小麦子粒湿面筋FT-NIRS分析模型的建立与研究

Research on FT-NIRS in Determination of Wet Gluten of Wheat Seed

用傅立叶变换近红外光谱(FT-NIRS)分析小麦子粒湿面筋含量。采用国标GB/T14608-1993法测试了216份小麦子粒的湿面筋含量,用近红外仪采集数据,选择113份建立了数学模型。结果:最佳主成分数(Rank)=3,内部交叉验证均方差(RMSECV)=2.33,决定系数(R2)=90.69。为了验证模型的可靠性,对预测集样品进行预测,结果湿面筋的预测均方差(RMSEP)=2.70;相对偏差(RSEP,%)=11.84。结果表明,该数学模型可初步用来快速、较准确、无污染、低消耗地测试小麦子粒的湿面筋含量。

In this study,gluten contents of 216 genotypes of Triticum aestivum were analyzed by GB/T14608-1993 method. Bruker fourier transform near-infrared spectroscopy（FT-NIRS） used 113 genotypesto establish math model of gluten. The result on this model showed that the Rank was 3 and the RM-SECV was 2.32 and R2 was 90.69. 35 samples were used to test this model. The result on model valida-tion showed that the RMSEP was 2.70 and the RSEP（%） was 11.84.So FT-NIR technique might be usedto determined accurately and fleetly gluten content with free pollution and depletion in short time.