摘要
目的:利用便携式近红外光谱技术对川贝母及其掺伪品进行快速鉴别及掺伪量快速检测,以建立系统的川贝母掺伪质量评价体系。方法:采集川贝母样品72批并制作不同类别掺伪品(浙贝母、平贝母、伊贝母、湖北贝母、光慈菇、面粉)570批;利用便携式近红外光谱仪采集样品近红外光谱数据;利用线性判别分析(LDA)及偏最小二乘法(PLS)分别建立川贝母-掺伪品、不同类别掺伪品定性校正模型及不同类别掺伪品掺伪量定量校正模型。结果:川贝母及其掺伪品的定性分析模型对川贝母及其掺伪品的识别率分别为99.49%(校正集),100.00%(验证集);在不同类别的掺伪品模型中,校正集和验证集的识别率分别为70.47%和73.68%;6个掺伪量比例的定量模型的验证集相关系数分别为0.8402(川贝母掺伪浙贝母),0.9602(川贝母掺伪平贝母),0.7657(川贝母掺伪伊贝母),0.9025(川贝母掺伪湖北贝母),0.9574(川贝母掺伪光慈菇),0.9761(川贝母掺伪面粉),预测均方根误差(RMSEP)分别为10.9485,5.4639,13.2564,8.5492,5.6553,4.2356;2个定性模型及6个定量模型的预测性能良好。结论:采用便携式近红外光谱技术可实时对川贝母及其掺伪品的快速鉴别和掺伪量快速测定,该方法快速准确,可满足川贝母现场无损真伪鉴别需求。
Objective:In order to establish a systematic quality evaluation system for Fritillariae Cirrhosae Bulbus adulteration,portable near-infrared(NIR)spectroscopy was used to identify Fritillariae Cirrhosae Bulbus and its adulterants and detect their adulteration quantity.Method:A total of 72 batches of Fritillariae Cirrhosae Bulbus samples were collected and 570 batches of adulterated products(dry bulbs of Fritillaria thunbergii,F.ussuriensis,F.pallidiflora and F.hupehensis,Bulbus Tulipae,flour)were prepared,NIR spectral data of samples were collected by the portable NIR spectrometer.Linear discriminant analysis(LDA)was used to establish the qualitative correction models of Fritillariae Cirrhosae Bulbus-adulterants and adulterants of different categories,partial least squares(PLS)was used to establish the quantitative correction models of adulteration quantity of different kinds of adulterants.Result:The recognition rates of qualitative analysis model of Fritillariae Cirrhosae Bulbus and its adulterants were 99.49%(calibration set)and 100.00%(validation set),respectively.In different adulterant models,the recognition rates of calibration set and validation set were 70.47%and 73.68%,respectively.Moreover,the correlation coefficients of validation set(R2 P)of the six quantitative models of adulteration ratio were 0.8402(Fritillariae Cirrhosae Bulbus adulterated with F.thunbergii dry bulbs),0.9602(Fritillariae Cirrhosae Bulbus adulterated with F.ussuriensis dry bulbs),0.7657(Fritillariae Cirrhosae Bulbus adulterated with F.pallidiflora dry bulbs),0.9025(Fritillariae Cirrhosae Bulbus adulterated with F.hupehensis dry bulbs),0.9574(Fritillariae Cirrhosae Bulbus adulterated with Bulbus Tulipae),0.9761(Fritillariae Cirrhosae Bulbus adulterated with flour),the root mean square error of prediction(RMSEP)were 10.9485,5.4639,13.2564,8.5492,5.6553,4.2356,respectively.The two qualitative models and six quantitative models showed good prediction performance.Conclusion:The portable NIR spectroscopy can be used to identify Fritillar
作者
范林宏
何林
谭超群
田瑶
张春玲
吴纯洁
黄永亮
FAN Lin-hong;HE Lin;TAN Chao-qun;TIAN Yao;ZHANG Chun-ling;WU Chun-jie;HUANG Yong-liang(College of Pharmacy,Chengdu University of Traditional Chinese Medicine(TCM),Chengdu 611137,China;National Key Laboratory of Fundamental Science on Synthetic Vision,College of Computer Science,Sichuan University,Chengdu 610065,China;Affiliated Hospital of Chengdu University of TCM,Chengdu 610072,China)
出处
《中国实验方剂学杂志》
CAS
CSCD
北大核心
2022年第3期131-138,共8页
Chinese Journal of Experimental Traditional Medical Formulae
基金
四川省重大科技专项(2018TZDZX0007)。
关键词
便携式近红外光谱仪
川贝母
真伪
定性分析
定量分析
线性判别分析
偏最小二乘法
portable near infrared spectrometer
Fritillariae Cirrhosae Bulbus
authenticity
qualitative analysis
quantitative analysis
linear discriminant analysis
partial least squares
