稻谷脂肪近红外光谱特征筛选及检测模型构建  

Establishment of a selection and detection model of fat in rice by near infrared spectrum characteristics

作　　者：李路[1] 黄汉英[1] 李毅 赵思明[2] 杨素仙[1] 

机构地区：[1]华中农业大学工学院,湖北武汉430070 [2]华中农业大学食品科技学院,湖北武汉430070

出　　处：《食品与发酵工业》2018年第2期87-91,共5页Food and Fermentation Industries

基　　金：中央高校基本科研业务费专项(2662015PY078);湖北省重大科技创新计划(2014ABC009)

摘　　要：应用近红外光谱技术对稻谷脂肪含量进行检测。采集了90个稻谷样本的漫反射近红外光谱,运用Kennard-Stone法选取校正集及预测集样本。对比研究了归一化、一阶导、二阶导、一阶导+归一化等4种预处理方法对模型性能的影响,确定一阶导为最佳预处理方法。运用竞争性自适应重加权采样技术筛选出与稻谷脂肪含量检测相关的特征波长,再用多元线性回归对特征波长进行优选,最终得到30个特征波长。其中最典型的特征波长为1 343、1 489和1 583 nm,反映了稻谷脂肪中大量存在的—CH和—OH基团。所建立的基于近红外光谱分析技术的稻谷脂肪含量检测模型具的决定系数为0.958 9,定标标准差RMSEC为0.223 6,相对偏差为5.53%。Near Infrared(NIR) spectrum was used to detect fat content in rice. NIR spectra of 90 rice samples were measured. Kennard-Stone method was used to select the calibration set and prediction set samples. The effects of different pretreatment(normalize,first derivative and second derivative methods) have been compared for the accuracy of the models. The best pretreatment method is the first derivative. The competitive self-adaptive weighted sampling technology is used to screen the key wavelengths associated with sample properties. Finally,thirty key wavelengths are selected by Multiple Linear Regression further. The most typical key wavelengths are 1 343 nm,1 489 nm and 1 583 nm which related to the groups of —CH and —OH in rice fat. The detection model of fat content of rice based on near infrared spectroscopy has higher precision with the coefficient of determination,root mean square error of calibration and relative deviation are 0. 958 9,0. 223 6 and 5. 53%,respectively.

关 键 词：近红外光谱 稻谷 脂肪 竞争性自适应重加权采样 多元线性回归 

分 类 号：O657.33[理学—分析化学]