摘要
以93份燕麦样品为研究对象,比较了不同光谱预处理方法和不同回归方法对定标模型的影响,建立了定量分析燕麦中蛋白质含量的合理模型。结果表明,最佳的预处理方法:光谱散射处理为标准化处理(SNV),数学处理为2441;最佳的回归方法为改进偏最小二乘法(MPLS)。在此条件下模型对验证集的测定值与预测值的决定系数为0.954 3,均方根误差为0.160 7,模型的预测准确性良好。建立的近红外模型对燕麦中蛋白质快速测定有一定参考价值。
93 oat samples collected from China were used as the study object to establish the calibration model of protein content in oat after discussing the influence of the different spectrum preprocessing and regression methods on the calibration model. The results showed that the spectrum scattering of SNV, mathematics derivative processing of 2441 was the optimum preprocessing method of calibration model, and modified partial least square （MPLS）regres- sion was the optimal regression method. The correlation coefficient of the true value and the prediction value was 0. 954 3, and the root mean square deviation was 0. 160 7, which showed that the calibration model had better predic- tive accuracy. The method provided a reference value to quickly test the protein content of oat.
出处
《中国粮油学报》
EI
CAS
CSCD
北大核心
2016年第8期138-142,共5页
Journal of the Chinese Cereals and Oils Association
基金
国家燕麦荞麦产业技术体系(CARS-08-D)
西安市科技计划项目现代农业推进计划[NC1207(1)]
关键词
近红外光谱技术
燕麦
蛋白质含量
光谱预处理
回归方法
near - infrared spectroscopy, oat, protein content, spectrum preprocessing, regression method
