美引新品种烟草常规化学成分NIR预测模型的建立  

Establishment of NIR predictive model of routine chemicals components in flue-cured tobacco varieties introduced from the USA

作　　者：蔡伟[1] 和智君[1] 简彬[1] 向明[1] 任一鹏[1] 田丽梅[1] 张天剑[1] 

机构地区：[1]云南烟草科学研究院,云南昆明650106

出　　处：《云南大学学报（自然科学版）》2010年第S1期51-56,共6页Journal of Yunnan University(Natural Sciences Edition)

基　　金：云南中烟工业公司科技资助项目(2006YL01-2)

摘　　要：采集4个美引品种NC102,NC297,NC55,GL350共3a初烤烟叶样品205份,按样品不同地区、不同品种将样品分为180份校正集和25份验证集.通过行业标准方法检测烟叶中总糖、还原糖、总氮、烟碱、蛋白质、钾含量.利用近红外(NlR)光谱建立近红外预测模型.通过验证,以上6项化学成分预测模型的相关系数(R2)分别为96.95%,94.43%,96.1%,98.1%,95.3%,94.0%.均方差(RMSECV)分别为0.895,0.876,0.06698,0.0743,0.397,0.0717.平均相对偏差分别为2.97%,3.13%,5.12%,3.47%,6.03%,5.28%.结果表明该方法与标准方法测定的结果无显著性差异,证明该方法可靠、快捷.Two hundreds and five curing tobacco leaf samples of four American breeds:NC102,NC297,NC55,GL350 were collected and were divided two sets:the calibration set and the validation set according to different species and area.Total sugar,reducing sugar,total nitrogen,nicotine,protein and potassium contents were detected by tobacco industry standard methods.NIR predictive model of chemical components mentioned was established.Through outer validation,the correlation coefficient(R2) was 96.95%,94.43%,96.10%,98.10%,95.3% and 94.0% respectively.The mean square deviation(RMSECV) was 0.895,0.876,0.066 98,0.074 3,0.397,0.071 7 respectively,and the average relative deviation was 2.97%,3.13%,5.12%,3.47%,6.03% and 5.28% respectively.The results indicated that this methods has non-distinguished differences compare to standard method and is quickly and reliably.

关 键 词：近红外光谱预测模型 烟草 美引新品种 化学成分 

分 类 号：TS424[农业科学—烟草工业]