基于机载LiDAR数据,分析哑变量对林分蓄积量估测精度的影响。以广西高峰林场为研究对象,借助机载激光雷达点云数据和96个样地数据,将样地数据按7∶3的比例随机划分为建模样本和测试样本,采用随机森林模型(RFR)和支持向量机模型(SVR)对建模样本与对应的点云特征回归建模,将树种组(针叶林和阔叶林)和龄组分别作为哑变量引入到回归模型。利用测试样本的估测精度评价模型的估测精度,引入树种组哑变量,随机森林模型决定系数R²从0.59提高到0.64,支持向量机模型决定系数R²从0.49提高到0.50。引入龄组哑变量,随机森林模型决定系数R²从0.59提高到0.65,支持向量机模型决定系数R²从0.45提高到0.55。根据模型的建模精度和验证精度结果得出,引入哑变量对蓄积量估测模型的精度提升是相对有效的。龄组哑变量对模型精度提升效果优于树种组哑变量。

Based on the airborne LiDAR data,the influence of dummy variables on the estimation accuracy of stand volume is analyzed.In this study,Guangxi Gaofeng Forest Farm is taken as the research object,and based on the airborne lidar point cloud data and 96 plot data,the plot data is randomly divided into training samples and test samples are at a ratio of 7∶3.The samples and the corresponding point cloud features are estimated by regression modeling through the use of random forest model(RFR) and support vector machine model(SVR),and the tree species group(coniferous forest and broad-leaved forest) and age group as dummy variables are introduced into the regression model.The study uses the estimation accuracy of the test samples to evaluate the estimation accuracy of the model,and introduces tree species group as dummy variables.The RFR determination coefficient R² increases from 0.59 to 0.64,and the SVR determination coefficient R ² increases from 0.49 to 0.50.With the introduction of dummy variables in the age group,the RFR determination coefficient R ² increases from 0.59 to 0.65 and the SVR determination coefficient R ² increases from 0.45 to 0.55.According to the modeling accuracy and verification accuracy results of the model,the introduction of dummy variables is relatively effective in improving the accuracy of the accumulation estimation model.The dummy variables of age group have better effect on model accuracy improvement than dummy variables of tree species group.