Automated wavelet analysis of low resolution gamma-ray spectra and peak area uncertainty

The accuracy of automated isotope identification from low resolution gamma-ray spectra can be significantly improved with better algorithms. The method based on the wavelet transform and non-negative least squares (NNLS) are discussed in this thesis. Several improvements are made for the wavelet algorithm itself and different options can be configured in the MATLAB code. The partial or whole spectrum can be sent to NNLS and analyzed with or without subtracting the continuum. The boundary effects are also discussed. Several methods are developed to determine the area uncertainty. The matrix form of wavelet transform and error propagation are used. The inversion of the basis matrix is obtained either by the Moore-Penrose pseudo inversion or by truncated singular value decomposition (TSVD). The results are compared with those given by OriginLab and Gaussian fitting in MATLAB, which are consistent with each other, while TSVD is shown to be more accurate. The wavelet algorithm using TSVD for the area uncertainty calculation works well for complicated spectrum continuum and for overlapping peaks.

Name thesis:degree_name

M.S.

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Nuclear, Plasma, Radiolgc Engr

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2015

Author dc:creator

• Xiong, Hao

• gamma spectrum
• automated isotope identification
• wavelet analysis

Statement dc:rights

• Copyright 2015 Hao Xiong

Language dc:language

en

Handle dc:identifier

http://hdl.handle.net/2142/78576

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/78576