Generation of Complex Polyhedral Building Models by Integrating Stereo-Aerial Imagery and Lidar Data

Abstract

The integration of lidar data and aerial imagery is a promising approach for accurate building model generation. In this research, lidar data and stereo-aerial imagery are incorporated in the generation of complex polyhedral building models whose rooftops are bounded by straight lines. The process starts by utilizing lidar data for deriving building hypotheses and the initial boundaries of the planar patches constituting the buildings' rooftops. The boundaries of these patches are then refined through the incorporation of stereo-aerial imagery while utilizing 3D geometric and spectral constraints. Precise building boundaries are derived through resolving feature matching problems and fully utilizing spectral information. Finally, an efficient manual mono-plotting procedure is introduced to remove incorrect and add missing boundaries. The performance of the developed procedures is evaluated through experimental results from real data where the correctness, completeness, and accuracy of the derived building models are evaluated through comparison with a manually generated DBM.