伏安电子舌对不同植物源蜂蜜的检测识别

Detection and identification of different floral honey by voltammetric electronic tongue

构建伏安电子舌系统为标准的三电极体系,并选用差分脉冲伏安法来检测获取蜂蜜样品伏安特性曲线,实现了对6种同种品牌、不同植物源的蜂蜜的识别。通过提取实验曲线的波峰波谷作为特征值,利用主成分分析法(PCA)对数据进行分析,并采用模糊K近邻分类器模式识别算法对数据进行处理。最后将识别结果与模糊自适应谐振理论和RBF神经网络相比较。结果表明伏安电子舌能够有效区分不同植物源蜂蜜,并且模糊K近邻的识别效果要优于模糊自适应谐振理论和RBF神经网络。

The electronic tongue was built to be the standard three electrodes system,and the differential pulse voltammetry was used to test for the honey sample voltammetric curve,which realized the identification of 6 different brands and floral sources of honey samples.The picks and valleys were analyzed by principal component analysis（PCA） and processed by fuzzy k-nearest neighbor（FKNN） algorithm.The results of FKNN were compared with fuzzy adaptive resonance theory（Fuzzy ARTMAP） and radial basis function（RBF）.Results show that voltammetric electronic tongue can distinguish six kinds of honey effectively and the recognition effect of Fuzzy KNN is better than that of Fuzzy ARTMAP and RBF neural network.