Seven kinds of Chinese herbs and processed products were rapidly identified by near infrared dif-fuse reflectance spectroscopy(NIRDRS)combined with PCA-Mahalanobis distance discriminant analysis.The near infrared diffuse reflectance spectra of seven kinds of Chinese herbs and processed products were pre-pro-cessed by multiplicative scatter correction(MSC)and followed by PCA-Mahalanobis distance discriminant anal-ysis treatment.Perfect classification ratio with 100% was obtained for the MSC fingerprint information of the NIRDRS based on PCA-Mahalanobis distance model,which indicated that the model had good classification pre-dictive ability.PCA-Mahalanobis distance discriminant analysis can extract otherness of MSC fingerprint infor-mation to identify the different kinds of similar Chinese herbs and processed products as a simple,rapid,accurate and green method.%运用近红外漫反射技术(NIRDRS)结合主成分-马氏距离判别分析方法(PCA-Mahalanobis distance discrim-inant analysis)对7种中药材及炮制品进行鉴别。采集7种中药材和炮制品的近红外漫反射光谱,经多元散射校正(MSC)后,运用主成分分析将多元散射光谱变量降维,再利用马氏距离判别分析方法建立模式识别模型。所得 PCA-Mahalanobis distance模型对7种中药材和炮制品的分类正确识别率均为100%,表明模型具有较好的预测分类能力。近红外漫反射技术结合主成分-马氏距离判别分析方法,为判别外观相近容易混淆的不同药材和炮制品提供了一种快速简便、准确可靠的鉴别方法和手段。
展开▼