欢迎访问《茶叶科学》,今天是

茶叶科学 ›› 2020, Vol. 40 ›› Issue (5): 689-695.doi: 10.13305/j.cnki.jts.2020.05.014

• 研究报告 • 上一篇    下一篇

小种红茶茶多酚和咖啡碱近红外定量分析模型的建立

卢莉1, 程曦1, 张渤1, 沈小霞1, 刘艳1, 熊丽1, 袁潇2, 李远华1,*, 黎星辉2,*   

  1. 1.武夷学院茶与食品学院,福建 武夷山 354300;
    2.南京农业大学园艺学院,江苏 南京 210095
  • 收稿日期:2020-06-12 修回日期:2020-07-25 出版日期:2020-10-15 发布日期:2020-10-10
  • 通讯作者: * yhli@wuyiu.edu.cn; lxh@njau.edu.cn
  • 作者简介:卢莉,女,讲师,主要从事茶树栽培育种与茶叶加工技术研究。
  • 基金资助:
    福建省自然科学基金(2017J01648)、福建省科技创新平台建设(2018N2004)、国家自然科学基金(31870682)

Establishment of Predictive Model for Quantitative Analysis of Tea Polyphenols and Caffeine of Souchong by Near Infrared Spectroscopy

LU Li1, CHENG Xi1, ZHANG Bo1, SHEN Xiaoxia1, LIU Yan1, XIONG Li1, YUAN Xiao2, LI Yuanhua1,*, LI Xinghui2,*   

  1. 1. College of Tea and Food Sciences, Wuyi University, Wuyishan 354300, China;
    2. College of Horticulture, Nanjing Agricultural University, Nanjing 210095, China
  • Received:2020-06-12 Revised:2020-07-25 Online:2020-10-15 Published:2020-10-10

摘要: 以76份有代表性的小种红茶为研究对象,采用现行国标方法测定的茶多酚和咖啡碱含量作为近红外预测模型的化学值,对应采集样品的近红外光谱值,分别建立小种红茶茶多酚和咖啡碱含量最佳偏最小二乘法(Partial least squares,PLS)模型。结果表明,所构建的茶多酚含量模型校正集决定系数(Coefficient of determination,R2)为97.59%,校正均方差(Root mean square error of calibration,RMSEC)为0.566%,验证集R2为95.06%,预测均方差(Root mean square error of prediction,RMSEP)为0.855%;咖啡碱含量模型校正集R2为96.98%,RESEC为0.110%,验证集R2为95.67%,RESEP为0.148%。茶多酚和咖啡碱含量定量分析模型效果均较好,可实现对小种红茶茶多酚和咖啡碱含量的快速检测。

关键词: 小种红茶, 近红外光谱, 茶多酚, 咖啡碱, 模型

Abstract: A total of 76 representative Souchong samples were studied. Quantitative models of tea polyphenols and caffeine of Souchong were established respectively by partial least squares (PLS). The chemical values of tea polyphenols and caffeine of 76 Souchong samples were measured by Chinese National Standard methods, combined with the near infrared spectrum. The results show that the R2 and RMSEC values of tea polyphenols of calibration set were 97.59% and 0.566%. And the R2 and RMSEP values of validation set were 95.06% and 0.855%. The R2 and RMSEC values of caffeine of calibration set were 96.98% and 0.110%. And the R2 and RMSEP values of validation set were 95.67% and 0.148%. The quantitative models had good prediction results, and could be used for rapid evaluation of the contents of tea polyphenols and caffeine in Souchong.

Key words: Souchong, near infrared spectroscopy (NIRS), tea polyphenols, caffeine, models

中图分类号: