信阳毛尖茶品质和化学成分快速检测方法

探索茶叶品质和化学成分的快速检测方法,利用电子鼻技术对3个品质等级的信阳毛尖茶的挥发性气味进行了分析.对电子鼻检测信号进行主成分分析和线形判别分析结果显示3个品质的茶叶能被很好地区分,各个类的集中性也较强.采用二次多项式逐步回归分析分别建立传感器信号和信阳毛尖茶的氨基酸、茶多酚和咖啡碱含量之间的预测模型.通过测试集对二次多项式逐步回归模型进行验证得到氨基酸、茶多酚和咖啡碱含量的预测值和测试值的相关系数分别为0.95、0.94和0.92,预测标准误差分别为0.08、0.11和0.8,平均误差分别为0.9%、2.8%和1.5%.结果表明,电子鼻技术可以用于茶叶理化成分的快速检测.

Method of Rapid Identification of Quality and Chemical Composition of Xinyangmaojian Tea

An electric nose was used to evaluate the quality and chemical composition of Xinyangmaojian tea at three different quality levels. The muhivariable analyses including principal component analysis and linear discrimination analysis were applied to distinguish the tea samples. A correct classification was achieved for the tea sample of three different quality levels and each group has strong convergence. The relationship between sensors signals and content of amino acid, tea polyphenols and caffeine for Xinyangmaojian tea were developed using the quadratic polynomial step regression analysis. The results showed that the quadratic polynomial step regression model represented good ability in predicting of chemical composition, with high correlation coefficients （R=0.95 for amino acid： R=0.94 for tea polyphenols; R=0.92 for caffeine, resrectively） between predicted and measured values, with relatively low standard error prediction （SEP） （0.08%, 0.11%, 0.8% for amino acid, tea polyphenols and caffeine, respectively） and a relatively low average percent error（ERR）（0.9%, 2.8%, 1.5% for amino acid, tea polyphenols and caffeine, respectively）. These results prove that electronic noses has the potential of becoming a rapid instrument to assess the tea chemical compositions.