拉曼光谱结合化学计量学方法测定蜂蜜的主要糖分  

Quantification of Main Sugar Content of Honey Using Raman Spectroscopy and Chemometrics

作　　者：苏东林[1,2,3] 张欣[2] 李高阳[3] 朱向荣[2,3] 刘伟[1,3] 张菊华[1] 谢秋涛[3] 单杨[1,2,3] 

机构地区：[1]湖南省农业科学院湖南省农产品加工研究所,长沙410125 [2]中南大学隆平分院,长沙410125 [3]湖南省食品测试分析中心,长沙410125

出　　处：《中国食品学报》2016年第11期263-272,共10页Journal of Chinese Institute Of Food Science and Technology

基　　金：湖湘青年培养对象项目(湘科人字〔2014〕76号);农业科研杰出人才及其创新团队(农办人〔2012〕64号);"十二五"国家科技支撑计划(2012BAK17B17);湖南省科技重大专项(2010FJ1009-2)

摘　　要：用便携式拉曼光谱仪测定蜂蜜中的主要糖分,研究不同光谱前处理与校正模型之间的关系(光谱范围175 cm^(-1)至2 600 cm^(-1))。结果表明,与最小偏二乘回归模型(PLS-RM)相比,支持向量机回归模型(SVM-RM)适用于蜂蜜中果糖、葡萄糖、蔗糖和麦芽糖含量的快速测定,其中果糖、葡萄糖和麦芽糖的预测值同参照值的决定系数(R2)分别为0.84,0.87和0.75,而蔗糖含量的定量精确性稍低。拉曼光谱可用于快速定量测定蜂蜜的主要糖分。Sugars in honey were investigated through portable Raman spectroscopy（spectral range between 175 cm^（-1）and 2 600 cm^（-1））. The performance of calibration models were also evaluated using different spectral pretreatments. An SVM（support vector machine） regression model was constructed to calibrate fructose, glucose, sucrose and maltose of honey, and this model was then compared with PLS（partial least squares） regression models. Fructose, glucose and maltose were accurately predicted as indicated by their coefficients of determination（R2） of 0.84, 0.87 and 0.75, respectively, between the predicted and reference values. However, the prediction accuracy for sucrose was little poor and unreliable. So, this research indicated that Raman spectroscopy can be used to rapidly determine the major sugar components of honey.

关 键 词：拉曼光谱 蜂蜜 化学计量学 主要糖分 定量 

分 类 号：S896.1[农业科学—特种经济动物饲养]