The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Publications Copernicus
Articles | Volume XXXVIII-4/C26
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XXXVIII-4/C26, 15–19, 2012
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XXXVIII-4/C26, 15–19, 2012

  07 Jun 2012

07 Jun 2012


B. Gorte and J. Lesparre B. Gorte and J. Lesparre
  • Optical and Laser Remote Sensing Delft University of Technology The Netherlands

Keywords: TIN, Point cloud, reconstruction, 3d data representation

Abstract. Point clouds obtained by aerial laser scanning are a convenient input source for high resolution 2.5d elevation models, such as the Dutch AHN-2. More challenging is the fully automatic reconstruction of 3d city models. An actual demand for a combined 2.5d terrain and 3d city model for an urban hydrology application led to the design of an extension to the well-known Delaunay triangulated irregular networks (TINs) as to accommodate vertical walls. In addition we introduce methods to generate and refine models adhering to our data structure. These are based on combining two approaches: a representation of the TIN using stars of vertices and triangles, together with segmenting the TIN on the basis of coplanarity of adjacent triangles. The approach is supposed to deliver the complete model including walls at the correct locations, without relying on additional map data, as these often lack completeness, actuality and accuracy, and moreover most of the time do not account for parts facades not going down to street level. However, automatic detection of height discontinuities to obtain the exact location of the walls is currently still under implementation.