Implementation of Roads in the VTP software

• the vtdata library support a network of roads (paved roads, unpaved, trails, railways, etc.)
  • this functionality it implemented in the class vtRoadMap
  • read centerlines data from DLG, TIGER, SHP, and other vector data formats
  • clean up the network topology
  • read and write RMF (road map files), an ad-hoc binary format which stores:
    • CRS (string, as OGC WKT)
    • extents of dataset (double left, right, bottom, top)
    • node locations (2D point)
    • link locations (2D polylines)
      • road number (short)
      • width in meters (float)
      • number of lanes (short)
      • surface type (short enum)
      • flags (short, contains direction of flow, presence of sidewalk and other flags)
      • sidewalk width (float)
      • curb height (float)
      • margin width (float)
      • lane width (float)
      • parking width (float)
      • topology (indices of from-node and to-node, int)
• the vtlib library can read RMF and generates simple triangle strips for the roads, draped on a terrain
  • this functionality it implemented in the class vtRoadMap3d
  • in addition to the roadway surface, can create additional triangle strips for: margins, parking areas, sidewalks
  • applies correct road marking (yellow and white stripes) based on number of lanes and traffic direction
  • some caveats:
    • Intersection (junction) geometry is rudimentary: a simple polygon connecting the adjoining roads, with a 'pavement' texture.  Better intersections would require more knowledge of how the lanes of flow move through the node - connectivity of the lanes, direction of flow per lane.  Even if there was an encoding to support it, this information is rarely available.
    • Draped roadways can sometimes have trouble with the terrain, see Z-Buffer Issues