Use of NIR Spectroscopy Combined with Recognition Methods for the Identification of Black Teas from Different Regions
2012 ◽
Vol 503-504
◽
pp. 1601-1604
◽
Keyword(s):
Near-infrared (NIR) spectroscopy, combined with pattern recognition, was applied in this study for the rapid identification of Black tea from different origins.The K-Nearest Neighbor model recognition method was used for the establishment of a tea origin recognition model, which involved optimization of the principal component factors (PCs) and the identification rate using a cross-validation method. The experimental results showed that, after standard normal variant spectral preprocessing, an optimized model was obtained when the PCs were equal to three, with the cross-validation recognition rate and the predicted recognition rate reaching 98.1% and 93.3%, respectively.