Discrimination of Neutron-Gamma Ray Pulses with Pileup Using Normalized Cross Correlation and Principal Component Analysis

During high event rates, discrimination of neutron and gamma-ray pulses can be challenging because of pulse pileup. In this paper, we develop a novel approach to deal with this problem. In our method, the normalized cross correlation (NCC) is used to characterize the behavior of typical neutron and gamma-ray pulses. The gradient of the NCC curve is shown to provide distinct features that can be used to distinguish neutron from gamma-ray pulses. Principal component analysis (PCA) is employed to extract features from the NCC gradient curve. We have employed the standard PCA approach and a modified PCA version to obtain unique features. The modified PCA method first extracts 20 Kolmogorov-Smirnov points and then computes the principal components of these 20 coefficients. We have exercised the technique on both simulated (for different pileup delays) and measurement data from a CLYC detector (for varying event rates). The modified PCA approach shows more promising results than the standard PCA approach with better figure of merit.