研究目的
To quantify 3D structure and occlusion in dense tropical and temperate forests using close-range LiDAR, specifically addressing how occlusion is distributed and how it affects forest reconstruction and modeling.
研究成果
Combining ground-based TLS with above-canopy measurements (UAVLS or canopy crane) allows for almost complete coverage of dense forest canopies with minimal occlusion (<2%), enabling high-resolution 3D forest modeling. This approach is recommended for applications requiring full canopy representation, such as radiative transfer modeling and ecological studies, while ground-only measurements suffice for stem reconstruction without top canopy details.
研究不足
The scan pattern in the temperate forest may have limited reduction of occlusion on top of canopy; additional scans outside the plot could have helped. UAVLS measurements had limited scan angles and may not be suitable for complete canopy reconstruction without ground complement. The method assumes random distribution of scatterers in voxels, which may not hold in all forest types. Computational demands for voxel traversal are high.
1：Experimental Design and Method Selection:
The study used a ray tracing approach with a voxel-traversal algorithm to trace laser pulses and determine sampled and occluded voxels in 3D space. This method was implemented to assess spatial coverage and occlusion patterns from ground-based TLS and above-canopy measurements (UAVLS or canopy crane TLS).
2：Sample Selection and Data Sources:
Data were collected from two dense forest sites: a tropical forest in Lambir Hills National Park, Malaysia, and a temperate forest in Laegern, Switzerland. Both sites had canopies of 30-50 m height, with specific plot sizes (60x60 m) selected for analysis.
3：List of Experimental Equipment and Materials:
Equipment included Riegl VZ-1000 and VUX-1UAV laser scanners, OxTS xNAV550 IMU/GPS, Aeroscout Scout B1-100 UAV helicopter, reflective cylindrical reference targets, and software such as Riscan Pro, RiPROCESS, and AMAPVox for data processing.
4：Experimental Procedures and Operational Workflow:
For ground measurements, TLS was performed with a dense scan pattern (10-15 m between scans), using reference targets for co-registration. Above-canopy measurements involved TLS from a canopy crane at multiple heights or UAVLS flights at 40-50 m above canopy. Point clouds were co-registered and processed to create voxel grids at 10 cm and 25 cm resolutions for occlusion mapping and plant area density estimation.
5：Data Analysis Methods:
Data were analyzed using voxel traversal algorithms to quantify occlusion, with plant area density profiles derived using AMAPVox. Statistical analysis included root mean squared error calculations to compare measurements from different setups.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
