用于能谱本底处理的阶数自适应型正交多项式模型法  

Order Self-adaptive Orthogonal Polynomial Model Method for Spectrum Background Processing

作　　者：杨万里 刘冲[1] 胡锐 YANG Wanli;LIU Chong;HU Rui(School of Electrical Engineering,University of South China,Hengyang,Hunan 421001,China;CGN Jiuyuan(Chengdu)Science and Technology Co.,Ltd.,Chengdu,Sichuan 610059,China)

机构地区：[1]南华大学电气工程学院,湖南衡阳421001 [2]中广核久源(成都)科技有限公司,四川成都610059

出　　处：《南华大学学报（自然科学版）》2021年第3期65-69,74,共6页Journal of University of South China：Science and Technology

基　　金：湖南省自然科学基金资助项目(2018JJ2316);国家重点研发计划公共安全项目(24102-01)。

摘　　要：在能谱分析中,参数准确且个数尽可能少的本底处理方法一直是研究的重点。针对复杂本底获取过程中参数获取较为困难,本底获取精度低等问题,设计了一种用阶数自适应的正交多项式拟合法拟合本底,以拟合出的正交多项式作为本底函数,结合高斯函数描述特征峰,采用L-M算法,在确定高斯函数中相关参数的同时自动调整正交多项式的阶数,获取更准确的本底。该方法具有无需参数设置、可用于复杂本底处理、处理本底区间宽等优点。并与SNIP算法、傅里叶变换法、线性多项式模型法等进行对比,证明该方法不仅在理论模型上更加完善,而且整体结果也更准确,具有实用价值。In spectrum analysis,accurate and minimal background processing methods have always been the focus of research.This work proposes to use the order-adaptive orthogonal polynomial fitting method to fit the background,and use the fitted orthogonal polynomial as the background function,combined with the Gaussian function to describe the characteristic peak,and use the L-M algorithm to determine the correlation of the Gaussian function.At the same time,adjust the coefficients of the orthogonal polynomial to obtain a more accurate background.This method has the advantages of no parameter setting,being used for complex background processing,and wide background interval.And compared with SNIP(Scale Normalization for Image Pyramids)algorithm,Fourier transform method,linear polynomial model method,etc.,it is proved that not only this method is more perfect in the theoretical model,but also the overall result is more accurate and has practical value.

关 键 词：能谱分析 本底处理 正交多项式拟合 L-M算法 

分 类 号：TL82[核科学技术—核技术及应用]