ABSTRACT
CONTENTS 13.1 Introduction ............................................................................................. 251 13.2 Building Detection and Extraction....................................................... 252
13.2.1 Edge Detection........................................................................... 252 13.2.2 Building Polygon Extraction.................................................... 256
13.3 Creation of DBM and DTM .................................................................. 256 13.3.1 Establish Relationship between Building
and LiDAR Data ........................................................................ 256 13.3.2 Interpolation Algorithm via Planar Equation....................... 258 13.3.3 Establish Digital Building Model............................................ 258 13.3.4 Create Digital Terrain Model .................................................. 259
13.4 Experiments ............................................................................................. 259 13.4.1 Datasets ....................................................................................... 259
13.4.1.1 Control Field.............................................................. 259 13.4.1.2 Aerial Image Data..................................................... 260 13.4.1.3 LiDAR Data ............................................................... 260
13.4.2 Experimental Results ................................................................ 261 13.4.3 Discussion................................................................................... 264
13.5 Conclusion ............................................................................................... 266 Acknowledgments ............................................................................................. 266 References ........................................................................................................... 266
The urban three-dimensional (3D) model has increasingly been needed for various applications such as town planning, microclimate investigation, and so on. Traditional photogrammetry is an important tool to acquire the 3D data. However, the photogrammetric method has encountered difficulties
for complex scenes in dense urban areas due to the failures of image matching, which are primarily caused by, for example, occlusions, depth discontinuities, shadows, poor or repeated textures, and the lack of a model of man-made objects (Zhou et al., 1999). For this reason, the extraction of buildings and digital terrain model (DTM) are currently carried out using human-guided interactive operations, such as stereo compilation from screen. This process is both costly and time-consuming.
