak matrices. Finally, we tested the model’s radionuclide identification capability on datasets with four simple and easily distinguishable sequences, five-category sequences, and eight complex-category sequences, and compared its performance against three other methods: BPNN+PCA (Back Propagation Neural Network + Principal Component Analysis), SVM+PCA (Support Vector Machine + Principal Component Analysis), and 2D-CNN+spectrum. [Results]: The 2D-CNN radionuclide identification results show that with only 300 nuclear pulse sequence points, it achieves an accuracy of 99.61% on four easily distinguishable sequence sets. For five category sequence sets, an accuracy of over 95% is achieved with just 400 pulse sequence points. Moreover, for eight complex category sequence sets, a stable recognition accuracy is attained with 400 pulse sequence points. Additionally, comparative experiments with different models indicate that the 2D-CNN achieves accuracies of 100%, 98.61%, and 84.45% for classifying four, five, and eight category sequence sets, respectively. This performance significantly surpasses that of the BPNN+PCA, SVM+PCA, and 2D-CNN+gamma spectrum methods, and it also outperforms these models in single-source generalization. [Conclusions]: The 2D-CNN model demonstrates feasibility in automatically extracting features from fixed-length nuclear pulse peak sequences. It effectively extracts pulse sequence features within a 40 cm detection range and performs rapid radionuclide identification. This method exhibits advantages in both accuracy and generalization, making it suitable for rapid radionuclide identification tasks with low count rates.1. The dataset contains 137Cs (2.94 × 105Bq), 60Co (9.9 × 104Bq), 155Eu-22Na (1.12 × 104Bq), 137Cs-60Co (3.93 × 105Bq), 60Co-155Eu-22Na (3.052 × 105Bq), 137Cs-55Eu- 22Na (1.102×105Bq), 137Cs-60Co-155Eu-22Na (4.042×105Bq), environmental background (0.01uSv/h) 8 kinds of radioactive sources nuclear pulse sequence point data and fixed-length nuclear pulse sequence set sample data.2. Acquisition conditions: The detection distances were from 5cm to 40cm with an interval of 5cm, a total of 8 detection distances. The collected nuclear pulse signals were extracted by DSPEC-50 digital multichannel spectrometer to get the γ pulse amplitude and the corresponding pulse arrival time, i.e., the γ pulse peak sequence data. The detector collected at least 2,000,000 sequence points for each combination of radioactive sources at each detection distance, for a total of at least 2,000,000 × 7 × 8 sequence points, and in addition, 7,412,376 sequence points of nuc7 were collected separately. Each sequence point contains nuclear pulse amplitude points and real-time arrival times.