论文部分内容阅读
面粉颗粒度是影响小麦食品加工品质的重要性状。以2 0 0 1- 2 0 0 2年度来自北部冬麦区、黄淮冬麦区、长江中下游冬麦区和西南冬麦区的2 5 6份小麦品种(系)为材料,用激光散射颗粒度分析仪和近红外透射光谱技术对面粉颗粒度进行了研究。结果表明,我国小麦面粉颗粒度分布特点为从北向南,硬质麦分布比例逐渐减少,软质麦分布比例逐渐增大。总体上硬质麦所占比例较高,为5 9. 4 % ,软质麦和混合麦所占比例分别为2 8 .1%和12 .5 %。近红外预测集模型决定系数为0 92 ,与激光散射颗粒度分析仪测试结果之间呈极显著正相关。在建立模型时,采用预测残差平方和及交叉验证处理方法进行了主成分数目选择,并对定标集样品中浓度超常样品和光谱超常样品做2次删除,删除样品数占总样品数的4 .1% ,使定标模型决定系数RSQ从0 . 82提高到0. 92 ,标准误差SEC从12 . 75下降到8 .5 4 ,提高了模型的预测精度,优化了定标模型。此外,根据我国小麦颗粒度大小分布的特点,建立了各麦区的定标模型,其可靠性优于以总参试样品建立的模型。
Flour grain size is an important trait that affects wheat food processing quality. A total of 256 wheat cultivars (lines) from winter wheat, northern Huanghuai winter wheat, middle and lower reaches of the Yangtze River and winter wheat from the southwest winter wheat were selected as materials from 2000 to 2002 by using laser scattering particle size Analyzer and near-infrared transmission spectroscopy of flour particle size were studied. The results showed that the grain size distribution of wheat flour in our country was from north to south, the distribution of hard wheat decreased gradually, and the distribution of soft wheat increased gradually. Hard corn accounted for a high proportion of 5.92%, while soft wheat and mixed wheat accounted for 28.1% and 12.5% respectively. The coefficient of determination of near-infrared prediction set model is 0 92, which is significantly and positively correlated with the laser scattering particle size analyzer test results. At the time of establishing the model, the number of principal components was selected by square sum of prediction residuals and cross-validation processing methods. The concentration of hypersensitivity samples and spectral aberration samples in calibration sample were deleted twice, and the total number of samples 4. 1%, so that the calibration coefficient RSQ increased from 0.82 to 0.92, the standard error SEC decreased from 12.75 to 8.54, which improved the prediction accuracy of the model and optimized the calibration model. In addition, according to the characteristics of grain size distribution of wheat in our country, a calibration model of each grain area is established, which is more reliable than the model established by the total sample.