Research on n - γ Identification of Scintillation Neutron Detector Using PCA-CNN

The generation of neutrons always comes with accompanying gamma rays. Therefore, it is necessary to eliminate the influence of gamma rays in neutron measurement and discriminate n-γ rays. The Cs2LiLaBr6 (CLLB) crystal is capable of detecting both neutrons and gamma rays and has excellent scintillation performance. To enhance its discriminating capabilities, this study suggests combining CLLB crystals with a convolutional neural network (CNN) and principal component analysis (PCA). The CNN is used to extract features of neutron and gamma rays, and then the PCA method automatically selects the fewest principal components with at least 95% variance, enabling effective n-γ discriminating. Experimental results demonstrate that the PCA-CNN model proposed in this study outperforms the CNN alone, achieving a discriminating accuracy of 99.07%.