基于时频域特征融合的龙井茶品质判定  

Quality Recognition of Longjing Tea Based on Time and Frequency-Domain Feature Fusion

作　　者：支瑞聪[1,2] 赵镭 高海燕[4] 戴悦雯[4] Zhi Ruicong;Zhao Lei;Gao Haiyan;Dai Yuewen(School of Computer and Communication Engineering,University of Science and Technology Beijing,Beijing 100083;Beijing Key Laboratory of Knowledge Engineering for Materials Science,Beijing 100083;Institute of Food and Agriculture Standardization,China National Institute of Standardization,Beijing 100191;School of Life Science,Shanghai University,Shanghai 200444)

机构地区：[1]北京科技大学计算机与通信工程学院,北京100083 [2]材料领域知识工程北京市重点实验室,北京100083 [3]中国标准化研究院食品与农业标准化研究所,北京100191 [4]上海大学生命科学学院,上海200444

出　　处：《中国食品学报》2018年第9期303-310,共8页Journal of Chinese Institute Of Food Science and Technology

基　　金："十三五"国家重点研发计划重点专项(2017YF D0400100);国家自然科学基金项目(61673052);国家863计划项目(2011AA1008047)

摘　　要：智能感官分析技术是模拟人的感官获取茶叶滋味特征信息,用于茶叶品质自动检测的有效方法。本文采用电子舌智能感官分析仪器采集不同等级西湖龙井茶的智能味觉指纹图谱,从时域和频域两个角度提取电子舌传感器响应信号参数,并将时域特征参数和频域特征参数进行特征融合,然后茶叶样品特征分别采用线性降维方法(主成分分析、线性判别分析)和非线性降维方法(核主成分分析、核线性判别分析)进行维数约简,采用最近邻分类器判定茶叶等级。对单特征参数和多特征参数以及线性降维和非线性降维方法的算法效果进行比较,结果龙井茶等级判定的最高正确识别率在95%以上,实现了电子舌对不同等级龙井茶样品的自动模式分类。Intelligent sensory analysis could represent the tasting characteristic information by simulating human senses, and it has been applied to tea quality identification successfully. In this paper, electronic nose（E-tongue） was employed for quality classification of Xihu-Longjing tea by collecting the sensing signal information. The signal information was processed from time-domain and frequency-domain independently, and the fusion features with two kinds of features were utilized for dimensionality reduction, by linear dimensional reduction methods（PCA LDA） and nonlinear dimensional reduction methods（KPCA KLDA）. Comparison was conducted between single feature and fusion feature, linear dimensional reduction and nonlinear dimensional reduction, and the highest recognition accuracy achieved higher than95%.

关 键 词：电子舌 特征融合 时频域 非线性降维 龙井茶 

分 类 号：TS272.5[农业科学—茶叶生产加工]