recognition;
Fourier-transform infrared spectroscopy;
green tea;
correlation coefficient;
average deviation;
D O I:
10.1109/ICMLC.2008.4620639
中图分类号:
TP3 [计算技术、计算机技术];
学科分类号:
0812 ;
摘要:
It is urgent to think up a quick and precise method for the identification of tea varieties and grades. The Fourier-transform infrared spectra of three grades of Xihu Longjing tea are measured, and in the spectra region of 400-4000cm(-1), which are build separately by the Savitzky-Golay smoothing and standard normal variate preprocessing method. It can be shown that the three different grades of Xihu Longjing have nearly the same characteristic peaks in conventional Fourier-transform infrared spectra. And the data analysis shows that the difference of correction coefficient in different grades is small. So only front the spectrum characteristic or correction coefficient the three grades of green tea can not be effectively recognized. The further analysis shows that the average derivation of the spectra can identify effectively and separate the different grades of tea quickly. Based on the analysis, a way of tea recognition which need not relate to the complicated component analysis of tea is given.