闪烁探测器信号波形识别的人工神经网络实现

Pulse discrimination of scintilator detector with artificial neural network

根据闪烁探测器中激发荧光的快、慢成分的差异,定义了代表信号波形的三个特征物理量:即信号幅度、信号在上升时间段的积分以及中频段的频谱积分。研究了人工神经网络对不同波形的识别方法。用计算机模拟了实际情况下的信号波形,用神经网络对这些信号波形进行识别,得到了较好的识别结果并进行了讨论。

The features of signal for scintillator detectors are analyzed. According to the difference in the fraction of slow and fast scintillation for different particles, three intrinsic parameters （signal amplitude, integration of signal during rinsing, integration of frequency spectrum of signals in middle frequencies） of signals are defined. The artificial neural network method for pulse discrimination of scintillator detector is studied. The signals with different shapes under real condition are simulated with computer, and discriminated by the method. Results of discrimination are gotten and discussed.