应用NIR及主成分回归预测落叶松密度的研究

Study on Modeling Larch Density by NIR and Principle Component

运用近红外光谱主成分回归法对落叶松样品密度进行研究,校正集的相关系数(R)为0.86,校正集标准误差(SEC)为0.01,预测集的相关系数(R)为0.89,预测集标准误差(SEP)为0.02,对未参与建模的12个未知样品进行密度预测,相关系数达0.95。研究表明,近红外光谱能够快速、准确地对落叶松样品密度进行预测,这为快速检测落叶松木材材性提供了一种新方法。

Using Principle Component Regression （PCR）, the Near -Infrared （NIR） spectroscopy from Larch samples were analyzed for the modeling of density. The correlation coefficient of calibration set （R） was 0. 86 and the standard error of calibration （SEC） was 0. 01. The model was used for the density prediction of unknown samples with the correlation coefficient of prediction set （R） and the standard error of prediction （SEP） of 0. 89 and O. 02, respectively. Study showed that the density of Larch sample could be predicted accurately using NIR. This study provides a way for density fast prediction for Larch wood.