Master of Science
Dr. Jinfei Wang
Buildings play an essential role in urban intra-construction, urban planning, climate studies and disaster management. The precise knowledge of buildings not only serves as a primary source for interpreting complex urban characteristics, but also provides decision makers with more realistic and multidimensional scenarios for urban management. In this thesis, the 2D extraction and 3D reconstruction methods are proposed to map and visualize urban buildings. Chapter 2 presents an object-based method for extraction of building footprints using LiDAR derived NDTI (Normalized Difference Tree Index) and intensity data. The overall accuracy of 94.0% and commission error of 6.3% in building extraction is achieved with the Kappa of 0.84. Chapter 3 presents a GIS-based 3D building reconstruction method. The results indicate that the method is effective for generating 3D building models. The 91.4% completeness of roof plane identification is achieved, and the overall accuracy of the flat and pitched roof plane classification is 88.81%, with the user’s accuracy of the flat roof plane 97.75% and pitched roof plane 100%.
Zhao, Ting, "Object-based Urban Building Footprint Extraction and 3D Building Reconstruction from Airborne LiDAR Data" (2013). Electronic Thesis and Dissertation Repository. 1211.