Geographical Information Systems (GIS) are used for mapping and analysis of data pertaining to geographic locations. The location data may consist of vectors, rasters, or points.
Vector data are typically used to represent boundaries of discrete political entities, zoning, or land use categories.
Raster data are often used to represent geographic properties that vary continuously over a 2D area, such as terrain elevation. Each raster represents a small rectangular finite element of information projected onto a regular 2D grid. It’s simple to construct a triangulated mesh from such data.
Unstructured point clouds are often acquired by LIDAR or other scanning techniques. Dense clouds of points can create a fuzzy visual impression of 3D surfaces of terrain, vegetation, and structures. Unlike raster data, point clouds can represent concave, undercut surfaces, but it’s harder to construct a triangulated mesh from such data.
The MIDEN demo collection includes test cases for generic loaders of all three of these types. The vector tests include boundaries of roads and wooded areas in Ann Arbor projected onto a 2D map, and national boundaries projected onto the surface of a globe. The raster test is a 3D terrain mesh for a section of the Grand Canyon. The point test is a LIDAR scan of a fault line in the Grand Tetons.