-
-
Notifications
You must be signed in to change notification settings - Fork 37
Open
Description
mvt.PrepareGeometry (and mvt as whole) was written to be srid agnostic, but the follwing lines would fail in the edge case that the projection had axis oriented differently from the common web mercator and lat/longs, such as:
These types of coordinate systems are also refered to as Lo coordinate systems.
The following lines are relevant the mvt code:
Lines 66 to 67 in f76f5a1
| px := int64((g.X() - tile.MinX()) / tile.XSpan() * pixelExtent) | |
| py := int64((tile.MaxY() - g.Y()) / tile.YSpan() * pixelExtent) |
However, I think the change would need a new or re-designed types. One thing that comes to mind is tweaking the
Extent elements to be left, top, right, bottom instead of min/max X/Y. This would not require any changes under our current assumptions in tegola of only using epsg 3857 and 4326.
Here are some relevant links to discussions on this topic:
https://community.esri.com/thread/185667-why-is-the-projection-upside-down
https://isis.astrogeology.usgs.gov/IsisSupport/index.php?topic=370.0
http://www.ngi.gov.za/index.php/technical-information/geodesy-and-gps/datum-s-and-coordinate-systems
gdey
Metadata
Metadata
Assignees
Labels
No labels