Identification of Tea Based on CARS-SWR Variable Optimization of Visible / Near Infrared Spectrum.

BACKGROUND The identification of tea varieties is the basis of realizing high quality and high price of tea. In order to quickly and non-destructive identify tea varieties and fight against counterfeit and inferior products in the tea market, a new method of visible/near infrared spectrum processing based on competitive adaptive reweighting algorithms-stepwise regression analysis (CARS-SWR) variable optimization is proposed to identify tea varieties in this paper. RESULTS The spectral data of five different tea varieties were obtained by visible / near infrared spectrometer, and the spectral data were preprocessed by the multivariate scattering correction (MSC) algorithm. Firstly, 20 wavelength variables were selected by CARS, and then 6 optimal wavelength variables were selected by using SWR method based on the optimal variables of CARS. Finally, the generalized regression neural network (GRNN) classification model and probabilistic neural network (PNN) classification model were established respectively based on the spectral information of full wavelength, the spectral information of CARS preferred wavelength variable, the spectral information of SWR preferred wavelength variable and the spectral information of CARS-SWR preferred wavelength variable. CONCLUSION It was found that the CARS-SWR-PNN model has the best classification effect by comparing different modeling results, and the classification accuracy of its training set and test set reached 100%. This shows that the method of CARS-SWR preferred variable combined with visible/near infrared spectrum is feasible for rapid and non-destructive identification of tea varieties. This article is protected by copyright. All rights reserved.