食品科学

• 成分分析 • 上一篇    下一篇

基于龙井茶香气风味特性的品质判定

戴悦雯1,支瑞聪2,*,赵 镭2,高海燕1,史波林2,汪厚银2   

  1. 1.上海大学生命科学学院,上海 200444;2.中国标准化研究院食品与农业标准化研究所,北京 100088
  • 出版日期:2015-05-25 发布日期:2015-05-08
  • 通讯作者: 支瑞聪
  • 基金资助:

    国家高技术研究发展计划(863计划)项目(2011AA1008047);国家自然科学基金青年科学基金项目(31201358);
    北京市优秀人才培养资助项目(2012D009999000001)

Evaluation of Longjing Tea Quality Based on Aroma Characteristics

DAI Yuewen1, ZHI Ruicong2,*, ZHAO Lei2, GAO Haiyan1, SHI Bolin2, WANG Houyin2   

  1. 1. College of Life Science, Shanghai University, Shanghai 200444, China;
    2. Institute of Food and Agriculture Standardization, China National Institute of Standardization, Beijing 100088, China
  • Online:2015-05-25 Published:2015-05-08
  • Contact: ZHI Ruicong

摘要:

结合人工感官审评和智能感官分析对4 个等级西湖龙井茶进行识别判定。通过相关性分析和主成分分析,先后建立龙井茶香气分属性权重及龙井茶香气分属性与电子鼻传感器关联性。根据龙井茶香气分属性权重及香气分属性与传感器关联性结果,对电子鼻传感器进行筛选。通过核Fisher判别分析法和K-最近邻算法进行进一步特征提取和模式分类,实现了对于训练集样本100%和测试集样本97.5%的正确识别。

关键词: 西湖龙井茶, 人工感官审评, 智能感官分析, 核Fisher判别分析

Abstract:

A combination of artificial sensory evaluation with intelligent sensory analysis was used in discrimination and
identification of 4 grades of Xihu Longjing tea. Weights of aroma attributes and the correlation between aroma attributes and
electronic nose sensors were established by correlation analysis and principal component analysis (PCA) and according to
the results obtained, the optimal sensors of electronic nose were selected. Kernel Fisher discriminant analysis (KFDA) and
K-nearest neighbor (KNN) were utilized for further feature extraction and pattern recognition, and the correction coefficient
of the training and test sets were 100% and 97.5%, respectively.

Key words: Xihu Longjing tea, artificial sensory evaluation, intelligent sensory analysis, kernel Fisher discriminant analysis

中图分类号: