基于序贯Bayes分析的中低能量分辨率γ能谱解析技术研究

Low and medium energy resolution γ spectrum analysis technique is widely used in the field of nuclear security, environmental monitoring and effluent monitoring. In these application areas, we need automatic and on-line γ spectrum analysis. Known from Laboratory γ spectrum measurement, in these application areas we faced with many difficult problems that uncertain measurement condition, limited measurement time, and on-line characteristic of measurement data etc. The result of recent investigation for these difficulties indicates that the essential problem is the difficulties of the clear statistical statement about the data processing. In recent years the analysis techniques for high and medium level energy resolution γ spectrum based on the sequential Bayesian approach is realized successfully. But this method cannot directly applicable to low level or highly overlapping medium level energy resolution γ spectrum analysis. Corresponding to the above difficulties, we will research on two direction. First, research on the uncertainty estimation methodology for conventional analysis techniques. Second, research on Low and medium energy resolution γ spectrum analysis techniques based on sequential Bayesian analysis will perform in this project; and finally realize sequentialization of algorithm and clear statistical statement about the data processing.

中低能量分辨率γ能谱解析技术广泛应用于核安保、环境监测、流出物监测等领域。这些应用领域，都需要自动化程度较高的在线式γ能谱解析技术。区别于实验室当中的γ能谱测量，这些领域内现场实时监测面临着测量条件不确定、测量时间受限、测量数据的实时性等诸多特殊问题。面对这些困难，最近几年所研究结果表明，最根本的困难就是很多情况下对于数据处理结果无法给出明确的统计陈述。近年来基于序贯Bayes分析的中高能量分辨率γ能谱解析技术已得以实现，但其应用于低能量分辨率或重峰较严重的中能量分辨率γ能谱解析方面还有一些问题。本项目将围绕这些问题，具体从以下两个方面出发来研究这些问题。一，研究传统解谱算法中各种算法的不确定度评估技术，以便每一步处理结果都可得到明确的统计陈述。二，开展基于序贯Bayes分析的中低能量分辨率γ能谱解析技术研究，进而实现能谱数据处理的序贯化，并能给出明确的统计陈述。

本项目针对中低能量分辨率探测器γ能谱技术处理研究中，针对于NaI探测器实现了序贯化的反卷和核素识别。其中序贯化的反卷算法即使在测量条件不确定的情况下，也能有效增强能量分辨率；序贯化的核素识别算法可在测量条件确定的情况下进行实时地识别核素，因此在环境监测和流出物监测等领域具有较大的潜在应用价值。