Building is important element of the city which affects human life in these days. Detecting the shape of roofs is necessary to describe buildings. 2D Images have traditionally been used for acquiring 3D information with stereo viewing. However it has difficulty for acquisition of 3D information because 3D information has been acquired with 2D images. LiDAR system can archive directly 3D information of object with point data type because LiDAR system use laser pulse, that reach object and archive distance between aircraft and object. Automatic detection of roof components for describing building is performed in this study using LiDAR system which can archive 3D information directly. For the purpose, segment the LiDAR data to each patch using octree space segmentation method. Then if patches have same space characteristic merge the patches with analyzing the space characteristic of each patch. As a result LiDAR data is classified to the components of roof.
展开▼
