论文部分内容阅读
目的:建立银杏叶中总黄酮醇苷的近红外定量模型。方法:收集52批不同产地不同批号银杏叶样品的近红外光谱图,运用TQ Analyst 8.0软件,将其经二阶导数预处理,在6 250~4 000 cm-1范围内,选择4个主成分数,采用偏最小二乘法(PLS)建立银杏叶中总黄酮醇苷含量的近红外光定量模型。结果:所建模型预测集相关系数r=0.9541,校正集均方根偏差(RMSEC)=0.00799,预测均方根偏差(RMSEP)=0.0707,验证集预测回收率为100.77%(n=7)。结论:该研究所建立的模型性能较好。近红外光谱法具有快速、简便、准确、无损的优点,可以应用于银杏叶药材的质量控制及后续产品的开发利用。
Objective: To establish a quantitative model of total flavonol glycosides in Ginkgo biloba leaves by near infrared spectroscopy. Methods: Near infrared spectra of 52 batch samples of Ginkgo biloba from different producing areas were collected. TQ Analyst 8.0 software was used to pretreat the sample with Ginkgo biloba leaf extract in the range of 6 250-4 000 cm-1. Four principal components The PLS method was used to establish a quantitative model of total flavonol glycosides in Ginkgo biloba leaves by near-infrared spectroscopy. Results: The correlation coefficient of prediction model set was r = 0.9541, root mean square error of calibration (RMSEC) = 0.00799, root mean square error of prediction (RMSEP) = 0.0707, and predictive recovery rate of validation set was 100.77% (n = 7). Conclusion: The model established by this research has good performance. Near infrared spectroscopy has the advantages of fast, simple, accurate, non-destructive, can be applied to the quality control of Ginkgo biloba medicine and the follow-up product development and utilization.