Workshop #3:
The Representation of
Geographical Information in VRML.


Monday, February 16, 1998.  Monterey CA

I gave a short intro, in which I described a little bit about the genesis of the GeoVRML WG and the issues we are interested in discussing.  We then had a short intro session in which the attendees introduced themselves.  During this time, a contact sheet was distributed. 

Theresa-Marie Rhyne

The first presenter was Theresa-Marie Rhyne <>, the Director of the SIGGRAPH Carto Project., which was formed and funded by SIGGRAPH to attempt to bridge the perceived gap between the cartographic and computer graphics communities.  She gave a short description of the Carto Project and of a study that they commissioned to survey the cartographic community on its access to and use of computer graphics and visualization tools. They concluded that this community did have access to a wide variety of tools (primarily at the low-end) and was very actively interested in more advanced data presentation techniques. Theresa handled a few questions and pledged cooperation between the 
Carto Project and the GeoVRML WG. 

Don Brutzman

The second presenter was Don Brutzman <>, a professor at the Naval Postgraduate School in Monterey, and a member of the VRML Consortium Board of Directors (and after Monterey, the V.P. Technology of the VRML Consortium).  He passed the presentation on to a student, Greg Leaver <>, who presented the Monterey Bay Modeling Group's "Virtual Monterey Bay," a virtual world derived from real bathymetric data measuring the ocean floor in the Monterey Bay, one of the richest areas of marine life in the US.  This VRML world is in the planning stages and is being designed to be useful to both scientists and the general public (they are collaborating with the Monterey Bay Aquarium). 

He then described their plan, which involves tiling the bathymetric data over a 150x150 km area.  The underwater terrain is represented in VRML as an LOD hierarchy of tiled terrain in which each tile is <100K. 



The next presentation was by Farid Mamaghani <> who described the SEDRIS project. This is a US Defense Department standardization effort for the representation and manipulation of datarelevant to simulation and training environments. SEDRIS stands for "Synthetic Environment Data Representation & Interchange Specification." Reading "synthetic environment" as "virtual reality," we can see the relevance to our community in that they are standardizing georeference and space-time sequencing of real-world data. To that end, they are developing and freely distributing tools for managing their standardized data formats.  We were then given a short overview class on geodesy and geographical coordinate systems by Ralph Toms <> who is developing the coordinate system library for SEDRIS. He discussed Earth Reference Models (ERMs), World Coordinate Systems, and the difficulties and limitations in converting from one to others. He gave a three to four month timeline before the high-performance, high-accuracy software he's working on will be completed and publicly available. 

Martin Reddy

The last speaker was Martin Reddy <>, who spoke about TerraVision, the terrain visualization system we're developing at SRI. It is a high speed, very high bandwith system for visualizing extremely large (10-100GB) terrain data sets (typically hundreds of square kilometers with elevations at 30m resolution and imagery at 1m). The first version of it was built using OpenGL and portable C on high-end SGI workstations using an application-specific data format. The second version, which he discussed also uses OpenGL, but we have adopted VRML as the external data representation and have developed our own EXTERNPROTOs for georeferencing and managing topology in very deep level-of-detail hierarchies. We too are managing tiled terrain, but by explicitly representing the neighbourhood topology of tiles, we can dynamically modify the 
geometry (in TerraVision) we avoid the surface tears when LOD levels change within a scene.  A CD of one of our sample VRML tile sets was distributed at the meeting, and was snapped up like hotcakes. 
Lee Iverson <>

Last updated: Tuesday, 21-Jul-1998 15:13:05 PDT