A novel adaptive-scale Continuous Wavelet Transform method for weak peak search in low-resolution gamma spectrum

During measurements of Low- and Intermediate-Level Radioactive Waste (LILW), multiple factors can result in a significant number of weak peaks being partially or entirely obscured in the gamma spectrum. This phenomenon complicates accurate nuclide identification. To leverage the high detection efficiency of the LaBr3(Ce) detector for the on-site LILW measurements, research has been conducted to address the challenge posed by its low energy resolution, which hinders the precise search of weak peaks. A novel Continuous Wavelet Transform (CWT) weak peak search method has been proposed based on the analysis of CWT characteristics of gamma spectrum with different low signal-to-noise ratios (SNRs). The proposed method introduces a non-uniform scale parameter model for the mother wavelet, enabling adaptive adjustment based on energy resolution. Compared to conventional CWT, this approach significantly reduces energy error in peak search, with a maximum decrease of 97.9%. Furthermore, it improves multiplet peak resolution by 52.3%. Applying this method to peak search in gamma spectrum of nuclides (Cs-137, Co-60, Ba-133 and Eu-152) in LILW. The results show an absolute energy error of less than 1.5 keV, with an uncertainty of about 1 keV. Compared to Gamma Vision software, this method reduces absolute energy error by 95.15% and uncertainty by 91.95% in weak peak search. Compared with the traditional CWT, symmetric zero-area, covariance and Fourier transform methods, the proposed approach exhibits superior peak search performance on low-resolution complex spectrum, achieving true positive rate (TPR) improvements of up to 0.77 and false discovery rate (FDR) reductions of up to 0.85. This research provides a theoretical foundation and essential technical support for the application of LaBr₃(Ce) in the field of LILW measurement.