2026 Volume 79 Pages 45-49
We integrated Airborne Laser Scanning-derived Digital Elevation Model (DEM), Unmanned Aerial Vehicle (UAV)-based Structure-from-Motion (SfM), and Terrestrial Laser Scanning (TLS) point clouds to reconstruct the three-dimensional structure of a Japanese cedar (Cryptomeria japonica) plantation minimizing vertical gaps. Ground alignment between the DEM and UAV-SfM point clouds was solved using least-squares optimization with ground control points. Subsequently, the Iterative Closest Point registration algorithm was used to align the TLS data with the resulting DEM-SfM composite. Point density was homogenized via voxel grid resampling. Accuracy was assessed using RMS for (i) ground surface alignment relative to the DEM and (ii) aboveground features based on cloud-to-cloud nearest-neighbor distances. Ground RMS prior to alignment was 30.29 m, decreasing to 0.66 m after registration. Aboveground RMS decreased from 1.92 m to 1.14 m (Stand 1) and from 2.06 m to 0.57 m (Stand 2). Vertical distribution histograms revealed complementary coverage, with UAV-SfM capturing the upper canopy and TLS capturing stems and understory, thereby bridging the middle-story gap and yielding continuous coverage from terrain to canopy tops. These results demonstrate that the proposed integration approach provides robust baseline data for quantitative analysis of forest vertical structure and supports applications like silvicultural effect assessment and biomass estimation.
