In X-ray spectrum analysis, the pulse pile-up is a long-standing issue which deteriorates the energy resolution and count rates of the radiation detection systems. In this study, a novel pulse pile-up identification method based on particle swarm optimization and double-layer parameter identification model (PSO-DLPIM) is proposed. Different Gaussian pile-up waveforms are realized by exponential pulse through Sallen-Key (S-K) low-pass filtering. Then, the proposed model recognizes the parameters of each sub-Gaussian pulse. Especially, it can be used to modelling the pulse indirectly without a certain model parameter and overcomes the model mismatch troubles. Finally, computer simulations and experimental tests are carried out and the results show that this method has higher accuracy for the recognition of pile-up pulses. The example shows that the minimum distance between pulses that can be identified by this method is 0.05 μs. And when the pulse generation time is known and the environmental noise is low, the relative error of the amplitude of pulse pile-up recognition is as low as 0.15%. Therefore, this method can greatly improve the resolution of the X-ray spectrum.

在 X 射线光谱分析中，脉冲堆积是一个长期存在的问题，它会降低辐射检测系统的能量分辨率和计数率。在这项研究中，提出了一种基于粒子群优化和双层参数识别模型（PSO-DLPIM）的新型脉冲堆积识别方法。不同的高斯堆积波形通过Sallen-Key(SK)低通滤波由指数脉冲实现。然后，所提出的模型识别每个亚高斯脉冲的参数。特别是，它可以在没有一定模型参数的情况下对脉冲进行间接建模，克服了模型不匹配的问题。最后进行了计算机模拟和实验测试，结果表明该方法对堆积脉冲的识别具有较高的准确率。该示例表明，该方法可以识别的脉冲之间的最小距离为 0.05 μs。并且当脉冲产生时间已知且环境噪声较低时，脉冲堆积识别幅度的相对误差低至0.15%。因此，这种方法可以大大提高X射线光谱的分辨率。