Application of Higham and Savitzky-Golay Filters to Nuclear Spectra

Abstract

A nuclear spectrum is a Type II digital signal. It is processed to extract the qualitative and the quantitative information about the radioisotopes. Noise reduction is usually a step in processing a spectrum, and smoothing filters are employed for this. Perhaps the most widely used filters for the spectral filter-ing since the last 50 years have been Savitzky-Golay (S-G) filters. Incidentally S-G filters are not well known to the digital signal processing community. In this paper the authors have employed rather unknown high fidelity Higham filters used by actuaries of the 19th century for filtering real observed nuclear spectra, and compared their performance with that of an optimal S-G filter. It was observed that the performance of the Higham filters and S-G filters compared favorably amongst each other. This paper describes theory of S-G and Higham filters, their application to nuclear spectra, and presents the results. Higham filters have potential for more applications. There is a scope of exploring more treasure of the past, and merging tools from various disciplines, for more interdisciplinary research.