Treetop Detection in Mountainous Forests Using UAV Terrain Awareness Function

Unmanned aerial vehicles (UAVs) are becoming essential tools for surveying and monitoring forest ecosystems. However, most forests found on steep slopes, where capturing individual tree characteristics might be compromised by the difference in ground sampling distance (GSD) between slopes. Thus, we tested performance of treetop detection using two algorithms canopy height models (CHMs) obtained with a commercial UAV (Mavic 2 Pro) terrain awareness function (TAF). The area surveyed was slope covered predominantly fir (Abies mariesii) trees, flown following (TAF) not (NTAF). Results showed that when TAF used, trees were clearly delimited, lower branches visible orthomosaic, regardless position. As result, dense point clouds (DPCs) denser more homogenously distributed along than NTAF. Two applied detection: (connected components), (morphological operators). components) 5% improvement accuracy (86.55%), comparison to NTAF (81.55%), at minimum matching error 1 m. In contrast, operators), reached 76.23% 62.06% alone, can sufficient sophisticated algorithms. higher number repeated points, leading an overestimation detected treetop.