Field tree measurement to collect ground truth data is labor intensive. To make it efficient to collect the field data in a large scale, Unmanned Aerial Vehicle is introduced in this study. An automated photogrammetric technique called Structure from Motion (SfM) is used to create 3D data out of photos collected by UAV. Tree structure created by SfM is compared with the 3D data obtained by several Terrestrial Laser Scanners to see the difference in vertical forest structure. Canopy structure is mainly captured by UAV-SfM and there is more point density around tree canopy than 3D data obtained by airborne laser. The point density of UAV around tree canopy is equivalent to that of TLS data, but tree trunk underneath the canopy cannot be reconstructed well, because the trunk is obscured by tree canopy and it is not visible on photos.