Hi Frank & Ed, Thanks for your prompt and efficient answers!!!
You're right, the pbm seems to come from the prime meridiam mismatch (and not decimal degree conversion issue as I thought before). I was able to confim this by manually doing some conversion tests using the Convers useful little tool (http://vtopo.free.fr/convers.htm ). As per example, for a WGS84 GPS point in Paris: longitude=2.38388 latitude=48.85562 In Lambert II Etendu, I have today (wrong value) : X=775063.135566471 Y=2431242.54750943 which is given using "Paris" as origin in Convers and which corresponds to the postgis transform call : psql> select public.transform(GeometryFromText('POINT (2.38388 48.85562)', 4326), 27582); SRID=27582;POINT(775063.135566471 2431242.54750943) So I have to find other EPSG values to get the correct X/Y that should be: X=603478.33780249 Y=2428587.76696934. I didn't find it yet, but I keep investigating in the spatial_ref_sys postgis (0.9.1) table!!! Regards, FP -----Message d'origine----- De : UMN MapServer Users List [mailto:[EMAIL PROTECTED] De la part de Ed McNierney Envoyé : mercredi 11 janvier 2006 16:47 À : [email protected] Objet : Re: [UMN_MAPSERVER-USERS] Projection WGS84 in decimal degree Frank & FP - Good catch, Frank - that's an excellent possibility. FP, make sure your EPSG file contains an accurate definition of <27582>, including the prime meridian (primem) of 2.337229166666667. - Ed Ed McNierney President and Chief Mapmaker TopoZone.com / Maps a la carte, Inc. 73 Princeton Street, Suite 305 North Chelmsford, MA 01863 [EMAIL PROTECTED] (978) 251-4242 -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Frank Warmerdam Sent: Wednesday, January 11, 2006 10:16 AM To: Ed McNierney Cc: [email protected] Subject: Re: [UMN_MAPSERVER-USERS] Projection WGS84 in decimal degree On 1/11/06, Ed McNierney <[EMAIL PROTECTED]> wrote: > > FP - > > Your first GPS layer definition should be fine. You don't say what > data source you are using for your GPS data, but if it is a shapefile > or inline features you will be using decimal degrees and those will > work fine with the "init=epsg:4326" definition. > > The first thing I would suggest is that you confirm that the > EPSG:27582 projection is in fact correct. The related Paris / Lambert > projections (27581-27584, for example) are all rather similar and can > certainly look the same. Can you check the coordinates of a few > points to make sure they're accurate? An error of 200km is far too > large to be a datum error, but could easily be due to the different > false easting/northing values used in these similar projections. FP, I would add to Ed's comments that the PROJ definitions files for EPSG zones with alternate prime meridians are often messed up, and a horizontal offset of 250km is approximately the distance of the paris meridian from greenwich. Best regards, -- ---------------------------------------+-------------------------------- ---------------------------------------+------ I set the clouds in motion - turn up | Frank Warmerdam, [EMAIL PROTECTED] light and sound - activate the windows | http://pobox.com/~warmerdam and watch the world go round - Rush | Geospatial Programmer for Rent
