近红外光谱技术在陈皮道地性分析中的应用

Genuineness Identification of Citrus reticulata by Near Infrared Reflectance Spectroscopy

目的:建立以近红外光谱技术鉴别陈皮药材道地性的方法。方法:对不同产地的陈皮样品进行近红外光谱数据采集,用模式识别方法聚类分析,建立判别模型并对模型进行验证。结果:道地陈皮样品(广陈皮)与非道地样品之间有较好的分离度;所建模型对已知训练集样本的分类精度为99.98%,对未知样本的预测精度为100%,该模型具有很好的预测性能,可准确鉴别陈皮药材的道地性。结论:近红外光谱技术性能优良,是一种简便、快速、不破坏样品的绿色分析方法。

OBJECTIVE： To establish a near infrared reflectance spectroscopic （NIRS） method for the identification of Citri reticulata. METHODS： The near infrared spectra of C. reticulata from different habitats were treated by cluster analysis which is a method of pattern recognition. The discrimination model was established to validate the stability of this model. RESULTS： Genuine C. reticulata was well-separated from non-genuine C. reticulata; after modeling, the classification precision of training set reached up to 99.98%, and the forecasting precision also reached 100%. This model possessed good predictability that demonstrated the method can identify genuineness of C. reticulata. CONCLUSION：NIRS is green omalysis method, and it is excellent, simple, fast and nondestructive, which have potential application.