Grand Canyon, AZ
As part of his View Dependent Progressive Mesh work, Hugues
Hoppe evaluated his system with a terrain model of the Grand
Canyon. This includes a 4097x2049 grid and 4096x2048 texture,
both in BMP format.
Puget Sound, WA
Peter Lindstrom and Valerio Pascucci's IEEE Visualization 2001 paper
used a large terrain model of Puget Sound in the Washington state. The
data is from the US Geological Survey and the University of Washington,
totalling 16,385 x 16,385 pixels and provided in PNG format.
Martin Reddy demonstrated
his perceptually modulated LOD terrain rendering system using a model of
Yosemite Valley in California (IEEE C&GA 2001).
These data include 1 m satellite imagery,
21,135 x 27,090 pixels, and 32 m elevation data, 661 x 847 postings.
SRI International have produced a number of terrain datasets for
the TerraVision terrain visualization system. This particular dataset
is a mosaic of the USGS GTOPO30 datasets to produce a global 30-arc
second elevation grid for the world (roughly 1-km resolution at the
equator). This gives 43,200 x 21,600 grid postings, for a total of
almost a billion points.
Virtual Terrain Project (VTP) data
The VTP site is maintained by Ben Discoe with the goal to foster the
creation of tools for easily constructing any part of the real world
in interactive, 3D digital form. The Open Source implementation
includes demo data the consists of elevation and satellite imagery
for several areas including Hawai'i.
ROAM sample data
Mark Duchaineau provides some sample terrain models along with the
source code for the ROAM (Real-time Optimally Adapting
Meshes) algorithm. These use a simple ASCII format to define the