近红外分析技术在稻米淀粉RVA谱特征值测定中的应用  

Determination of RVA profile parameters of rice by near infrared spectrometer

作　　者：刘义富[1] 卢义宣[2] 奎丽梅[2] 郭咏梅[2] 谭春艳[2] 刘小莉[2] 

机构地区：[1]云南农业大学农学与生物技术学院,云南昆明650201 [2]云南省农业科学院粮食作物研究所,云南昆明650205

出　　处：《西南农业学报》2007年第5期974-978,共5页Southwest China Journal of Agricultural Sciences

基　　金：农业部"948"项目(2006-G1);农业科技攻关项目(2006NG03)

摘　　要：采用近红外光谱分析技术和化学计量方法建立了稻米淀粉RVA谱3项(BDV、SBV、CSV)与稻米其它品质相关显著的特征值的近红外分析模型,并对模型进行了预测准确性评价。结果显示BDV、SBV和CSV的模型校正决定系数Rc2分别为0.9707,0.9966和0.9943,校正标准差RMSEE分别为4.12,2.41和1.72;内部交叉检验的决定系数Rcv2分别为0.942,0.9942和0.9913;标准差RMSECV分别为5.4,2.87和1.99。验证决定系数除DBV外均达到0.99以上,模型准确性较高,具有实用价值。Three RVA profile parameters of rice predicted mathematic models with the technique of near infrared reflectance spectroscopy（NIRS）were established.The result indicated that the models determination coefficients of calibration（Rc2）for BDV,SBV and CSV were 0.9707,0.9966 and 0.9943,respectively.The root mean square errors of calibration（RMSEE） were 4.12,2.41 and 1.72.The veracity of models was estimated by the determination coefficients（Rcv2） and the root mean square errors（RMSECV） of cross-validation.The Rcv2 of NIR models of brown rice RVA profile parameters were 0.9420,0.9942 and 0.9913,respectively.The RMSECV were 5.4,2.87 and 1.99.It indicated that the models were usefully for the Rcv2 of all the models were reached 99 percent except the BDV.

关 键 词：近红外光谱 RVA谱 内部交叉验证 外部验证 

分 类 号：S511.033[农业科学—作物学]