基于电子鼻的星座图聚类分析法在烟叶分类中的应用

Classifying Flue-cured Tobacco Leaves with Electronic Nose-based Constellation Diagram Clustering Analysis

摘要: 为探索区分不同产地或不同等级烤烟烟叶的有效方法,利用电子鼻检测了广东、福建、重庆和四川4个产地,C2F、B2F和X2F 3个等级以及不同品种的24个烟叶样品,并对电子鼻采集的数据进行星座图聚类分析。结果表明:电子鼻结合星座图聚类分析方法可以有效区分广东、福建、重庆和四川4个产地16个不同的烟叶样品,4个产地烟叶的综合值依次为:重庆 >广东 >四川 >福建;同时可以有效区分四川产地红花大金元品种3个等级的12个烟叶样品,3个等级烟叶的综合值依次为C2F >B2F >X2F。为人工烟叶分级鉴定提供了一种辅助方法。

Abstract: In order to effectively discriminate tobacco leaf samples of different grades or from different areas,24 leaf samples of C2F,B2F and X2F grades of different cultivars from Guangdong,Fujian,Chongqing and Sichuan were tested by electronic nose,and the test data were analyzed with constellation diagram clustering method.The results showed that the said method could effectively distinguish 16 tobacco leaf samples by their growing areas,the leaf samples in the order of their comprehensive values were Chongqing >Guangdong >Sichuan >Fujian.The 12 samples of cv.Honghuadajinyuan from Sichuan were effectively distinguished by their grades,the order of their comprehensive values were C2F >B2F >X2F.This method provides an assistant means to subjective grading of tobacco leaves.