应用地基雷达和机载激光雷达数据反演落叶松冠层叶面积密度

Inverting the Leaf Area Density of Larix gmelinii Canopy Using Ground-Based LiDAR and Airborne LiDAR Data

为了实现落叶松冠层叶面积密度盲区互补的协同反演,以塞罕坝机械林场为研究区,根据地基雷达和机载激光雷达数据对落叶松林冠层点云信息进行提取;利用点云分割算法对点云信息进行枝叶分离,并对冠层点云进行体素建模,分析落叶松林最优体素和接触频率的相关性;采用体素模型冠层分析法(VCP)分别对机载雷达和地基雷达数据绘制冠层叶面积密度曲线,实现对区域林木叶面积密度的反演。结果表明:运用冠层分析法适用于估算落叶松的叶面积密度,机载雷达点云数据和地基雷达点云数据协同估算的相对误差最小(平均相对误差1.95%),机载雷达数据估算次之(平均相对误差5.09%),地基雷达数据估算误差最大(平均相对误差9.37%)。因此,机载和地基雷达数据协同反演可以提升落叶松林叶面积密度的估算精度。

In order to realize the collaborative inversion of the blind area complementary Larix gmelinii canopy leaf area density,Saihanba Mechanical Forest Farm was chosen as the research area.Ground-based radar and airborne LiDAR data were used to extract point cloud information of the L.gmelini forest canopy.A point cloud segmentation algorithm was utilized to separate branches and leaves in the point cloud information,and the canopy point cloud was voxelized.The correlation between the optimal voxels and contact frequency of L.gmelini forest was analyzed.The VCP method was employed to draw the leaf area density curve of the canopy using both airborne radar and ground-based radar data,achieving inversion of regional forest leaf area density.The results indicated that the VCP method is suitable for estimating the leaf area density of L.gmelini.The relative error of the collaborative estimation of airborne radar point cloud data and ground-based radar point cloud data was the smallest(average relative error is 1.95%),followed by the estimation of airborne radar data(average relative error is 5.09%),and the estimation error of ground-based radar data was the largest(average relative error is 9.37%).Therefore,collaborative inversion of airborne and ground-based radar data can improve the estimating accuracy of L.gmelini forest leaf area density.