Application PCA-ANN method to fast discrimination of tea varieties using visible/near infrared spectroscopy

A new method for the discrimination of varieties of tea by means of visible/near infrared spectroscopy (Vis/NIRS) (325-1 075 nm) was developed. A relation has been established between the reflectance spectra and the tea varieties. The data set consists of a total of 150 samples of tea. First, the data was analyzed with principal component analysis (PCA). It appeared to provide the reasonable clustering of the varieties of tea. Meanwhile PCA compressed hundreds of spectral data into a small quantity of principal components which described the body of the spectra; the first 6 principal components computed by PCA were applied as inputs to a back propagation neural network with one hidden layer. One hundred twenty five samples from five varieties were selected randomly, then they were used to build BP-ANN model. This model has been used to predict the varieties of 25 unknown samples; the residual error for the calibration samples is 1. 267 × 10-4. The recognition rate of 100% was achieved. This model is reliable and practicable. So this paper could offer a new approach to the fast discrimination of varieties of tea.