Digital Forest Inventory Using Fused UAV and PLS Point Cloud Data

Abstract

Accurate digital forest inventory (DFI) is essential for sustainable forest management, yet single-sensor LiDAR approaches often fall short in capturing the full vertical structure of forest stands. This study evaluates the performance of a fused dataset combining UAV-based LiDAR (ULS) and personal laser scanner (PLS) data to overcome platform-specific limitations. A mixed forest stand in northeastern Germany was surveyed using both ULS and PLS under consistent scanning patterns, supported by ground control points for georeferencing. Point cloud fusion was achieved through a dynamic marker-based alignment and refined using partial iterative closest point (ICP) registration. The fused dataset, processed in TreeLS, enabled detailed stem reconstruction and vertical canopy characterization. Comparative analysis against field-measured tree metrics revealed a mean deviation of -2.5% for diameter at breast height (DBH) and +4.9% for tree height, with RMSE values of 2.9 cm and 3.84 m, respectively. These results highlight the complementary strengths of ULS and PLS platforms, demonstrating that their integration significantly enhances the accuracy, completeness, and efficiency of forest inventories. The presented workflow supports scalable, repeatable, and ecologically informative forest assessments, offering substantial potential for precision forestry and long-term monitoring applications.