High-resolution gamma spectroscopy shift-invariant wavelet de-noising

TL;DR

This paper introduces a shift-invariant wavelet de-noising algorithm for high-resolution gamma spectra that effectively reduces noise while preserving peak shapes, overcoming pseudo-Gibbs artifacts common in traditional methods.

Contribution

A novel shift-invariant wavelet de-noising method is proposed for gamma spectra, improving peak preservation and reducing artifacts compared to traditional wavelet techniques.

Findings

Reduces pseudo-Gibbs artifacts in gamma spectrum de-noising

Maintains characteristic peak shapes effectively

Enhances peak detection and localization quality

Abstract

For the multi-resolution of wavelet transform, it used to filter the complex gamma Spectrum, the fluctuation is filtered, while the detector resolution is still kept well, which has been demonstrated as a new, promising technique for gamma spectrum de-noising. However, both side of the peak where the data rapidly changing area, the reconstruction spectrum will be artificial fluctuation of pseudo-Gibbs, when de-noising high-resolution gamma spectrum. To solve these problems, a novel shift-invariant wavelet de-noising algorithm is proposed to treat the gamma spectrum which measured by HPGe detector of the segment gamma scanning system. It has a high resolution, a short measuring time, severe statistical fluctuation and scattering characteristics. The original spectrum was cycle spinning, de-noising by soft threshold, reconstructed. And then it was reversed cycle spinning, while the result was averaged. The algorithm not only overcomes the pseudo-Gibbs in the high-resolution gamma spectrum de-noising by the traditional wavelet, but also keeps the shape of the characteristic peak well. Emphasis will be placed on quantifying the merits of this algorithm to traditional wavelet by exploring the quality of peak detection and localization.