显微红外光谱法无损识别林下山参与园参

显微红外光谱分析技术是将红外光谱仪和显微镜结合发展而来的分析技术。该技术样品处理简单，可用于微区分析和微量分析，并且能反映样品的本质特征。人参包括林下山参、园参和野山参，而我国禁止采挖野山参。因此，药典只收录了林下山参和园参。该工作使用显微红外光谱技术，结合判别分析建立了对林下山参和园参的无损识别模型。收集了20个林下山参和24个园参后，采集了其显微红外光谱。其中，33个样品被随机划分为校正集，11个为验证集。建立模型时，优化了预处理方法、主成分数、建模波段和扫描部位。最后，使用多元散射校正+Savitzky-Golay平滑的预处理方法、3932.14～669.18 cm-1的波段、4个主成分数和芦头部位建立了最优的判别分析模型，准确率达到100%。结果表明，结合判别分析的显微红外光谱技术，样品处理简单、快速、无损、有效，可用于林下山参和园参的无损识别。

Nondestructive Recognization of Mountain Cultivated Ginseng and Garden Cultivated Ginseng by FTIR Microspectroscopy

Fourier transform infrared (FTIR) microspectroscopy technology is the combination of the FTIR spectrometer and the microscope. This technology is of simple preparation of the samples, can be used in micro-area analysis and micro-samples, and reflect the nature of the samples spectra. Panax ginseng include mountain cultivated ginseng (MCG), garden cultivated ginseng (GCG) and mountain wild ginseng (MWG), but the excavation of MWG is prohibited in China. So, only MCG and GCG were collected and recorded in Chinese pharmacopoeia. In this study, we developed a discriminant analysis (DA) method for recognition of MCG and GCG using FTIR microspectroscopy technology. Twenty MCG samples and twenty four GCG samples were obtained, and their spectra of IR microspectroscopy were collected. Then 33 samples were randomly selected into calibration set and the remaining 11 of the samples were selected into validation set. The authors optimized the pretreatment method, the principal components, the modeling region and the scanning parts when developing the models. The optimized model of discriminant analysis was developed using the pretreatment multiplicative scatter correction (MSC)+Savitzky-Golay filter (SG) smoothing, the region 3 932.14～669.18 cm-1, 4 principal components and the rhizome part. The accuracy of the optimized model got up to 100%. The result demonstrated that infrared microspectroscopy technology combined with DA is of simple operation, rapid, nondestructive and effective, and can be applied to recognize MCG and GCG.