A Crown Morphology-Based Approach to Individual Tree Detection in Subtropical Mixed Broadleaf Urban Forests Using UAV LiDAR Data

Owing to the complex forest structure and large variation in crown size, individual tree detection subtropical mixed broadleaf forests urban scenes is a great challenge. Unmanned aerial vehicle (UAV) light ranging (LiDAR) powerful tool for due its ability acquire high density point cloud that can reveal three-dimensional structure. Tree based on local maximum (LM) filter, which applied canopy height model (CHM) generated from LiDAR data, popular method simplicity. However, it difficult determine optimal LM filter window size prior knowledge usually needed estimate size. In this paper, novel approach morphology information proposed. approach, LMs are firstly extracted using whose determined by minimum then identified Gi* statistics out caused surface irregularities contained CHM. The retained final results represent treetops. was two test sites characterized different structures UAV data. sensitivity of parameter setting analyzed rules were On first site irregular distribution rate F-score derived combination values 72.9% 73.7%, respectively. second regular relatively small 87.2% 93.2%, comparison with variable-size algorithm, both rates proposed higher.