Résumé
In order to discriminate the crude and sweated Dipsaci Radix correctly and rapidly, the crude and sweated Dipsaci Radix were scanned by the NIR spectrometer, and an identifying model was developed by near infrared spectroscopy combined with principal component-Mahalanobis distance pattern recognition method. The pretreated spectra data of 129 crude samples and 86 sweated ones were analyzed through principal component analysis (PCA). The identifying model was developed by choosing the spectrum for 9 881.46-4 119.20 cm(-1) and "SNV + spectrum + S-G" to the original spectral preprocessing with 14 principal components, and then was verified by prediction set, identifying with 100% accuracy. The rapid identification model of the crude and sweated Dipsaci Radix by NIR is feasible and efficient, and could be used as an assistant means for identifying the crude and sweated Dipsaci Radix.