Discrimination of varieties of tea using near infrared spectroscopy

This work is aim to present a new approach for discrimination of varieties of tea by means of infrared spectroscopy (NIRS) (325-1075nm). The relationship has been established between the reflectance spectra and tea varieties. The data set consists of a total of 150 samples of tea. First, the spectra data was analyzed with principal component analysis. It appeared to provide the reasonable clustering of the varieties of tea. PCA compressed thousands of spectral data into a small quantity of principal components and described the body of spectra the scores of the first 6 principal components computed by PCA had been applied as inputs to a back propagation neural network with one hidden layer. 125 samples of five varieties were selected randomly, which were used to build BP-ANN model. This model had been used to predict the varieties of 25 unknown samples; the residual error for the calibration samples is 1.267 x 10-4. The recognition rate of 100% was achieved. This model is reliable and practicable. So this paper put forward a new method to the fast discrimination of varieties of tea.