基于红外光谱技术建立宁夏中宁产区枸杞子判别模型的应用研究

Application of Discriminant Model of Chinese Wolfberry in Ningxia Based on Infrared Spectroscopy

[目的]采用傅里叶变换红外光谱技术实现对宁夏中宁产区枸杞子的快速鉴别。[方法]分别采集宁夏、甘肃、青海、内蒙古4个不同产地40份枸杞子样品进行扫描,对其红外光谱分别进行聚类分析和主成分判别分析,建立中宁产区枸杞子产地判别模型。[结果]在900~1 700 cm^(-1)波数范围内,建立主成分分析判别模型,对样品的识别率均达到100%,模型预测效果好;采用组间均联法,利用马氏距离作为样品的测度,进行聚类分析,可将不同产区枸杞子进行区分,样品判别率达到100%。[结论]采用聚类分析和主成分分析对不同产区枸杞子红外光谱进行分析,该方法在枸杞子产地判别中具有可行性,能够快捷、准确地鉴定中宁枸杞子。

[Objective] The rapid identification of Chinese wolfberry from Zhongning in Ningxia was realized by Fourier transform infrared spectroscopy（FTIR）. [Methods] To establish a discriminant model of the origin of Chinese wolfberry from Zhongning based on infrared spectra by clustering analysis and principal component analysis. The samples of 40 wolfberries were scanned in 4 different producing areas of Ningxia, Gansu, Qinghai and Inner Mongolia. [Results] The model was established by principal component analysis model within the range of 900 ~ 1 700 cm^（-1） wave. The results showed that the recognition rate of samples was 100% and the model predicted good results; cluster analysis could distinguish the wolfberries in different areas using the method of group and the Mahalanobis distance to measure the sample. The sample discriminant rate reached 100%. [Conclusion] Cluster analysis and principal component analysis were used to analyze the infrared spectroscopy of Chinese wolfberry in different areas. This method is feasible in origin discrimination of Chinese wolfberry, which can quickly and accurately identify the Chinese wolfberry from Zhongning.