荧光光谱法结合Fisher判别分析在西洋参鉴别中的应用

发布时间：2018-08-31 18:10

【摘要】：旨在建立可靠的Fisher判别模型,以实现西洋参及其常见伪品饮片的快速、客观、准确鉴别,采用自组的凝视式光谱成像仪,对90份不同市售来源的中药材饮片(西洋参、人参、桔梗各30份)进行了荧光光谱成像实验,波长范围为400~720nm,成像间隔为5nm。采用标准正态变量(SNV)变换对原的光谱数据进行预处理,以减少光谱数据中的噪声干扰。比较了主成分分析(PCA)与逐步判别分析(SDA)的原理特点及对模型的优化效果,联合这两种分析方法,首先,应用PCA对预处理后的光谱数据进行处理,使光谱数据中的主要信息集中分布在前面的主成分中,然后应用SDA从65个主成分中筛选出判别能力较强的12个主成分建立Fisher判别模型。由所建模型的两个判别函数作样品得分散点图,各类样品在图中表现出良好的聚类现象。以待判样品点与各种类中心点之间的欧氏距离作为依据,得出模型的准确判别结果。结果显示,所建Fisher判别模型在训练集和预测集中的判别正确率分别为98.33%和96.67%,具有较高的可信度与准确度,因此,荧光光谱法结合Fisher判别分析可用于快速鉴别西洋参及其伪品饮片。
[Abstract]:The purpose of this study was to establish a reliable Fisher discriminant model for rapid, objective and accurate identification of Panax quinquefolium and its common adulterants. 90 samples of Chinese herbal medicines (Panax quinquefolium, Panax ginseng, Panax quinquefolium) from different marketable sources were determined by a self-assembled staring spectral imager. Fluorescence spectral imaging experiments were carried out in 30 samples each of Platycodon grandiflorum. The wavelength range was 400 ~ 720nm and the imaging interval was 5 nm. The standard normal variable (SNV) transform is used to preprocess the original spectral data in order to reduce the noise interference in the spectral data. The principle and characteristics of principal component analysis (PCA) and stepwise discriminant analysis (SDA) are compared, and the optimization results of the model are compared. The two analytical methods are combined. Firstly, the pretreatment spectral data are processed by PCA. The main information in the spectral data is distributed in the front principal component, and then the Fisher discriminant model is established by using SDA to screen 12 principal components with strong discriminant ability from 65 principal components. The two discriminant functions of the model are used as the sample score scatter plot, and all kinds of samples show a good clustering phenomenon in the graph. Based on the Euclidean distance between the sample points and the center points of various classes, the accurate discriminant results of the model are obtained. The results show that the accuracy of the Fisher discriminant model is 98.33% and 96.67% in the training set and prediction set, respectively. Fluorescence spectrum method combined with Fisher discriminant analysis can be used to quickly identify American ginseng and its counterfeit pieces.
【作者单位】：暨南大学物理系思源实验室;南方医科大学中医药学院;
【基金】：国家自然科学基金项目(61475065) 广东省自然科学基金项目(2015A030310458)资助
【分类号】：R282.5;O433

【相似文献】