<![CDATA[Urban trees provide various ecosystem benefits, including carbon sequestration, urban heat island mitigation, and air quality improvement. Accurate tree inventory is important for sustainable urban tree management. This study presents individual tree segmentation from point cloud data acquired using a handheld Mobile Laser Scanner (MLS). Individual tree segmentation was performed using the TreeIso plugin in CloudCompare. Three parameter scenarios were evaluated on 52 manually digitized reference trees. The results show that Scenario 3, with the highest parameter values (K₁=20, K₂=40, λ₁=10, λ₂=40, ρzmax=2, w=2), achieved the best performance with a precision of 0.61, recall of 0.83, and F1-score of 0.70. All scenarios exhibited over-segmentation, attributed to high point cloud density and occlusion effects during scanning. The relationship between parameter values and accuracy was non-linear, higher parameter values did not consistently improve recall but proved effective in increasing precision. This study demonstrates that TreeIso can reduce processing time from 24 hours of manual segmentation to only a few minutes, providing practical guidance for the application of MLS point cloud data in urban tree inventory.]]>
Copyrights © 2026
