西湖龙井茶品质的智能嗅觉识别  

Quality Recognition of Xihu Longjing Tea Based on Intelligent Olfactory

作　　者：史波林[1] 赵镭[1] 支瑞聪[1] 席兴军[1] 朱大洲 

机构地区：[1]中国标准化研究院食品与农业标准化研究所,北京100088 [2]国家农业智能装备工程技术研究中心,北京100097

出　　处：《农业机械学报》2012年第12期130-135,共6页Transactions of the Chinese Society for Agricultural Machinery

基　　金：国家高技术研究发展计划(863计划)资助项目(2011AA1008047);北京市优秀人才培养资助项目(2012D009999000001)

摘　　要：通过电子鼻采集不同等级、树种和产地西湖龙井茶的智能嗅觉指纹图谱,利用主成分分析得分矩阵研究等级、树种和产地指纹信息对茶叶品质的影响程度,基于软独立模型分类分析方法建立茶叶等级、树种和产地的3类智能嗅觉判别模型。结果表明,不同等级西湖龙井茶的电子鼻信号差异最大;在涵盖不同树种和产地信息的样品中,电子鼻能准确预测品质相近的高档等级(精品、特级和一级)茶叶,等级判别正确率基本达到100%。树种与产地特征对于茶叶品质的影响程度比较接近,并且同一等级、同一产地下不同树种模型和同一等级、同一树种下不同产地模型的判别正确率基本都达到92%以上。在此基础上,提出了首先利用电子鼻进行等级划分,然后在同一等级下进行树种鉴定和产地判别的西湖龙井茶品质智能嗅觉快速检测策略。The intelligent olfactory spectrograms of Xihu Longjing tea were collected by electronic nose. The spectrograms scores matrix that obtained by principal component analysis was applied. The influence of the information from grading, tree varieties and producing areas on quality of tea was analyzed. The soft independent modeling of class analogy was used to establish three kinds of models for grading, tree varieties and producing areas, respectively. Results showed that the intelligent olfactory spectrograms from different grading tea had the most difference than the other two elements. The grading discrimination model had well performance with about 100% correct rate to predict top grade teas. Meanwhile, the influences of tree varieties and location on the tea quality were seemed to be similar. Under the same grade and same producing area, the tree varieties discrimination model of tea was built. Also, the producing area discrimination model of tea was built. The correct rates of the two models were all over 92%. On the basis of above study, the quick detection strategy of Xihu Longjing tea by intelligent olfactory technology was presented, that is, grading demarcation, tree varieties identification and producing area distinguishing under the same grading were carried out in turn.

关 键 词：西湖龙井茶 电子鼻 智能嗅觉 特征识别 

分 类 号：TP242.64[自动化与计算机技术—检测技术与自动化装置] TS272.7[自动化与计算机技术—控制科学与工程]