How to rasterize walls for grid calculations

From Tygron Preview Support Wiki
Revision as of 10:56, 19 September 2023 by Rudolf@tygron.nl (talk | contribs)
Jump to navigation Jump to search

For some calculations, it may be relevant to know where the walls of structures exist. This is a trivial consideration, but a bit more complex to implement in a geographical calculation. However, by using some localized checks on the behavior and local changes of elevation models, it is possible to create a spatial calculation which can find the locations of walls of structures.

Editor → Current Situation (Ribbon tab) → Overlays (Ribbon bar)
How to rasterize walls for grid calculations:
  1. Add a Heightmap Overlay, configured as DTM, to the Project (if it does not exist yet)
  2. Add a Heightmap Overlay, configured as DSM, to the Project (if it does not exist yet)
  3. Dsm averaging.jpg
    Add an Average Overlay to the Project with the following configuration
    Name: DSM Smoothing
    Set the averaging to the DSM Heightmap Overlay.
    Set the averaging distance to approximately 2.5 times the grid cell size.
  4. Wall combo overlay.jpg
    Add a Combo Overlay to the Project with the following configuration:
    Name: Walls
    Input A: DTM Overlay
    Input B:DSM Overlay
    Input C:"DSM Smoothing" Overlay
    Formula: AND(GT(SUB(C, A), 1), GT(SUB(C, B), 1))
  5. Wall combo overlay legend.jpg
    Modify the legend of the Combo Overlay to have only the following values:
    Value 0: No wall
    Value 1: Wall

Notes

  • This calculations performs an estimation of the locations of walls, and may miss spots. This calculation should therefor not be taken at absolute value, but can be used as a valid estimation in other calculations, such as the percentage of walls which also overlaps with specific sightlines.
  • The formula of the Combo Overlay checks whether the localized average DSM exceeds the absolute height by at least 1 meter, which generally means that the height of a wall should be at least 2 meters high. By changing the "1" into a lower or higher number, the approximation can be made more or less sensitive respectively.