An adaptive rational filter for interpretation of spectrometric data

The computer-based interpretation of spectrometric data is of great importance for applications of various kinds of spectroscopy in science, biomedical engineering, environmental engineering, and industry. It is based on the use of algorithms of (generalized) deconvolution for correcting spectrometric data, i.e., compensating for the instrumental effects and natural bandwidth effects distorting those data. Among many sophisticated methods currently used for this purpose, no one method is able to properly deal with the irregularities in the data such as the variable shape of peaks of which the interpreted spectrum is composed. In this paper, a flexible and efficient algorithm of generalized deconvolution is proposed, viz. an adaptive rational filter. It is proven to satisfactorily solve this problem by recursive adaptation of its parameters to the widths of the consecutive peaks of which the interpreted spectrum is composed, as well as to the relative distances between peaks along the wavelength axis.