2

I have troubles using several GRASS GIS tools from QGIS 3.16 (win 10) - I encountered the same problem applying v.clean as well as v.generalize tools (I tried Douglas Peuker algorithm). I tried a lot of different values for parameters. I always got an error and no output:

ERROR 1: osgeo::proj::io::WKTParser::Private::buildEllipsoid: non double value

ERROR: Unable to create OGR spatial reference

When reprojecting these polygon layers to EPSG:4326, the GRASS tools run without problem. I can set different values for the parameters (like threshold from 0.0001 to 1000), all work fine.

Why do I get errors when using the same tool/algorithm with the same layer in projected CRS?

markusN
  • 12,976
  • 31
  • 48
Babel
  • 71,072
  • 14
  • 78
  • 208
  • Distance calculations is Web Mercator are always wrong, so it is particularly unsuited for generalization algorithms. Geographic coordinate systems are also problematic, because Cartesian degrees are a less than useful comparison units as well. You really ought to use a PCS appropriate for distance calculation. – Vince Nov 21 '20 at 15:10
  • I'm aware that Web Mercator is not ideal for distance relecant proceedings. However, geographic CRS with degrees are much less suited for distance calculations than web mercator as a projected CRS with meters as units. However, it's exactely the geographic CRS that works perfectely. – Babel Nov 21 '20 at 15:30
  • 2
    Web Mercator is more than "not ideal", it's always wrong. Generalization in degrees works, but is more precise (less effective) the further you get from the poles. Web Mercator simply fails because distances approach infinity as you reach the poles. – Vince Nov 21 '20 at 15:54
  • @Vince I think this deserves an appropriate answer so other people can find it. Changing to WGS84 (4326) solved my problem.. – JMarcelino Dec 11 '20 at 16:56

0 Answers0