Hasso-Plattner-Institut
  
Hasso-Plattner-Institut
Prof. Dr. Jürgen Döllner
  
 

Research Challenges for GeoVE

The research group is developing concepts, methods, and techniques for the visualization of complex, massive geovirtual 3D environments (GeoVE). Research topics include:

  • Multiresolution terrain rendering and texturing
  • Multiresolution building and site modeling and rendering
  • Photorealistic 3D rendering techniques for GeoVE
  • Non-photorealistic 3D rendering techniques for GeoVE
  • Assistive 3D navigation and interaction techniques for GeoVE
  • Non-linear 3D projections and panorama maps for GeoVE
  • Visibility and visual contact 3D analysis based on GeoVE
  • 3D Generalization of complex GeoVE
  • 3D Geodata generation and synthesis
  • Service-oriented architectures for GeoVE systems and applications
  • Cartographic, perceptual, and cognitive aspects of GeoVE
  • Computational and conceptual aspects of CityGML

GeoVE as Spatial Integration Platform

GeoVEs exhibit an outstanding characteristic: they are excellent platforms for the integration of heterogenous, distributed geodata and georeferenced data. Most important, the integration takes place at "visualization level" so that individual formats, databases, and model schemata of data don't have to be unified; in the framework of the GeoVE the multitude of data and objects appear to be seamlessly merged. In the image, part of the goverment district of Berlin is shown; this virtual 3D city model integrates various 3D building sources such as VRML models, X3D models, and cadastre-based models as well as many offical 2D geodata sources.  

The Berlin-3D City Model as Integration Platform.

Automated Generation of GeoVE

A long-term goal of HPI's research in 3D geoinformation and geovisualization represent methods and technique for the automated generation and maintenance of large-scale, massive GeoVEs. In the image, an automation strategy is illustrated: roof textures are automatically assigned to roof geometry of building models. More advanced methods take into account object semantics and heuristics to reduce the manual work load needed to set up and maintain GeoVEs. Automation and maintenance are inherently related: only if generation processes can be automated, maintenance becomes feasible for large-scale GeoVE.

Content Management System for 3D City Models.

GeoVE for Spatial Data Mining

GeoVEs and in particular virtual 3D city models enable manifold applications beyond traditional fields. One of these applications represents spatial data mining. The virtual 3D city model, roughly speaking, serves as panorama to project any georeferenced thematic data, e.g., taken from enterprise systems. This way, even massive georeferenced data sets can be explored within the virtual city model, using the different "graphics variables" of city models as channels to visualize multi-dimensional data sets.

Urban Data Mining by Virtual 3D City Models.

Interaction with GeoVE and Usability

In our research, we also focus on innovative interaction concepts and techniques. In particular, we develop techniques that take advantage of the semantics of geospatial objects and their affordances with respect to interaction and behavior.  

For example, one approach is based on two aligned concepts: 1) All navigation techniques are controlled by constraints that ensure user orientation and avoid "getting lost" situations. 2) All navigation techniques are handled in a time-coherent way achieving steady, continuous user movements using a physically-based motion model. Based on these concepts, we have demonstrated several ways to improve commonly used navigation techniques for GeoVEs.