【摘要】 为了快速分类鉴别紫石英及其炮制品。首先检测了外貌非常相似的紫石英、方解石、白石英的拉曼光谱，不同产地生紫石英，紫石英的生品与炮制品的拉曼光谱。然后建立了拉曼光谱技术结合主成分分析(PCA)-支持向量机(SVM)算法的分类鉴别模型，并对拉曼光谱数据进行了分类鉴别。结果表明，对于外形相似的紫石英、方解石、白石英而言，拉曼光谱有明显的差异，因此根据拉曼光谱的明显差异通过肉眼或PCA-SVM算法都能够进行准确的分类鉴别。但是对于不同产地的生紫石英，紫石英的生品与炮制品而言，其拉曼光谱非常相似，用肉眼几乎无法进行区分，但是通过PCA-SVM算法根据拉曼光谱数据的微小差异也能够进行准确的分类鉴别，且准确率可达到100%。该方法具有快速、准确、无损、方便、仪器设备便携、成本低等优点，对矿物药的分类鉴别及其质量监控具有重要的应用价值。更多还原

【Abstract】 In order to rapidly classify and identify fluorite and its processed products, the Raman spectra of fluorite, calcite, and quartz, which have very similar appearances, were first detected. Then, the Raman spectra of fluorite from different origins, as well as the raw and processed products of fluorite, were analyzed. A classification and identification model combining Raman spectroscopy with Principal Component Analysis(PCA)-Support Vector Machine(SVM) algorithm was established, and the Raman spectral data were classified and identified. The results show that for the similar appearance of fluorite, calcite, and quartz, there are obvious differences in Raman spectra, so these different kinds of mineral medicines can be accurately classified and identified by the naked eye observing or the PCA-SVM algorithm identifying the obvious differences in the Raman spectra. Although the Raman spectra of the fluorite from different origins, the Raman spectra of fluorite and the processed fluorite are so similar that it is almost impossible to differentiate them with the naked eye, the PCA-SVM algorithm is also able to accurately classify and identify based on small differences in the Raman spectral data. Meanwhile, the accuracy rate can reach 100%. This method has the advantages of being fast, accurate, non-destructive, convenient, portable, and low-cost, which is of great value for the classification and identification of mineral medicines and their quality control.更多还原