Full
Outline
To investigate the applicability of unmanned aerial vehicle (UAV) LiDAR, handheld LiDAR, and their fused point cloud data for individual tree parameter extraction and aboveground biomass estimation, and to improve the accuracy of single-tree scale modeling in complex forest stands.
The study was conducted in five plots of Pinus kesiya at the Puwen experimental forest of the Yunnan Academy of Forestry and Grassland. UAV LiDAR point cloud data and handheld LiDAR point cloud data were collected separately as two independent data sources, followed by data fusion of the two datasets. The watershed algorithm was applied to segment individual trees from the UAV LiDAR point cloud, whereas a distance-based clustering algorithm was used for the handheld and fused point clouds. Then, a total of 71 variables were extracted from individual tree point clouds, including 4 structural features, 40 height-related variables, 10 density variables, and 17 voxel-based features. The Boruta algorithm and random forest feature importance ranking were used for variable selection, and key variables were then used to build random forest models for AGB estimation. The modeling accuracy of different data sources was compared.
1) The overall segmentation F-scores for UAV, handheld, and fused point cloud data were 0.90, 0.95 and 0.97, respectively; 2) The R2 values of tree height estimation from UAV, handheld, and fused data were 0.93, 0.94 and 0.94, respectively, with RMSEs of 1.43 m, 1.74 m, and 1.68 m. The fused data achieved a DBH estimation accuracy of R2=0.95 and RMSE is 2.13 cm; 3) For the three types of point cloud data, 71 variables were extracted and independently selected using the Boruta algorithm and Random forest. For each dataset, eight key variables were identified separately and used for single-tree aboveground biomass modeling; 4) The estimation accuracy of individual tree above-ground biomass based on UAV point cloud data was R2=0.83, RMSE is 85.69 kg/plant and rRMSE is 44.53%; based on handheld point cloud data, R2=0.87, RMSE is 72.02 kg/plant and rRMSE is 36.34%; and based on fused point cloud data, R2=0.86, RMSE is 70.65 kg/plant and rRMSE is 36.72%.
The individual tree biomass model constructed based on the random forest method demonstrated good predictive performance across different data sources, with the fused point cloud data from UAV and handheld LiDAR showing superior accuracy and stability.
21
4
| 科 Family | 属数 Number of genus | 种数 Number of species | 占总种数比例 Percentage of total species (%) | 属 Genus | 种数 Number of species | 占总种数比例 Percentage of total species (%) |
|---|---|---|---|---|---|---|
| 鹅膏菌科Amanitaceae | 2 | 11 | 5.26 | 鹅膏菌属 Amanita | 10 | 4.78 |
| 小菇科 Mycenaceae | 2 | 12 | 5.74 | 丝盖伞属 Inocybe | 5 | 2.39 |
| 多孔菌科 Polyporaceae | 8 | 14 | 6.70 | 蜡蘑属 Laccaria | 5 | 2.39 |
| 红菇科 Russulaceae | 3 | 23 | 11.00 | 小皮伞属 Marasmius | 6 | 2.87 |
| 小菇属 Mycena | 11 | 5.26 | ||||
| 光柄菇属 Pluteus | 5 | 2.39 | ||||
| 红菇属 Russula | 17 | 8.13 | ||||
| 栓菌属 Trametes | 5 | 2.39 |
关闭全屏
No. 86 Xueyuan South Road, Haidian District, Beijing
010-62662699(Operation and Technical Support)
qkjq@cast.org.cn
Copyright © 2025 China Association for Science and Technology. All rights reserved. For all open access content, the relevant licensing terms apply.
Sponsored by the Office of the Leading Group for Cybersecurity and Informatization of CAST, and supported by Science and Technology Review Publishing House
