FT-NIR法测定青烟叶品质参数的研究

利用傅立叶变换近红外光谱仪漫反射技术对青烟叶的总糖、还原糖、尼古丁、总氮、总氮等5种主要品质参数进行分析研究，并建立关系模型，经内部交叉证实验证模型并预测样品，结果为：总糖、还原糖、尼古丁、总氮、总氯等各组分的决定系数R^2分别为97.93%,97.01%,97.43%,98.96%,95.06%；近红外预测值与实测值之间的平均相对误差分别为4.78%,5.64%,4.15%,3.86%,18.60%.该技术具有检测快速、制样简便、无污染等优点，可初步用于青烟品质参数的快速定量分析，并且测试对象可以不断扩展。

DETERMINATION OF QUALITY FACTORS OF GREEN TOBACCO LEAVES WITH FOURIER TRANSFORMATION NEAR- INFRARED SPECTROSCOPY

Fourier transformation near-infrared spectroscopy was performed to determine total sugars, reducing sugar, nicotine, total nitrogen and total chlorine in green tobacco leaves, and correlation models were established. The models were checked up with inner chiasma confirmation and predictions were made of the samples with them. Their determination coefficient (R 2) was estimated to be 97.93%, 97.91%, 97.43%, 98.96% and 95.96%, and their average relative error between NIR-predicted data and the actually measured results was 4.78%, 5.64%, 4.15% , 3.86% and 18.60%, respectively. It is thus tentatively concluded that NIR spectroscopic analysis can be applied for rapid quantitative analysis of the quality factors of green tobacco leaves.