基于电子舌技术的西凤酒类别鉴定及掺假检测研究

Category Identification and Adulteration Detection of Fengxiang Liquor Based on Electronic Tongue Technology

研究以凤香型白酒的代表“西凤”为研究对象,利用电子舌对11种西凤酒的味觉进行采集,建立西凤酒传感器雷达图。以采集信号为依据,运用多种分析方法对西凤酒进行品牌、档次、年份的鉴定,并对西凤六年进行掺假检测的定性分析。结果表明:采用主成分分析(PCA)贡献率高达98.1186%,可以有效地将11种酒样进行区分;运用主成分分析结合临近算法(PCA-KNN)对西凤酒高价位段子品牌建立分类模型,模型准确率为1。在年份鉴定中,PCA分析方法优于线性判别分析(LDA)。采用PCA-SVM建立的西凤六年掺假检测模型准确率为96.67%。可以看出,电子舌可以用于西凤酒的品种鉴定和真假检验,且在西凤酒质量的检测和管理方面有很大的应用潜力。

In this study,the representatives of Fengxiang liquor,“Xifeng”were used as the research objects,the taste sensations of 11 kinds of Xifeng liquor were collected,and the radar map of the Xifeng liquor sensor was established by using electronic tongues.Based on the collection of signals,a variety of analysis methods were used to identify the brand,price segment and year of Xifeng liquor,and the qualitative analysis of adulteration detection was carried out in the sixth years of Xifeng.The results showed that the contribution rate of principal component analysis(PCA)was as high as 98.1186%,which could effectively distinguish 11 wine samples,the main component analysis combined with the K Nearest Neighbors(PCA-KNN)was used to establish a classification model for the high-priced duan sub-brands of Xifeng wine,and the model accuracy was 1.In vintage identification,the PCA analysis method was superior to linear discriminant analysis(LDA).The accuracy rates of the six-year adulteration detection models established by PCA-SVM were 96.67%.It could be seen that the electronic tongue could be used for the variety identification and authenticity inspection of Xifeng liquor,and had great application potential in detection and management of the quality of Xifeng liquor.