期刊名称：ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
印刷版ISSN：2194-9042
电子版ISSN：2194-9050
出版年度：2003
卷号：XXXIV-3/W13
出版社：Copernicus Publications
摘要：An automated method for boundary representation of building objects has been considered as a core processor for 3D city modelling. Since the reconstruction of generic building shape fundamentally depends on geometric features extracted from data sources, it suffers difficulties especially when a monocular imagery with high scene complexity is solely used. The research described in this paper aims to develop an automated method for building extraction, in which individual building object is localized and boundaries of polyhedral building shape are delineated with a less specific building model. The developed technique focuses on an exploitation of synergy of Ikonos imagery combined with a LIDAR DEM. Individual buildings are localized with rectangle polygon by a hierarchical segmentation of LIDAR DEM and Ikonos multi-spectral information. This polygon is recursively partitioned by linear features extracted from Ikonos image and LIDAR space, which results in a set of convex polygons. Only polygons comprising "significant" parts of building shape are verified and aggregated. Finally, polyhedral building shapes are reconstructed. Several results are presented with a discussion of evaluation and limitations of our method
