高光谱成像技术结合化学计量学方法鉴别大花红景天和狭叶红景天  

Identification of Rhodiola crenulata and R.kirilowii by Hyperspectral imaging technology combined with chemometrics

作　　者：曲明亮 李涛[1] 钟玉琴 QU Mingliang;LI Tao;ZHONG Yuqin(West China School of Pharmacy,Sichuan University,Chengdu,Sichuan,610041 P.R.China)

机构地区：[1]四川大学华西药学院,四川成都610041

出　　处：《华西药学杂志》2021年第3期303-307,共5页West China Journal of Pharmaceutical Sciences

基　　金：四川省科学技术厅应用基础研究计划项目(编号:2020YJ0275);四川省中医药管理局全国第四次中药资源普查科技专项项目(编号:2019PC003);四川大学泸州市人民政府战略合作资金项目(编号:2017CDLZ-S24)。

摘　　要：目的采用高光谱成像技术结合化学计量学方法鉴别大花红景天和狭叶红景天。方法利用高光谱成像系统,采集大花红景天和狭叶红景天药材粉末在900~1700 nm的高光谱图像,利用标准正态变换的方法进行光谱图像预处理,基于Kennard-Stone算法将样本划分为训练集和预测集,采用连续投影算法、竞争性自适应加权采样从全波长中选取特征波长,分别建立了基于全波长段和特征波长段的支持向量机和极限学习机的判别模型。以预测集的分类准确率、均方根误差、平方相关系数为主要衡量标准,比较分析各模型的性能,选择最优的判别模型。结果不同判别模型的分类准确率都在85%以上,其中,基于竞争性自适应加权采样特征波长选择方法的极限学习机模型的准确率达到了100%。结论采用高光谱成像技术结合化学计量学方法,能够实现对大花红景天和狭叶红景天的在线快速、无损鉴别,可为红景天药材的品种鉴别、质量控制和品质评价提供参考。OBJECTIVE To identify Rhodiola crenulata and R.kirilowii by Hyperspectral imaging technology combined with chemometrics.METHODS The hyperspectral images of R.crenulata and R.kirilowii powders at 900-1700 nm were collected using a hyperspectral imaging system.Spectral image preprocessing was performed using the standard normal variate(SNV)method,and the samples were divided into a training set and a prediction set based on the Kennard-Stone(KS)algorithm.SPA and CARS was used to select characteristic wavelengths from the full wavelength.The discriminant models of SVM and ELM based on the full wavelength and characteristic wavelength were established,respectively.The classification accuracy,RMSE,and R2 of the prediction set were used as the main measurement criteria to compare and analyze the performance of models,and the optimal discriminant model was selected.RESULTS The classification accuracy of different discriminant models was more than 85%,among which the accuracy of the CARS-ELM model reached 100%.CONCLUSION The use of hyperspectral imaging technology combined with chemometrics method can realize the fast and non-destructive identification of R.crenulata and R.kirilowii,which can provide a reference for the identification,quality control and quality evaluation of Rhodiola.

关 键 词：大花红景天 狭叶红景天 高光谱成像技术 支持向量机 极限学习机 化学计量学 连续投影算法 竞争性自适应加权采样 质量控制 品质评价 

分 类 号：R917[医药卫生—药物分析学]