红外光谱法结合系统聚类和SIMCA模式识别法快速鉴别羌活种子

目的 采用傅里叶变换红外光谱法鉴别羌活种子。方法 在4 000～400 cm^?1内测定羌活种子光谱吸收峰，应用系统聚类和SIMCA模式识别法鉴别羌活和宽叶羌活的种子。结果 运用系统聚类法，当聚类距离为15时，两种羌活可明显分为2类；运用SIMCA模式识别法，所建模型对羌活和宽叶羌活的种子识别率分别达96%和97%，拒绝率均为100%，两者盲样检测准确率也在90%以上，所建模型可用于样品检测。结论 应用红外光谱法结合系统聚类和SIMCA模式识别法，可以简单、快速、准确地鉴别羌活种子，该方法为鉴别羌活种子的来源及真伪提供了一种新的技术手段。

Rapid seed identification of Notopterygii Rhizoma et Radix by SIMCA based on FTIR

Objective To identify the seeds of Notopterygii Rhizoma et Radix by Fourier transform infrared spectrometry (FTIR). Methods Based on the fingerprint infrared spectrum from 4 000 to 400 cm^?1, hierarchical cluster and Soft Independent Modeling of Class Analogy (SIMCA) analyses were developed to identify the seeds of Notopterygium incisum and N. franchetii. Results The seeds could be divided into two categories with 15 Euclidean distance by using hierarchical cluster analysis. Using SIMCA analysis, the recognition rate of N. incisum and N. franchetii reached to 96% and 97%, respectively, and the rejection rate reached to 100%. When testing with the blind sample which the authors picked out from the chosen samples, the accuracy rate reached over 90%. The established model could be used for sample testing. Conclusion Both clustering analysis and FTIR combined with SIMCA provide an effective way to evaluate the seeds of Notopterygii Rhizoma et Radix rapidly and accurately. Therefore, this method provides a new technique for evaluating the source and authenticity of the seeds of Notopterygii Rhizoma et Radix.