近红外光谱快速测定植物纤维原料中甲氧基含量的研究

Study on Rapid Determination of the Methoxyl Group Content in Plant Fibrous Material with Near-Infrared Spectroscopy

采取偏最小二乘法(PLS)建立起近红外光谱法快速测量植物纤维原料的甲氧基含量的模型。分别使用小波变换处理前后的光谱数据建立了植物纤维原料的甲氧基含量的PLS法测量模型,建立的两模型的平方相关系数分别为0.986和0.995,并用6个样本进行了预测,预测标准偏差分别为0.75和0.71。结果表明,该模块具有较高的学习和预测精度,可以用于大量植物纤维原料样本的甲氧基含量的快速测定,这将为甲氧基含量的测量提供了新的理论和方法。

Rapid determination model of the methoxyl group content in plant fibrous material was established by the integration of near-infrared spectroscopy and partial least squares （PLS）. Two models were obtained by PLS with spectral data of non-pretreatment and wavelet transformation pretreatment. The R2 of the obtained models were 0. 986 and 0. 995 respectively. The standard error of prediction （SEP） were 0.75 and 0.71 respectively for 6 test samples. The results showed the models had satisfactory training and forecasting precision and could be used as a powerful, nondestructive analytical technique for rapid and high-efficient determination of methoxyl group content in plant fibrous materials.