A peak-search method based on spectrum convolution
A peak-search method based on spectrum convolution
复制标题
DOI：
10.1088/0022-3727/36/15/323
复制
发表时间：
2003-08-07
影响因子：
3.4
通讯作者：
Vidmar, T
中科院分区：
文献类型：
--
作者：
Likar, A;Vidmar, T
本文提出了一种完全基于谱卷积的高分辨γ能谱寻峰方法。峰的检测和峰位置、宽度和面积的确定可以在连续的步骤中彼此独立地实现。为了实现每个步骤，频谱与适当的接近最优的函数进行卷积。可以导出位置、面积和宽度的不确定度的解析表达式。为了分辨双峰，使用标准的最小二乘技术，但在卷积光谱上，避免了与背景分量相关的问题。基于所提出的方法的软件包的性能通过国际原子能机构的两组标准测试谱成功测试，并与其他峰值搜索软件包进行了比较，证明所提出的方法是鲁棒性和可靠的。
A peak-search method for high resolution gamma-ray spectroscopy, based entirely on spectrum convolution, is presented. The detection of the peak and determination of the peak position, width and area can be realized in successive steps, independent of each other. To realize each of the steps, the spectrum is convoluted with an appropriate near-optimal function. Analytical expressions for the uncertainties of positions, areas and widths can be derived. For resolving doublets the standard least-square technique is used, but on the convoluted spectra, avoiding the problems associated with the background component. The performance of the software package based on the proposed method was successfully tested against two standard sets of test spectra of the International Atomic Energy Agency and intercomparison with other peak-search packages proves that the proposed approach is robust and reliable.