本文在针对提取率和异构化现象优化了样品制备程序的基础上,对杭州、丽水和新昌三个地区以龙井43、群体种和迎霜品种加工制作的9类扁形绿茶产品,尝试了由两个色谱图采用多元信息融合方法进行象素级融合,制作出各自的数字化多元化学指纹模式图谱,并用独立样本进行了验证,分析了不同类样本模式图谱间的相似性程度。结果表明,所建立的化学指纹模式图谱较好地代表了其来源样本的特征属性,同一品种不同产地或同一产地不同品种绿茶样本指纹图谱间,其相似度存在显著差异,但前者间的差异不如后者间差异大,此结果表明特色名优产品应该考虑由一种或几种相类似的品种来制作,从而保证其风味品质的相对统一。

The digital multiple chemical fingerprints of nine groups of flat-shaped green tea produced in different district of Hangzhou, Lishui and Xinchang city, and made from three cultivars, including Longjing 43, jat and Yingshuang, were made by conjoining of two HPLC chromatograms, after the sample preparation procedure being optimized for better extraction and less isomerization. Also validation analysis was made by using 14 tea samples collected independently from above mentioned districts, and similarities between those 9 kinds of pattern fingerprints were analyzed. The result showed that those established pattern fingerprints well represented the characteristic attributes of those origin samples respectively. Distinct differences were found between the pattern fingerprints of samples from different district using same cultivar, or from different cultivar made in same district, but the difference between the former was not obviously as much as that between the latter. So it was reminded that premier green teas should be made from one optimum cultivar or at least several similar suitable cultivars to guarantee the unification of the flavor and the character.