摘要
建立了基于FOSS近红外谷物分析仪快速测定小麦发芽率的模型,对光学处理和数学处理等因素对模型的影响进行讨论,并对模型进行内部和外部检验。结果表明最佳的建模参数为:光学处理采用标准正常化处理(SNV and Detrend),数学处理技术采用"2,1,1,1"。得到的定标方程的定标相关系数(RSQ)为0.9544,定标标准偏差(SEC)为0.3762,交叉验证相关系数(1-VR)为0.891 8,交叉检验标准偏差(SECV)为0.3802,模型验证相关系数(R)为0.938,预测标准偏差(SEP)为0.373 9。
The model of determining wheat germination rate was built based on near infrared spectroscopy （NIRS） and the FOSS system as the analyzer. The effect of the factors, such as the mathematics methods and optics treatment methods, on the model was studied. The calibrations of both the interior and the exte- rior of the model were also performed. The results showed that the best factors were SNV and Detrend for optics treatment method and＂2,1,1,1＂ for mathematics method. The coefficient of correlation（RSQ） was 0.954 4, the standard error of calibration （SEC） was 0. 376 2, the determination coefficient of cross - vali- dation（ 1 - VR） was 0. 891 8, the standard error of cross - validation （ SECV ） was 0. 380 2, regression squared （R） was 0. 938, square error of prediction （SEP） was 0.373 9.
出处
《粮油食品科技》
北大核心
2013年第6期73-75,共3页
Science and Technology of Cereals,Oils and Foods
基金
国家863课题(2012AA101705)
关键词
小麦
近红外光谱
发芽率测定模型
wheat
near infrared spectroscopy
wheat
germination rate model
