Nav view search





3D City Model of Aachen for the "Route Charlemagne"
Martin Habbecke, Arne Schmitz and Leif Kobbelt

The Route Charlemagne is a cultural project of the city of Aachen. It is named after Charlemagne, King of the Franks and Emperor of the Romans around 800AD, who chose Aachen as his favourite seat of government. The route leads through the center of Aachen, connecting historical sites with modern buildings hosting parts of the university and the local government. Commissioned by the Aachen city council, we constructed, animated and rendered a 3D city model of Aachen for the virtual tour of the Route Charlemagne



Open3DCity: Collaborative 3D City Modeling from Aerial Images
Martin Weusten, Martin Habbecke and Leif Kobbelt

In this project, we have developed an image-based modeling system for the construction of 3D building models from oblique aerial images. Based on our earlier work on precisely registering oblique aerial images with a cadastral map, all modeling operations take place in image-overlay, such that it is an easy task to recover the true dimensions of a building. The main design goal was simplicity of the user interface to enable users unfamiliar with (usually very complex) standard 3D modeling tools to successfully use our system. The resulting system employs a geometry representation by simple, generic building blocks that can be stacked and attached to one another in order to construct more complex shapes. We only provide a small set of editing operations specifically tailored to the domain of 3D building models, thereby lowering the error rate of novice users. In his diploma thesis project, Martin Weusten has turned this system into a multi-user, web-based collaboration system. Please see the project page (in German) for more details and tutorial videos.



Tool Path Generation from Geodesic Distance Fields
Martin Habbecke and Leif Kobbelt

Milling and sheet metal forming are important manufacturing techniques in industrial environments. While the digital design of production parts using CAD/CAM systems has become commonplace, the generation of suitable tool paths for a given surface to control a milling machine is a challenging problem. To run a milling machine at optimal efficiency, the tool paths are required to have several properties: To minimize the manufacturing time, it is desirable to minimize the number of required re-positionings of the milling head and hence cover the part's surface with as few disconnected tool paths as possible. The variation of the distance between neighboring path segments is supposed to be small to remove the optimal (constant) amount of material at each time step of the milling process. In addition, the tool paths need to be smooth without sharp corners to allow the milling head to move with maximal velocity.

We have developed a tool path generation algorithm for arbitrary polygonal surfaces that aims at optimizing each of the above three criterions. Our method is based on the computation of a geodesic distance field in a surface region delimited by a set of given boundary curves. By tracing iso-contours in the distance field, we generate curves with constant pairwise distances. These curves are linearly blended to form a single, continuous tool path.  To guarantee smooth tool paths, our algorithm performs error-bounded smoothing of the resulting paths on the surface as well as in surface normal direction in case of a noisy input geometry. The pictures below show an exemplary input geometry, the computed geodesic distance field and the resulting tool paths.