A dataset of UAV-LiDAR and MLS collected in three different forest types from Southern China

The region has an annual mean temperature of 21.7℃ and receives 1,624–1,890 mm of precipitation annually.2)       The Gaofeng study site (GF) is a 900 m2 (30 m´30 m) rectangular plot located in the state-owned Gaofeng Forest Farm in the northern Nanning City, Guangxi Zhuang Autonomous Region, China (22.9960°N, 108.3734°E). The forest was a 6-year-old second-rotation coppice plantation of Eucalyptus urophylla, characterized by a sparse understory and numerous dead branches on tree trunks.

Canopy closure, derived from the ULS data, is 0.83. The terrain is mountainous, with an elevation of 327–344 m. The region has an average annual temperature of 21.6℃, and an average annual precipitation of 1,300 mm.3)       The Shangsi study site (SS) can be found in the Shiwandashan National Nature Reserve in Shangsi County, Guangxi Zhuang Autonomous Region, China.

The polygonal area of 770 m2 (21.9117°N, 107.8989°E) covers the ecotone between the northern tropics and southern subtropics. The plot is covered with mixed coniferous- broadleaved forests, with the main species being Pinus massoniana, Castanopsis hystrix, Schima argentea, and Mytilaria laosensis. The understory contains scattered shrubs, bamboos, and climbing vines.

The canopy closure is 0.94. It is noteworthy that all Pinus massoniana individuals within the plot have undergone resin tapping. The plot’s terrain is quite flat, with an elevation range of 263–265 m, despite being situated in a generally mountainous region.

The climate has a mean annual temperature of 21.7°C and a mean annual rainfall of 1,217 mm.2. Field data Field measurement data were collected independently at the GZ, GF, and SS study sites on Aug. 4, 2024, Aug. 28, 2024, and Oct. 27, 2025. DBHs (at 1.3 m above ground level) of all live trees were measured with a diameter tape; trees with DBHs of less than 5 cm were not recorded.

Their trunk base centers were surveyed using a network of real-time kinematic global navigation satellite systems (RTK-GNSS). Within the ArcGIS environment, the tree distribution map was overlaid with the trunk point clouds (ranging from 1.0 m to 1.6 m). After that, manual adjustments were made to the tree positions until they covered the centers of the point cloud clusters, targeting a positioning error of no more than 5 cm.

The outcome of field measurement included six files:l Tree_location_GZ.shp: An ArcGISS shp file showing the tree distributions in the GZ study site.l Tree_location_GF.shp: An ArcGISS shp file showing the tree distributions in the GF study site.l Tree_location_SS.shp: An ArcGISS shp file showing the tree distributions in the SS study site.The projection system for all of the aforementioned tree distribution files is WGS_1984_World_Mercator.l DBH_field_GZ.xlsx: Records of measurements of diameter at breast height in the GZ study site.l DBH_field_GF.xlsx: Records of measurements of diameter at breast height in the GF study site.l DBH_field_GZ.xlsx: Records of measurements of diameter at breast height in the SS study site.3.

LiDAR data Three MLS systems and three ULS systems were deployed to acquire eight datasets of laser point clouds across the three study areas. These devices and their manufacturers are listed below: l The NavVis VLX2 wearable mobile mapping system (NavVis GmbH, Blutenburgstr., Germany).l The R6-16 backpack SLAM system (OmniSLAM Technology Co. Ltd., Beijing, China, oslamtec.com/R6.html.), l The handheld SLAM100 laser scanner (Shenzhen Feima Robotics Co.

Ltd., Shenzhen, China, feimarobotics.com/en/productDetailSlam100.). l The Fly2Map Done SF-1200 UAS integrated with an AcuteLas SA130 laser scanner (South Surveying & Mapping, Guangzhou, China), The URL for the equipment presentation is southinstrument.com/product/details/pro_tid/1/id/211.html and southinstrument.com/product/details/pro_tid/1/id/270.html.l The DJI Matrice 300 RTK UAS equipped with a Zenmuse L2 sensor (Shenzhen Dajiang Innovation Technology Co., Ltd., Shenzhen, China), The URL for the equipment presentation is enterprice.dji.com/cn//matrice-350-rtk?site=enterprise&from=nav and enterprice.dji.com/cn/zenmuse-l2?site=enterprise&from=nav.l The DJI Matrice 400 RTK UAS carrying a Zenmuse L3 sensor.

(Shenzhen Dajiang Innovation Technology Co., Ltd., Shenzhen, China). A total of eight MLS and ULS datasets were collected from three study sites. All eight datasets were systematically named using abbreviations representing the study sites and laser sensors.

For instance, GZ-NavVis and GF-SA130 refer to datasets collected with the NavVis VLX2 and SA130 sensors at the GZ and GF study sites, respectively. The UAV flight settings, as well as the point cloud density and data acquisition times for the datasets, are shown in Table 1. l GZ-NavVis: Point density (Including the entire point cloud; the same applies below): 64,686 pts/m2; Date of collection data: Aug. 4, 2024.l GF-R6: Point density: 212,388 pts/m2; Date of collection data: Sep. 5, 2024.l GF-SLAM: Point density: 12,020 pts/m2; Date of collection data: Sep. 5, 2024l GF-SA130: Flight altitude above ground level (AGL): 60 m; flight speed: 3 m/s; point density: 72,80 pts/m2; Date of collection data: Nov. 04, 2024.l GF-L2: Flight altitude AGL: 60 m; flight speed: 3 m/s; point density: 65,927 pts/m2; Date of collection data: Nov. 4, 2024.l SS-NavVis: point density: 77,846 pts/m2; Date of collection data: Oct. 27, 2025.l SS-SA130: Flight altitude AGL: 80 m; flight speed: 3 m/s; point density:104,588 pts/m2; Date of collection data: Oct. 27, 202. l SS-L3: Flight altitude AGL: 80 m; flight speed: 3 m/s; point density: 37,488 pts/m2; Date of collection data: Dec. 03, 2025.

The datasets were initially preprocessed using manufacturer-provided software. Point cloud classification, digital elevation model (DEM) generation, and elevation normalization were subsequently performed using Terrasolid (Terrasolid, Helsinki, Finland) and LiDAR360 version 5.2 software (GreenValley, Beijing, China). There were four datasets at the GF study site, and three at the SS study site.

Thus, these datasets required fine registrations. The GF-NavVis and SS-NavVis datasets were used as the reference datasets, respectively, and the remaining datasets were registered with the reference datasets using the iterative closest point (ICP) algorithm in CloudCompare (Version 2.13.2) (cloudcompare.org, accessed on Dec. 20, 2023). The ICP parameters were configured as follows: RMSE difference = 1.0 ´ 105; overlap = 100%; random sampling limit = 50,000; rotation = XYZ; and translation = TxTy).

All LiDAR data in this dataset consist solely of point clouds with elevations ranging from 0.2 to 2.4 meters, and all were projected in the WGS_World_Mercator coordinate system. The eight LiDAR point cloud files are listed bellow:l GZ_NavVis_20_240cm.lasl GF_R6_20_240cm.lasl GF_SLAM_20_240cm.lasl GF_SA130_20_240cm.lasl GF_L2_20_240cm.lasl SS_NavVis_20_240cm.lasl SS_SA130_20_240cm.lasl SS_L3_20_240cm.las.