二组分食用调和油组成的近红外光谱检测定量分析  

Quantitative Determination of Components in Binary Blended Edible Oil

作　　者：陈华才[1] 王志岚[1] 刘福莉[1] 

机构地区：[1]中国计量学院,浙江省现代计量测试技术与仪器重点实验室,杭州310018

出　　处：《中国粮油学报》2008年第5期180-182,共3页Journal of the Chinese Cereals and Oils Association

基　　金：国家高科技发展计划“863”项目(2007AA10Z216)

摘　　要：分别取花生油、大豆油、玉米油和葵花籽油调和成不同比例的二组分食用调和油样品，采集样品在4000cm^-1～10000cm^-1范围内的近红外漫反射光谱，光谱经一阶导数等预处理后，采用偏最小二乘法建立二纽份调和油中各纽分含量的定标模型，并用Leave—one—out内部交叉验证法对模型进行验证。四种食用油定标模型的相关系数R分别为0．9996，0．9931，0．9985和0．9988，校正均方根RMSEC分别为0．83％，3．81％，1．76％和1．57％。采用不同品牌的玉米油和大豆油配制成二组分调和油样品，建立的PLS定标模型其R和RMSEC分别为0．9985和1．42％。研究结果表明，采用近红外光谱法可以实现二组分调和食用油中不同成分含量的准确定量检测，检测结果不受食用油生产厂家的影响。Peanut oil, soybean oil, corn oil and sunflower oil were used to prepare binary blended edible oil samples. Near infrared diffuse reflectance spectra of the samples were collected over 4 000 cm^- 1 - 10 000 cm^- 1 region. The calibration models for predicting the contents of peanut oil, soybean oil, corn oil, and sunflower oil in binary blended oil were established with PLS after the spectra were pretreated with the first derivative and other methods. Results show： The correlation coefficients （R） of the four models are 0. 999 6, 0. 993 1,0. 9985, 0. 998 8, respectively. The root mean square errors of leave - one - out cross validation （RMSEC） of them are 0.83 % ,3.81%, 1.76 % and 1.57 %, respectively. The R and RMSEC of the model built with samples prepared by using corn oil and soybean oil from different producers are 0.998 5 and 1 - 42 %, respectively. Therefore it is feasible to quantitatively determine the components in binary blended edible oil with the near infrared diffusive spectroscopy method.

关 键 词：食用调和油 近红外漫反射光谱 偏最小二乘法 定量分析 定标模型 

分 类 号：TS225.19[轻工技术与工程—粮食、油脂及植物蛋白工程] S816.43[轻工技术与工程—食品科学与工程]