伏安型电子舌对不同种类绿茶的辨识

Idenfication of Green Tea Using Voltammetric Electronic Tongue

应用伏安型电子舌对碧螺春、观音王、毛峰和雀舌4种绿茶进行辨识,采用金电极和铂金电极对4种绿茶的64个样品进行测试。激励信号采用循环伏安法,电压范围设定为-0.2～0.8 V,循环2个周期,获取每个样品的电化学指纹图谱。对获取电流信号进行不同方式的预处理,并对预处理后的数据采用类间类内指数比和主成分分析(PCA)进行评估,采用偏最小二乘判别(PLS-DA)对校正集和预测集进行建模和比较。结果表明,在同一循环伏安周期相同电势下的响应电流差值归一化预处理后,金电极响应数据的类间类内指数比最大,此时PLS-DA建模效果最好,校正集和预测集分辨率分别达100%和93.75%。

About 4 kinds green tea used, such as Biluochun, Guanyinwang, Maofeng and Queshe, were distinguished by voltammetric electronic tongue. The detection was investigated using gold and platinum as working electrode and cyclic voltammery measuring 64 samples of 4 kinds green tea, 2 periods voltammogram of each sample was recorded by cyclic voltammery in potential range from-0.2-0.8 V, and the resultant current from the potentiostat was considered for data analysis. The different data preprocessing techniques were compared by criterion of class separability and principal component analysis（PCA）. The classification performances were evaluated using partial least square discriminant analysis（PLS-DA） model for identification of different green tea. The results showed that the ratio of class separability from response data of gold electrode is highest using normalized pretreatment of the difference of response current at the same cyclic voltammetric period and the same potential, and PLS-DA model achieved 100% and 93.75% classification rate in train set and test set, respectively.