Adaptive Hit-Quality Index for Raman Spectrum Identification

The recognition capability of the identification system using Raman spectroscopy is increasing with the demands in the field. Among the various approaches that determine the identity of a target, signal correlation using a moving window is one of the most effective and intuitive methods. In this paper, we report a new correlation method that is robust to spectral intensity variations. Using the peak distribution of a given spectrum, this method adaptively determines meaningful spectral regions for the identification target. Three commercial Raman spectrometer and a 14 033 library were included in the study, which was used for a library-based chemical discrimination test and mixed material analysis experiments. According to the identification experimental results, the proposed method correctly identified all of the spectra and maintained a mean correlation score above 0.95 while maintaining the correlation score of nontarget materials as low as possible.