基于风味指纹谱的庐山云雾茶品质等级研究  

Study on Quality Grading of Lushan Cloud-fog Tea Based on Flavor Fingerprints

作　　者：祁兴普[1] 刘纯友[2] 佀再勇 刘萍[1] 傅晓雨 战旭梅[1] 陈通 QI Xing-pu;LIU Chun-you;SI Zai-yong;LIU Ping;FU Xiao-yu;ZHAN Xu-mei;CHEN Tong(School of Food Science and Technology,Jiangsu Agri-animal Husbandry Vocational College,Taizhou 225300,Jiangsu,China;School of Biological and Chemical Engineering,Guangxi University of Science and Technology,Liuzhou 545006,Guangxi,China)

机构地区：[1]江苏农牧科技职业学院食品科技学院,江苏泰州225300 [2]广西科技大学生物与化学工程学院,广西柳州545006

出　　处：《食品研究与开发》2021年第14期152-157,共6页Food Research and Development

基　　金：泰州311高层次人才计划;广西科技大学博士基金项目(校科博20Z34)。

摘　　要：为建立一种快速、无损的庐山云雾茶等级判别方法,采用气相离子迁移谱(gas chromatography-ion mobility spectrometry,GC-IMS)联用设备对3个等级共63个庐山云雾茶样的挥发性有机成分进行分析检测,并采用Otsu自动阈值分割算法对GC-IMS二维谱图中特征峰进行特征提取,以特征峰的峰面积为变量进行主成分分析,再结合K-最邻近(K-nearest neighbor,KNN)算法对主成分得分进行模式识别。结果表明,采用KNN方法能够很好地区分不同等级的庐山云雾茶,预测集样品识别率可达94.73%。To identify a method for discriminating among three different grades of Lushan Cloud-fog tea,a rapid and nondestructive method for analysis of volatile organic compounds was established using gas chromatography combined with ion mobility spectrometry(GC-IMS). In this experiment,a total of 63 samples representing three different grades were evaluated by GC-IMS,and the Otsu algorithm was used to extract characteristic peaks from two-dimensional data profiles. Areas under the selected peaks were used as characteristic variables for principal component analysis,and a K-nearest neighbor(KNN)algorithm was used for pattern recognition.The results showed that the KNN pattern recognition method could effectively distinguish between different grades of Lushan Cloud-fog tea samples,achieving a discrimination rate of 94.73% in the prediction set.

关 键 词：庐山云雾茶 气相离子迁移谱 挥发性有机成分 等级判别 风味指纹 

分 类 号：TS272.7[农业科学—茶叶生产加工] O657[轻工技术与工程—农产品加工及贮藏工程]