水平衰减全反射傅里叶变换红外光谱法结合化学计量学鉴别问荆与节节草

Recognition of Equisetum Arvense L.and Hippochaete Ramosissima (Desf.) Boerner.Based on Fourier Transform Infrared Spectrometry Combined with Chemometrics

采用水平衰减全反射傅里叶变换红外光谱法(HATR-FT-IR)测定中药问荆和同科植物节节草的FT-IR谱,运用连续小波变换(CWT)多分辨率分析法对吸收较为相似的问荆及节节草的FT-IR进行特征提取。选择第8、9、10分解层数下的特征值作为分析的基础,采用FT-IR-CWT-SVM法建立了问荆和节节草识别的模型。通过学习训练,对120个预测样品的识别准确率为90%以上。当采用径向基函数作为核函数时,识别准确率达100%。样品的FT-IR经小波特征提取后的特征值有所差异,采用SVM进行识别可以很好地把两者分类。通过对样品的FT-IR小波变换后所得特征值进行SVM的分类,能够有效地进行区别鉴定形态较为相似的同科植物问荆及节节草。

Fourier transform infrared （FT-IR） and horizontal attenuated total reflectance （HATR） techniques were used to obtain the FT-IR of Equisetum arvense L. and Hippochaete ramosissima （Desf.） Boerner. Features of their similar absorptiong were extracted by continuous wavelet transform （CWT）. The features at decomposition level 8, 9 and 10 are used as basic data. A model of their discrimination was established by the FT-IR-CWT-SVM support vector machine. The accuracy of discrimination for the 120 predicable samples was all above 90% by training, and when radial basic function （RBF） was used to be kernel, its accuracy of discrimination was exactly 100% by training. As the eigenvalue of the FT-IR of the samples were different after continuous wavelet feature extraction, it is well to classify the two plants by adopting the SVM to identify. By using SVM to classifiy the eigenvalues those were extrated by continuous wavelet, we could effectively identify similar plants in morphology Equisetum arvense L. and Hippochaete ramosissima （Desf.） Boerner.