[GRASS-user] Invalid region and coordinates when trying to reproject.

Hamish hamish_b at yahoo.com
Mon Jul 2 04:07:35 PDT 2012


Marcello:        
>>>>    N:  89:55:00.138825N   S:  89:55:00.138825
>>>>    E: 179:55:00.069428E    W: 179:55:00.069428W

>>> This inf looks very suspicious, the latlon vector extents very
>>> probably too large and could not be represented in the
>>> projection of the target location.

>> Too much for mercator? I am really lost here....

Markus M: 
> I think for mercator this is fine, but not for transverse
> mercator.

I think so too; if it helps visualize it:

Mercator is often shown with its +lat_0 at the equator, and the
distortion gets worse as you get away from that central lat,
until Greenland is huge and the north and south poles are beyond
the math. East-west along a low latitude should be ok, zoom out
all the way in Google maps or OSM.org for example.

Transverse Mercator is just the Mercator math rotated by 90
degrees AFAIU. So along +lon_0 distortion is nil from north to
south pole, but if you go more than 90 degrees east or west
of +lon_0 it blows up. (see attached* for UTM zone 42, here
Western Africa becomes like Greenland does in the Mercator proj.
The Pacific Ocean explodes too, but it is harder to see due to
lack of land)


Now, there's a bug that v.proj doesn't try both sides of 180 lon
to pull features, but that just leaves you with empty white areas
on one side, not a projection error message AFAICT.


[*] 20deg reprojected grid lines are a bit ugly, apparently we
need to extend the v.mkgrid breaks= option to work on vertical
lines as well as the horizontal ones.


Hamish
-------------- next part --------------
A non-text attachment was scrubbed...
Name: east_hemis_utm42.png
Type: image/png
Size: 11963 bytes
Desc: not available
URL: <http://lists.osgeo.org/pipermail/grass-user/attachments/20120702/066007f4/attachment.png>


More information about the grass-user mailing list