近红外光谱法快速测定烟草中淀粉含量

[目的]快速测定烟草中淀粉的含量。[方法]采用近红外(NIR)漫反射光谱仪扫描140个烤烟烟叶样品,连续流动法测定其淀粉的含量,多种参数方法对谱图进行处理,最后选择消除常数法对光谱进行预处理和偏最小二乘法对数据进行拟合,建立了烤烟中淀粉的NIR预测模型,并对这些模型进行了外部验证。[结果]淀粉的预测值与连续流动法测定值的平均相对偏差均在5%以内,精密度RSD均小于5%。[结论]该模型可快速测定烟叶中淀粉的含量,对控制烟草调制过程中淀粉的转化和烟草制品中的淀粉含量具有重要意义。

Rapid Determination of Starch content in Tobacco by Near Infrared Spectroscopy

Objective] To rapidly detect the starch content in tobacco.[Method] Near infrared (NIR) diffuse reflectance spectrometer was used to scan 140 samples of flue-cured tobacco leaves.Starch content was detected by continuous flow method .Spectrogram was processed by various parameters method.Finally,constant elimination method was used for pretreatment of spectrum ,and partial least squares method was used to fit the data.The NIR prediction model of starch in flue-cured tobacco was established.And external validation was carried out for these models.[Result] The average relative deviations between the prediction value of starch and the detection value by continuous flow method were less than 5% with RSD smaller than 5%.[Conclusion] The model is applicable to rapid detect the starch in tobacco samples,which is of important practical signifi-cance to control the starch transformation in tobacco curing process and the starch content in tobacco products .