光谱匹配算法应用于鸡血藤与大血藤的太赫兹光谱区分

发布时间：2018-03-25 08:32

本文选题：太赫兹时域光谱　切入点：鸡血藤　出处：《光谱学与光谱分析》2017年01期

【摘要】：鸡血藤和大血藤是药效、产地不同而外形极相似的两种不同的藤茎类药材,使用时极易混淆。利用太赫兹时域光谱技术获得测试样品太赫兹频谱,并结合光谱匹配算法分辨五种不同产地、不同批次的鸡血藤和大血藤。因中药化学成分的复杂性导致其不同成分的特征频谱发生相互的混叠,使得鸡血藤和大血藤的太赫兹时域数谱均没有出现明显的特征峰。重复实验测试同一样品同一位置10次,通过计算得到不同频率下光谱数据的标准差。数据在0.2~1.3THz范围内有较好的稳定性。对光谱数据进行二阶微分处理,使光谱间的微小差异得以显现。选取0.2~1.3THz以及0.2~2范围内的光谱数据进行对比研究,分别使用相关系数法、分段编码法、光谱角法和信息散度法四种光谱匹配算法对实验数据所属类别进行判定。数据分析结果:在0.2~1.3THz范围内,光谱角算法对鸡血藤与大血藤的太赫兹光谱数据有最好的分类效果,对两类中药样品总计100组数据的分类正确率达到95%。为易混中药的鉴定区分提供了一种新的技术与方法。
[Abstract]:It is easy to be confused when using two different kinds of vine stem medicinal materials with different origin and very similar appearance. Terahertz time-domain spectroscopy is used to obtain terahertz spectrum of test samples. The spectral matching algorithm was used to identify five different producing areas, different batches and different batches of Chinese medicinal herbs. Due to the complexity of the chemical components of traditional Chinese medicine, the characteristic spectrum of the different components overlapped with each other. The terahertz time-domain spectrum of both the chicken and the big blood vine has no obvious characteristic peak. The same sample has been tested 10 times in the same position. The standard deviation of spectral data at different frequencies is obtained. The data has good stability in the 0.2~1.3THz range. The spectral data is processed by second-order differential processing. The spectral data in the range of 0.2~1.3THz and 0.2 ~ 2 are compared and studied respectively by using correlation coefficient method and piecewise coding method. Spectral angle method and information divergence method are four spectral matching algorithms to determine the classification of experimental data. Data analysis results: within the 0.2~1.3THz range, the spectral angle algorithm has the best classification effect on the terahertz spectral data of Hedysticum and Dioscorea. The classification accuracy of 100 groups of data from two kinds of Chinese medicine samples is 95%, which provides a new technique and method for the identification and differentiation of easily mixed Chinese medicine.
【作者单位】：天津大学精密测试技术及仪器国家重点实验室;天津大学太赫兹中心;
【基金】：科技部重大仪器专项(2011YQ150021)资助
【分类号】：R284;O433.4

【相似文献】