摘要
测试了30个品种206株石斛茎的红外光谱,光谱显示,石斛的主要物质成分为纤维素等多糖物质。用石斛的傅里叶变换红外光谱结合Wilks’Lambda逐步判别分析法对石斛品种进行识别研究,比较分析了不同的光谱范围和不同的训练样本数对识别结果的影响。对1 800~1 250cm-1光谱范围的287个变量进行品种判别分析,每个品种的训练样本分别为2,3,4,5,6个时,对训练样本的回判正确率都是100%;以余下的样本作为测试样本进行品种预测的正确率分别为79.4%,91.3%,93.0%,98.2%,100%。同时对1 800~1 500,1 500~1 250,1 250~600,1 250~950,950~600cm-1等不同光谱范围,按五种不同的训练样本情况,相同的判别分析方法进行判别分析比较,1 800~1 250,1 800~1 500和950~600cm-1光谱范围的变量更适宜进行石斛品种的判别分析;每个品种的训练样本达到3个及以上时,建立的模型判别分析结果较好。结果表明,傅里叶变换红外光谱结合逐步判别分析法对不同品种的石斛进行鉴别,能够较好的识别石斛的品种,为石斛品种的鉴别提供了一种简便、易行的新方法。
The difference of Dendrobium varieties were analyzed by Fourier transform infrared (FTIR) spectroscopy .The infra-red spectra of 206 stems from 30 Dendrobium varieties were obtained ,and showed that polysaccharides ,especially fiber ,were the main components in Dendrobium plants .FTIR combined with Wilks’ Lambda stepwise discriminative analysis was used to identify Dendrobium varieties .The effects of spectral range and number of training samples on the discrimination results were al-so analysed .Two hundred eighty seven variables in the spectral range of 1 800~1 250 cm -1 were studied ,and showed that the return discrimination is 100% correct when the training samples number of each species was 2 ,3 ,4 ,5 ,and 6 ,respectively , whereas for the remaining samples the correct rates of identification were equal to 79.4% ,91.3% ,93.0% ,98.2% ,and 100% , respectively .The same discriminative analyses on five different training samples in the spectral range of 1 800~1 500 ,1 500~1 250 ,1 250~600 ,1 250~950 and 950~650 cm-1 were compared ,which showed that the variables in the range of 1 800~1 250 ,1 800~1 500 and 950~600 cm -1 were more suitable for variety identification ,and one can obtain the satisfactory result for discriminative analysis when the training sample is more than 3 .Our results indicate that FTIR combined with stepwise discrimi-native analysis is an effective way to distinguish different Dendrobium varieties .
出处
《光谱学与光谱分析》
SCIE
EI
CAS
CSCD
北大核心
2014年第11期2968-2972,共5页
Spectroscopy and Spectral Analysis
基金
国家自然科学基金项目(81260608)
农业部公益性行业科研专项项目(201303117)资助
关键词
红外光谱
判别分析
石斛
识别
Infrared spectroscopy
Discriminative analysis
Dendrobium
Identification.
