以300个烟草样品的近红外光谱为基础,采用比值法和多元二次回归相结合的方法,建立了近红外光谱与烟草六组分(尼古丁、总糖、还原糖、总氮、钾、氯)含量间的一种测量模型,并采用交叉验证和样品实际测量的方法对模型的预测能力进行测试,结果表明该模型具有良好的预测能力,是一种快速高效检测烟草品质的方法,且具有测定波长少、计算简单、实现方便等特点,易于推广和产品化.%Based on the near - infrared spectrograph of 300 samples of tobacco, combining the ratio method and multiple quadric regression method, we build a model which reveals the relationship between the near - infrared spectrograph with the contents of six fractions of tobacco. The model is tested with cross validation and actual measurements, the results show that the model gives good prediction. It offers an efficient method for fast detection of tobacco quality. Since it needs only several measurements, and the calculation is simple and convenient to realized, this method is easy to popularized and commercialized.
展开▼