The task of substance identification is an actual issue in many areas of research related to quality control and life safety. Rapid identification and analysis of substances by sample-free methods can be used to solve the problems of drug distribution and prevention of terrorist acts. Raman spectroscopy is widely used for these purposes. Substances can be analyzed using mathematical methods for statistical comparison of experimental and standard spectra. For the identification of chemical compounds and the possibility of quantitative analysis of substances, generally, whole Raman spectra are used, that provides the greatest informativeness of the initial data. However, in several cases it is not possible to register the full Raman spectrum, and it is necessary to develop mathematical methods of analysis by the position of peaks of characteristic lines of the Raman spectrum. The method is based on the comparison of peak intensity values in the spectra of the investigated and reference substances. The performance of the proposed methods is investigated by means of correspondence matrices and ROC-analysis. The developed algorithms are compared with the Pearson correlation method. The study was carried out on the basis of a database consisting of 16 powdery substances using a Ventana-785L-Raman diffraction spectrometer equipped with a laser source with an excitation wavelength of 785 nm, maximum laser power up to 120 mW and power instability not more than 1%. Keywords: Raman spectroscopy, spectral analysis, identification, ROC curves.

Подсчитывается количество просмотров абстрактов ("html" на диаграммах) и полных версий статей ("pdf"). Просмотры с одинаковых IP-адресов засчитываются, если происходят с интервалом не менее 2-х часов.

Дата начала обработки статистических данных - 27 января 2016 г.