用于复合晶体探测器的数字波形甄别方法研究

Research on digital pulse shape discrimination methods for phoswich detector

本文设计了一个基于DSP的数字波形甄别系统,在此基础上研究了BP神经网络法、最佳线性法、等效宽度法、脉冲前后沿宽度法等数字甄别方法,给出了NaI(Tl)和CsI(Na)复合晶体探测器的甄别结果,并从甄别能力、运算速度、抗干扰能力等方面进行了比较和讨论,结果表明BP神经网络甄别方法具有最大的甄别能力和抗干扰能力,同时能够满足系统实时甄别的要求。

A liSP based digital pulse shape discrimination system for NaI（TI）/CsI（Na） plaoswich detector was designed, based on that several digital pulse shape analysis methods such as BP neural network, optimum linear discriminator, equivalent width and rise-fall pulse width are studied , and analysis results are given and compared from several aspects such as discrimination capability, execute time and noise sensitivity. It has been shown that among all these methods BP neural network has the best discrimination capability and can satisfy the real-time requirement.