Proposed features/Ground Control Point

From OpenStreetMap Wiki
Jump to navigation Jump to search
Ground control point
Proposal status: Draft (under way)
Tagging: man_made=ground_control_point
Applies to: type
Definition: A set of coordinates for a feature visible from above, used for georeferencing imagery.

Draft started: 2020-10-13

A ground control point (GCP) is a set of coordinates and metadata intended to allow the georeferencing and/or photogrammetry of raster data such as satellite, aerial, or drone imagery. It corresponds to a long-lasting feature visible from above, but unlike a man_made=survey_point it may not be—and in fact in most cases probably is not—a marker specifically placed for this purpose, but rather is simply some relatively permanent visible feature.

Uses of Ground Control Points

Ground Control Points are used to correlate real-world locations with digital features. Sometimes referred to as "tie points," they establish the physical location/coordinates of specific points of raster or vector data.

In many places, accurate real-world coordinates for visible features are difficult to obtain. Frequently new Ground Control Points need to be gathered each time new imagery is taken, often because the data from previous GCP collection is inaccessible—languishing in a Google Drive folder or on someone's hard drive—even if there is no other impediment (licensing or otherwise) to re-use. Ensuring that GCP data is stored with all of the ancillary data needed for high-quality georeferencing in the global commons of OpenStreetMap is valuable for surveyors and mappers around the world!

Common uses of GCPs are:

  • Manually aligning satellite or aerial imagery for digitization, either by simple offsets in JOSM or similar, or by more sophisticated georeferencing using a GIS platform
  •  photogrammetry, whereby images from drones or other aerial platforms are processed into orthorectified photos  orthophoto and/or Digital Elevation Models using software such as OpenDroneMap or proprietary photogrammetry packages.

A GCP contains significantly more information than an ordinary node, as georeferencing requires:

  • Coordinate reference systems involved
  • Estimated accuracy and precision of the coordinates
  • Equipment used to capture the coordinates
  • Date and time the survey was done
  • Description of the exact point for which coordinates have been taken
  • Photographs of the location, which allow recognition of the feature in imagery being georeferenced

As well as, in some cases:

  • A RINEX file (or equivalent) with the raw GNSS measurements
  • The report from PPP or PPK processing
  • Height of the GNSS antenna above the actual feature

See also man_made=survey_point  Triangulation station and  Benchmark (surveying).

How to map

Ground Control Points are created by field surveys; the first step is to actually visit the point and obtain coordinates (generally using a GNSS receiver, though traditional surveying tools such as a compass and chain or theodolite on Wikipedia may be used in some cases). GCPs generally require photographs showing the feature so that it can be recognized in the imagery being georeferenced; a series of photos is taken, and a description of the exact placement of the point (for example: "the north-west corner of the bridge deck" or "The apex of the roof on the south end of the building") is recorded.

Add a node at the location of the point (using the coordinates, not using aerial imagery) and use the man_made=ground_control_point as well as the necessary additional tags.

Additional tags

Possible Rendering

None at present

Additional tags