摘要
利用微型傅立叶光谱仪对高粱秆榨汁液化学模型进行糖分的定量分析,发现间隔-偏最小二乘算法的回归模型具有较高的拟合度。实验利用了中红外光谱仪器和高效液相色谱法检测出高粱杆榨汁液的糖分类型和含量,再根据其糖分特征制作了3种混合糖溶液模型。基于微型傅立叶光谱仪的1150 nm^2050 nm近红外线波段光谱曲线,比较了偏最小二乘法和间隔-偏最小二乘算法定量回归算法,结果发现对于多组分糖溶液间隔-偏最小二乘算法具有较好的拟合效果,其中蔗糖、果糖、葡萄糖回归系数分别为0.8583、0.7820和0.8625。
The article makes a quantitative analysis of the sugar of chemical model of sorghum stalk squeezed juice by using MEMS-FTIR,finds that the results using the i PLS regression model with high fitness degree. While analyzing the sugars content of sorghum stalk by using MIR and HPLC measurement,experiment produced three kinds of mixed sugar solution model based on the content of sugar. Using the MEMS-FT-NIR from 1150 nm to 2050 nm band of NIR spectrum,compared the result with the PLS model and the i PLS model,and found that the i PLS model of sugar solution is more efficient,the regression coefficient of sucrose,fructose,and glucose is 0.8583,0.7820 and 0.8625 respectively.
出处
《成都信息工程大学学报》
2016年第4期367-371,共5页
Journal of Chengdu University of Information Technology
基金
农业部948计划资助项目(2015-Z44)
