基于不同预处理的椴树木材基本密度近红外光谱估测及模型优化

【目的】木材的基本密度在木材质量等级评定中起着重要的作用,是木材分流及精细化利用的重要依据。利用近红外光谱技术,实时监测木材性质,掌握木材性质的变化,为进一步制定和改善林木培育方法提供理论依据。【方法】借助树木生长锥对椴树活立木取样,以椴树样品基本密度真值和近红外光谱数据为输入,分别通过卷积平滑、一阶导数和二阶导数预处理方法来实现近红外光谱数据的预处理,建立了基于偏最小二乘法(PLS)的椴树木材基本密度的近红外估测模型。【结果】在350~2 500 nm波段范围内,一阶导数预处理的椴树木材基本密度模型是最优的,校正集相关系数为0.964 8,校正均方根误差为0.002 7,验证集相关系数为0.943 2,预测均方根误差为0.003 3。在对近红外光谱数据进行去噪优化处理,构建椴树木材基本密度模型后,在500~2 300 nm波段范围内,一阶导数预处理椴树木材基本密度模型依旧最优,其校正集相关系数为0.987 1,校正均方根误差为0.001 6,验证集的相关系数是0.948 6,预测的均方根误差是0.002 1。【结论】选择特定的预处理方法,结合样本特征,建立椴树木材基本密度模型,可以显著降低建模成本,提高模型预测精度,快速测定椴树木材的基本密度。

Near infrared spectral estimation and model optimization of Tilia tuan based on different pretreatments

【Objective】The basic density of wood plays an important role in the quality assessment of wood and is an important basis for the diversion and refinement of wood. The use of near-infrared spectroscopy technology to monitor the properties of wood in real time, to grasp the changes in the properties of wood, to provide a theoretical basis for further development or improvement of forest cultivation methods.【Method】The tree growth cone was used to sample the eucalyptus standing tree, and the basic density true value and near-infrared spectral data of the eucalyptus sample were input, and the near-infrared spectroscopy was realized by convolution smoothing, first derivative and second derivative preprocessing respectively. A near-infrared estimation model based on partial least squares(PLS) for the basic density of Tilia tuan was established.【Result】The basic density model of Tilia tuan pretreated by first derivative is optimal in the range of 350-2 500 nm, the correlation coefficient of calibration set is 0.964 8, the corrected root mean square error is 0.002 7, and the correlation coefficient of verification set is 0.943 2. The predicted root mean square error is 0.003 3. After denoising and optimizing the near-infrared spectral data and constructing the basic density model of Tilia tuan, the basic density model of the first-order derivative pre-processed Tilia tuan is still optimal in the range of 500-2 300 nm, and the correlation coefficient of the calibration set is for 0.987 1, the corrected root mean square error is 0.001 6, the correlation coefficient of the verification set is 0.948 6, and the predicted root mean square error is 0.002 1.【Conclusion】The selection of specific pretreatment methods combined with sample characteristics, the establishment of Tilia tuan basic density model can significantly reduce the modeling cost, improve the model prediction accuracy, and quickly determine the basic density of Tilia tuan.