Integrated Similarity Indices for Component Recognition in Mixtures with Three-Dimensional Fluorescence Spectra

Shaohui Yu,Xue Xiao,Zhigang Wang,Kai Zhang,Gaofang Yin,Yujun Zhang,Wenqing Liu
DOI: https://doi.org/10.1366/10-06160
2011-01-01
Applied Spectroscopy
Abstract:Component recognition is a very important issue in the analysis of mixed three-dimensional fluorescence spectra and it can be realized by calculating the similarity index between the reference spectra and the computed spectra from the trilinear decomposition of the three-dimensional data arrays. However, the most popular similarity index available in the literature for processing three-dimensional fluorescence spectra takes advantage of only part of the information from the trilinear decomposition. It works well when there are clear differences between the component spectra, but it may fail when the spectra are severely overlapped. In order to overcome the shortcomings and to adapt to the rapid development of online monitoring, we propose a type of integrated similarity index (ISI) that is particularly superior for component recognition in mixtures with severely overlapped spectra. The ISI makes full use of as much information of the three-dimensional fluorescence spectra as possible, namely, of both the waveform and the characteristic peak wavelength, as well as both the emission spectra and the excitation spectra. With the ISI, the recognition process can be accomplished automatically and more accurately in extreme cases than the traditionally defined similarity indices that are based on only one specific feature. The feasibility of the ISI is demonstrated by experiments with mixtures of phenol, cresol, and thymol.
What problem does this paper attempt to address?