应用MDS-SVM算法和拉曼荧光光谱技术快速鉴定天麻的掺假行为  

Rapid identification of adulteration behavior of Gastrodia elata Blume by MDS-SVM algorithm and Raman fluorescence spectroscopy

作　　者：张翠萍[1] 张军惺[1] 刘业炜[2] 何学峰 周明辉 ZHANG Cuiping;ZHANG Junxing;LIU Yewei;HE Xuefeng;ZHOU Minghui(Heibei Sport University,Modern Educational Technology Center,Shijiazhuang 050041,China;Hebei Education Examination Institute,Shijiazhuang 050041,China;School of Artificial Intelligence,Hebei University of Technology,Tianjin 300401,China;College of Information Science and Technology,Hebei Agricultural Unwersity,Baoding 071001,China)

机构地区：[1]河北体育学院,现代教育技术中心,石家庄050041 [2]河北省教育考试院,石家庄050041 [3]河北工业大学人工智能与数据科学学院,天津300401 [4]河北农业大学信息科学与技术学院,保定071001

出　　处：《光散射学报》2025年第1期86-93,共8页The Journal of Light Scattering

基　　金：2023年度河北省“三三三人才工程”资助项目:基于大数据挖掘技术的射击训练智能决策支持模式的研究(C20231118)。

摘　　要：天麻是一种重要的中药材,天麻粉常被用于中药制剂、保健品和食品添加剂等领域,然而不法商贩为降低成本或提高重量会在天麻粉中掺入其他成分。为实现一种快速、无损、高灵敏检测天麻粉掺假行为的激光光谱检测技术,本文开展了天麻粉掺假的拉曼荧光光谱的实验研究,通过对纯天然天麻粉和淀粉的谱学研究,分析了淀粉拉曼光谱的特征,提出采用多维尺度变换结合支持向量机(MDS-SVM)的定量模型,通过测量模拟掺假粉末的拉曼荧光光谱,建立了天麻粉掺假预测的MDS-SVM定量模型,模型预测效果较好,其中模型对测试集的预测结果的决定系数为0.8933,均方根误差为0.0131。结果表明拉曼荧光光谱技术具有表征天麻粉成分信息的性能,且结合MDS-SVM算法可以实现快速鉴定天麻粉的掺假行为。本文为天麻粉的快速鉴别提供了一种快速、非侵入式的方法,为相关部门执法应用提供了参考。Gastrodia elata is an essential traditional Chinese medicine.Its powder is often used in traditional Chinese medicine preparations,health products,food additives,and other fields.However,illegal traders will add other ingredients to the powder to reduce costs or increase weight.To achieve a rapid,non-destructive,and susceptible laser spectral detection technology for detecting the adulteration behavior of Gastrodia elata powder,this paper conducted an experimental study on the Raman fluorescence spectrum of Gastrodia elata powder adulteration.Through the spectral research of pure natural Gastrodia elata powder and starch,the characteristics of the starch Raman spectrum were analyzed,and a quantitative model based on multidimensional scale transformation combined with a support vector machine(MDS-SVM)was proposed.By measuring the Raman fluorescence spectrum of simulated adulterated powder,the mds-svm quantitative model for Gastrodia elata powder adulteration prediction was established.The prediction effect of the model was good;the determination coefficient of the prediction results of the model on the test set was 0.8933,and the root mean square error was 0.0131.The results show that Raman fluorescence spectroscopy has the performance of characterizing the composition information of Gastrodia elata powder and combining it with the MDS-SVM algorithm,which can quickly identify the adulteration behavior of Gastrodia elata powder.This paper provides a rapid and non-invasive method for the rapid identification of Gastrodia elata powder and a reference for the law enforcement application of relevant departments.

关 键 词：天麻 拉曼光谱 荧光 多维尺度变换 支持向量机 

分 类 号：O433[机械工程—光学工程] O657.3[理学—光学]