基于FTIR的两种模式识别方法在扶芳藤产地鉴别中的应用

目的：建立快速鉴别扶芳藤产地的方法。方法：采用傅里叶变换红外光谱（FTIR）技术,选取1800～400cm-1红外指纹图谱区,采用软独立建模聚类分析（SIMCA）法和Fisher比值法建立产地分类模型,分别对广西4个产地的扶芳藤进行鉴别,并通过计算识别率和拒绝率来验证模型。结果：FTIR技术不仅能准确提取扶芳藤药材的整体信息,而且所建立的2种识别模型对未知扶芳藤样品鉴别的识别率和拒绝率均达到100%,能正确识别4个不同产地的扶芳藤样品。结论：所建模型具有较强的识别能力,采用自编计算程序及现有统计软件还可将模式识别模型实现实时在线检测与样品快速鉴别,大大提高了鉴别速度。

Application of 2 Recognition Patterns in the Identification of Producing Area of Euonymus fortunei Based on FTIR

OBJECTIVE： To establish a rapid method for the identification of producing area of Euonymus fortunei. METHODS： Using FTIR, the section of 1 800～400 cm-1 which was the IR fingerprint of E. fortunei were selected, and class analogies （SIMCA） and Fisher ratio were used to build a classification model. E. fortunei from 4 producing area of Guangxi region were identified respectively, at the same time, the model was corrected by calculating their recognition rate and the rejection rate. RESULTS： Not only FITR accurately extracted the overall information of E. fortunei, but also recognition rate and rejection rate of established 2 recognition patterns to unknown samples reached 100% and they could indentify the sample of E. fortunei from 4 producing areas. CONCLUSION： Established pattern has stronger capability of identification. The self-designed computation program and present software achieve real-time detection and rapid sample identification to improve the identification efficiency greatly.