Development of a highly scalable interactive visualisation environment for large, dynamic infrastructure projects.
Visualising large and dynamic infrastructure projects has become an important tool in planning, publicpresentation and maintenance. NGIP is used to consolidate andvisualise geo-referenced planning and inventory data (roads, noise barriers, etc.).NGIP can be closely linked with existing work processes and databases, and offers support for highly dynamicscene elements such as realistic renderings of road and rail traffic using recorded or simulated traffic statistics. The visualization of large scale, dynamic infrastructure models has become an important tool for planning, public information and discussion and maintenance of such projects. Our existing GEARViewer prototype - developed within Interactive Planning and other projects over several years in close cooperation with GEOCONSULT Wien ZT GmbH and ASFINAG - has proven to be a reliable tool in this process, and the NGIP project extends its functionality in two important fields: better integration in the existing workflow and central data warehouse of our project partner ASFINAG, and support for highly dynamic components, such as populating the environment with realistic traffic (based on computed or measured traffic statistics).
Additionally, the large user base at ASFINAG and possible public interest strongly suggests the creation of a portable viewer application. This project shall therefore also study the feasibility of supporting remote rendering or other portable display solutions for the given application. Such a solution may include remote rendering / streaming, portable applications or interfacing to existing portable solutions (such as Google Earth).
Since the data warehouse at ASFINAG is primarily designed for management tasks, and not for visualization, the stored objects are often not directly usable for geospatial visualization. For example, highways may only be represented by lane center lines; noise abatement structures lack information on the particular design, and so on. However, transferring these objects into 3D space is not only a valuable tool in presentations for the general public, but also allows experts much better insights into the quality of their data because conflicting or inconsistent information is much easier to see in 3D than in 2D, where duplicate entries are hard to distinguish. However, the lack of detailed geometric information poses several problems concerning interfacing to these data sources, data formats, and presentation. Although the system design shall be sufficiently generic to allow future extensions to additional sources and types of information, a number of key features have been identified to be of particular interest - for these object types, a flexible procedural model will be created to synthesize geometry from the information available within the data warehouse.
The inclusion of highly dynamic objects also creates unique challenges. Previous optimization approaches made heavy use of the static nature of the environment to group and reorder scene elements for optimal performance; these optimizations need to be extended or modified to also support dynamic elements. Additionally, the real-time realistic simulation (according to average or maximum daily traffic volume) of dense traffic with up to 1000-1500 vehicles is a complex and computationally intensive task - and just as in real life, even the smallest collisions attract considerable interest and should be avoided.