gpio Oxxce ha scritto:
Well, The reprojection has done by ArchInfo.
Attach you can find the 2 files that instructs ArchInfo of how project
srs4121-srs4326 and vice-versa.
Also attached you can find a shapefile in two different projections.
Hum, thanks for giving me a sample.
I looked at the files.
First thing that shows up is that the .prj file included in 4121 directory does not represent 4121, but a transverse mercator projection
based on it, apparently 2100 (GGRS87 / Greek Grid).
The second issue is that Geoserver is in dire straits there, because
your data has projection infos, and it'll trust them more than what
you declare. On one side this is good, because 4121 was completely
wrong, on the other side it's bad, because that .prj file does not
have the TOWGS84 section which is needed for the ellipsoid change.
That's why when you try to reproject from the file in 4121 folder
to 4326 it seems points aren't changed at all, the bursa-wolf parameters aren't there.
If you use the .prj file attached, which I generated using the
EPSG:2100 definition in Geoserver, you should get better results.
Having points become completely match is another story, since
ArcMap and Geoserver do use different programming languages
and maybe slightly different formulas.
Hope this helps
Cheers
Andrea
poi_irakliou.prj (767 Bytes)