Research On Identification Of Similar And Complex Substances Based On Near-infrared Two-Dimensional Correlation Spectroscopy

Spectroscopy is an effective carrier of material information and can reveal the structural information of matter at the molecular level.The chemometrics method is a discipline that combines spectral analysis with mathematical methods.Combining chemometrics and pattern recognition analysis methods can generate new methods that can be used for qualitative analysis and quantitative analysis of substances.These spectral analysis-based methods have been widely used in daily life and various industrial fields.The traditional one-dimensional spectrum is a spectrum with the wavelength or wavenumber as the abscissa and the absorbance as the ordinate.By analyzing the spectrum,the structural information of the substance can be obtained to determine the chemical information of the substance.Combining one-dimensional spectroscopy with pattern recognition methods enables the identification of species or the determination of the content of a substance.However,for some substances with high structural similarities,the information obtained from one-dimensional spectra is limited,and the difference between the two cannot be clearly expressed.The two-dimensional correlation spectrum evolved on the basis of one-dimensional spectrum.The two-dimensional correlation spectrum is an information enhancement technology that can reflect the higher-dimensional information contained in matter.Two-dimensional correlation spectroscopy continuously interferes with the material system with some reasonable external disturbance.During this process,the spectral data of multiple acquisitions of materials is used to construct the dynamic spectrum of the material.After analyzing the principle of two-dimensional correlation spectroscopy,this paper creatively introduces it into the identification of complex materials,extracts many different features from it,and designs multiple classifiers correspondingly.Finally,an improved information fusion method based on information entropy is used to achieve the unity of multiple classifiers,and the effectiveness of the method is verified through experiments.At present,there are many substances with a high degree of similarity in their structure and difficult to distinguish by one-dimensional spectrum,such as wool products and cashmere products,pure cotton and mercerized cotton products.Because of the large difference in the prices of such commodities and the difficulty in distinguishing them with the naked eye,the shoddy market often appears to deceive consumers and textile practitioners to profiteering.This thesis proposes a complex material identification method based on two-dimensional correlation spectroscopy,which is a non-destructive and rapid method for identification.The author's experiments results proved the effectiveness and practicality of the method.