3D city models
Prof. Dr. D. Fellner, Dipl.-Ing. M. Lancelle
Computer Graphics and Knowledge Visualization, Graz University of Technology, Austria, www.cgv.tugraz.at
- automatic 3D city model generation
- using already existing data
- covering a 200 km² area
- interactive visualization
Visualization
Focusing on the visualization purpose, the results were used as a feasibility study to be able to judge what is possible at the moment. Mainly the data of the City of Braunschweig was used. The data sets of high resolution orthophotos (aerial photos with a 10cm grid), LIDAR data (airborne acquired laser range surface scan) on a 1 metre grid, cadastral maps, soil usage and a database with street names and house numbers were converted into formats suited for a real-time visualization. Buildings and trees were automatically extracted as well. Very little manual work was involved during the conversion which made it possible to process the whole city area of 200 km2.
With the City Model Viewer one can explore the city interactively on a standard PC as well as on a stereo projection setup or in a CAVE. Camera and animation paths are provided and interactive features were implemented like creating a path at mouse pointer position or showing the street name and house number of the house under the mouse pointer.
street names, house numbers |
elevation model |
3D buildings |
aerial photos |
|
|---|---|---|---|---|
 |
 |
 |
 |
soil usage |
cadastral map, plants |
games, movies |
traffic noise |
|
|---|---|---|---|---|
 |
 |
 |
 |
city planning |
tourism, navigation |
flooding |
air flow |
radio network planning |
|---|---|---|---|---|
 |
 |
 |
 |
 |
Idea
This work uses already existing digital data to automatically generate a 3D model of a city. For interactive navigation through this 3D model the “City Model Viewer” has been developed.
3D city models can be used in a wide range of applications: information and navigation, real estate, Geo Information Systems (GIS), city planning, analyses and physical simulations, military applications and entertainment. All these different cases have various requirements. There is still a lot of work to do in terms of standardization so that the same models may be used for different applications and that the same software can be used for different cities.
Buildings with different, automatic assigned wall textures. |
Visualization of the LIDAR elevation model. |
|
|---|---|---|
 |
 |
 |
Automatic handling of whole cities
These screenshots above show the Braunschweig city model. The source data is provided for the whole region of Braunschweig (200 km²). All buildings are automatically generated from 2D CAD data, Airborne Laser Scanning (LIDAR) heightmaps (see upper screenshot) and aerial photos. Landmarks and moving objects can be inserted manually.
The east quarter of downtown (Braunschweig)
