用高效液相色谱法和模式识别技术定量表征烟叶颜色相似性研究

Study on Quantitative Characterization of Color Similarity among Flue-cured Tobacco Leaves by HPLC and Pattern Recognition

用高效液相色谱法(HPLC)分析了13种国产烤烟型烟叶中多酚类物质,并定量测定了其中7种主要烟草多酚物质,用模式识别方法处理了HPLC测定多酚的谱图轮廓数据,计算了这些烟叶之间的颜色相似性标度参数TK值,并与人工感官结果进行了比较。结果表明,通过用模式识别技术处理多酚的HPLC轮廓数据,可以定量估计不同产地、不同等级烟叶之间颜色的相似性。模式识别方法处理与感官评价结果基本一致,但前者更客观。因此,模式识别技术可以作为客观评价烟叶颜色质量、弥补人工感官方法的工具。

Polyphenols in thirteen Chinese flue-cured tobacco leaves were assayed by high performance liquid chromatography（HPLC）. The major seven tobacco polyphenols were quantitatively determined. The HPLC profiles of polyphenol were processed by pattern recognition method （PRM）. The values of color similarity criterion TK among these tobacco leaves were calculated and compared with human organoleptic color assessment. It was shown that the data from PRM based on HPLC profiles of polyphenols could be used to estimate the color similarity among tobacco leaves of different grades from different are as. The results obtained by PRM agreed quite well with those obtained by organoleptic assessment, but the former was more objective.