利用脉冲快热中子分析方法和神经网络识别爆炸物

Study of explosive detection with PFTNA method and neural network

研究利用PFTNA方法和人工神经网络来识别爆炸物的方法。该方法利用PFTNA来获取被检测样品的热中子俘获γ谱和快中子非弹性散射γ谱,然后通过已知能谱对人工神经网络进行训练后识别未知样品。利用这种方法,能够快速的对未知样品进行定性分析。

<Abstrcat> An explosive detection method was studied with pulsed fast/thermal neutron analysis (PFTNA) method and artificial neural network in this paper. PFTNA was used to get object's γ-ray spectrum of thermal capture reaction and spectrum of inelastic scattering reaction. With these two spectra, artificial neural network can be trained to recognize explosive from other objects. Some explosives and non-explosives were sampled to train and test neural network's recognition capability in this study.