摘要
本文以当归药材为研究对象,根据当归在电子鼻中各传感器的响应情况进行产地的鉴别分析。在获得各样品电子鼻气味信息数据的基础上,基于Loading分析法分析各传感器对样品的区分能力,结合多元统计分析法构建相关气味模型并进行评价。结果显示,经PCA分析后第一主成分和第二主成分的总贡献率达到99.87%,而经LDA分析后总贡献率为94.89%,且模型经DFA鉴别验证,正确判别率为100.00%,表明不同产地当归气味信息存在显著差异,4个产地当归区分效果明显。由此可见,基于电子鼻技术的多元统计分析方法对当归的特殊气味进行表征分析,可以实现不同产地当归药材的区分鉴别,效果显著,能够为中药材快速产地区分、品质分析及无损检测提供技术参考。
In this paper,Angelica sinensis is taken as the research object,and the origin of Angelica sinensis is identified and analyzed according to the response of each sensor in the electronic nose.On the basis of obtaining the electronic nose odor information data of each sample,the discrimination ability of each sensor to the sample is analyzed based on the loading analysis method,and the relevant odor model is constructed and evaluated combined with the multivariate statistical analysis method.The results showed that after PCA analysis,the total contribution rate of the first principal component and the second principal component reached 99.87%,while after LDA analysis,the total contribution rate was 94.89%,and the correct discrimination rate of the model was 100%after DFA identification,indicating that there were significant differences in the odor information of Angelica sinensis from different producing areas,and the differentiation effect of Angelica sinensis from four producing areas was obvious.It can be seen that the multivariate statistical analysis method based on electronic nose technology can characterize and analyze the special smell of Angelica sinensis,which can realize the differentiation and identification of Angelica sinensis from different producing areas,with remarkable effect,and can provide technical reference for rapid origin differentiation,quality analysis and nondestructive testing of traditional Chinese Medicine.
作者
刘阿静
王娟
王波
王慧珺
张红艳
丁欢
LIU Ajing;WANG Juan;WANG Bo;WANG Huijun;ZHANG Hongyan;DING Huan(Lanzhou Customs Technology Center,Lanzhou Gansu,730010,China)
出处
《质量安全与检验检测》
2022年第1期1-5,共5页
QUALITY SAFETY INSPECTION AND TESTING
基金
海关总署科研项目(2019HK110)。
关键词
电子鼻
当归
多元统计分析
Electronic Nose
Angelica Sinensis
Multivariate Statistical Analysis
