An investigation of tree extraction from UAV-based photogrammetric dense point cloud

摘要: Manually, tree detection with terrestrial field work is a nonprofit labor in terms of time, cost, and manpower. As rapid alternative, airborne laser scanners are widely use for data collecting. But this active remote sensing technology expensive, especially local small areas. At point, unmanned aerial vehicles stand as new opportunity collection platforms both large study This shows the usage vehicle platform to collect images. The images used generate 3D dense point clouds. cloud investigated random sample consensus algorithm order detect tree. trees assumed or cylinder which geometrically defined respect tree’s parameters such radius. According results, RANSAC successful from unclassified image-based cloud. 232 individual have managed extract rate 70.1% 3 different sites.