Fig 1.
Artificial rubber plantation and rubber latex.
(a) Artificial Rubber Plantation; (b) Rubber Latex.
Fig 2.
Overview map of the research area.
Fig 3.
Schematic diagram of the working principle of UAV-borne 3D laser scanning technology.
Table 1.
BB4 Quadcopter UAV with AU20 Lidar technical parameter sheet.
Fig 4.
(a) BB4 quadcopter UAV; (b) AU20 LiDAR.
Fig 5.
3D point clouds of the initial and experimental areas of the artificial rubber forest.
(a) top view of the initial point cloud; (b) top view of the point cloud of the experimental area; (c) iso-metric map in front of the initial point cloud; (d) isometric map in front of the point cloud of the experimental area.
Fig 6.
Flowchart of point cloud data processing for artificial rubber forest.
Fig 7.
Filter denoising based on Gaussian filter point cloud.
(a) Number of point clouds after filter denoising; (b) Detail comparison after filter denoising.
Fig 8.
Post-classification map of point cloud ground points.
(a) point cloud classification results in the experimental area; (b) point cloud of ground points.
Fig 9.
Shows the processing results of three models.
(a) DEM; (b) DSM; (c) CHM.
Fig 10.
Flowchart of CHM based single-tree segmentation.
Fig 11.
Flowchart of the point cloud based single-tree segmentation.
Table 2.
Segmentation accuracy statistics of the three segmentation methods.
Table 3.
Information table of rubber tree parameters extracted based on Lidar360 processing software.
Table 4.
The accuracy analysis table of rubber tree height and average crown diameter.
Table 5.
The accuracy analysis table of the north-south and east-west crown diameters of rubber trees.
Table 6.
Expression of linear regression model.
Table 7.
Actual measured parameters of DBH, north-south crown diameter, east-west crown diameter, and average crown diameter.
Table 8.
Expression of linear regression model.
Fig 12.
Plots of the results of the three segmentation methods.
(a) CHM based single tree seg-mentation; (b) direct point cloud based single tree segmentation; (c) Seed point-based single tree segmentation.
Fig 13.
Single tree segmentation based on deep learning.
(a) training samples; (b) Segmentation results; (c) Segmentation problem.
Fig 14.
The process of the sampling plot survey method.
Fig 15.
Shows the fitting curves of various regression models.
(a) the fitting curve of the DBH regression model estimated by E-W CD; (b) The fitting curve of the DBH regression model estimated by N-S CD; (c) A CD estimates the fitting curve of the DBH regression model.
Table 9.
Estimated DBH values using E-W CD and DBH regression models.
Table 10.
Estimated DBH values using N-S CD and DBH regression models.
Table 11.
Estimated DBH values using A CD and DBH regression models.
Table 12.
Maximum, minimum, mean error, and root mean square error of DBH estimation regression model for each parameter.
Table 13.
The rubber allometric growth equation for above-ground and below-ground biomass.
Table 14.
Calculation table for carbon stocks in rubber trees.