Application of 3D tree modeling using point cloud data by terrestrial laser scanner

Constructing 3D models for trees such as those found in Japanese gardens, which many species exist, requires the generation of tree shapes that combine characteristics tree's and natural diversity. Therefore, this study proposes a method constructing model with highly-accurate shape reproducibility from point cloud data acquired by TLS. As method, we attempted to construct using TreeQSM, is open source TLS-QSM method. However, since processing based on assumption consists related trunks branches, measuring leaves have fallen recommended. To solve problem, proposed an efficient classification process mainly uses thresholds deviation reflectance, are adjunct object can be laser measurement. Furthermore, verify accuracy model, position coordinates constructed were extracted. The extracted compared clarify extent was data. result, within standard 0.016 m also effective terms accuracy.