[Gvsig_english] trouble with EPSG 2154 and 27561
Juan Lucas Domínguez Rubio
juan_lucas_dr at yahoo.com
Sat Jul 17 16:42:43 CEST 2010
Hello, Michel:
I believe gvSIG OADE 2010 is not performing any transformation at least on Windows Vista. Manual, EPSG, and grid-based transformations are all ignored. I think this is why you got a 50 m error in your test. Which operating system are you using?
Official gvSIG 1.10 does not have this problem, and I have been able to use grid-based transformation with your shapefiles successfully. You can try this on gvSIG 1.9 or gvSIG 1.10:
- close gvSIG if it's running.
- download http://www.killetsoft.de/zip/ntv2/france.zip to your PC and unzip it.
- rename the file FRANCE.GSB to ntf_to_r93.gsb (the extension must be .gsb in lower case) and copy it to the folder:
[GVSIG_FOLDER]\bin\gvSIG\extensiones\org.gvsig.crs\data
- check that the size of ntf_to_r93.gsb in that folder is about 271 KB.
- start gvSIG and do the test in the same way as before, but instead of choosing 'Manual transformation', choose 'Grid transformation', then in the combo box, you should be able to choose ntf_to_r93.gsb and then -very important- you have to change the radio button under the combo box. Choose the one to the right 'CRS of the view'.
In this way, I get a much smaller error. Some vertices differ by less than one cm, others differ by 1.5 cm, etc.
Out of curiosity:
Where do your shapefiles come from?
One of them is the original and the other was derived very accurately?
Or maybe both were derived from a third (original) shapefile?
How were they derived? I thought they could have been derived using the same .gsb file I have used, but I think a divergence of 1.5 cm cannot be caused by accidental roundings.
Regards,
Juan Lucas Domínguez Rubio
http://gvsigmobileonopenmoko.wordpress.com
http://tellusproject.blogspot.com
--- On Fri, 7/16/10, Bouillot Michel <michel.bouillot at montreuil.fr> wrote:
From: Bouillot Michel <michel.bouillot at montreuil.fr>
Subject: Re: [Gvsig_english] trouble with EPSG 2154 and 27561
To: "Users and Developers mailing list" <gvsig_internacional at listserv.gva.es>
Date: Friday, July 16, 2010, 3:02 PM
#yiv701327610 p {margin:0;}Hello,
50m is with OADE 2010 !!!
With gvsig 1.9 it's only 1.4 m (that is right with the precision of the datum shift)
If I create an user CRS with a copy of 27561 where I only change the central meridian, I have the same result that with your proposal of user CRS. I think that there is a problem with meridian "non Grenwich" in lambert conformal conic 1SP.
Regards,
Michel Bouillot
----- Mail Original -----
De: "Bouillot Michel" <michel.bouillot at montreuil.fr>
À: "Users and Developers mailing list" <gvsig_internacional at listserv.gva.es>
Envoyé: Vendredi 16 Juillet 2010 11:08:16
Objet: Re: [Gvsig_english] trouble with EPSG 2154 and 27561
#yiv701327610 p {margin:0;}Hello,
Thanks for the answer.
The process works but I find an error of about 50m.
I take time to study it and I tell you.
Regards
Michel Bouillot
----- Mail Original -----
De: "Juan Lucas Dominguez Rubio" <jldominguez at prodevelop.es>
À: "Users and Developers mailing list" <gvsig_internacional at listserv.gva.es>, "Users and Developers mailing list" <gvsig_internacional at listserv.gva.es>
Envoyé: Vendredi 16 Juillet 2010 10:04:27
Objet: Re: [Gvsig_english] trouble with EPSG 2154 and 27561
#yiv701327610 p {margin:0;}
Hello, Michel:
I think there's some confusion about the meaning of the EPSG codes. According to spatialref.org, the PRJ file for EPSG:27561 is this:
http://spatialreference.org/ref/epsg/27561/prj/
But that file is not like the one that comes with your shapefile.
gvSIG is probably using the definition found in the spatialref.org website, so that's why your shapefiles do not match if you select EPSG:27561 when adding that file to your view.
According to spatialref.org, the CRS that matches your PRJ file is 'NTF Lambert Zone 1 France 2SP' (code SR-ORG:6910):
http://spatialreference.org/ref/sr-org/6910/prj/
So all you have to do is declare it as a user-defined CRS in gvSIG.
This has worked for me (I'll refer to your files as montreuil_2154.shp and montreuil_27561.shp):
- from the project manager, click 'New' to create a new view. Select it (without opening it) and click on 'Properties'
- clic on the [...] button to change the view's CRS
- choose 'Type: user CRS' and click on 'New'
- choose the 'From WKT' option and paste this paragraph in the text box:
PROJCS["unnamed",GEOGCS["unnamed ellipse",DATUM["unknown",SPHEROID["unnamed",6378249.2,293.4660200000053],TOWGS84[-168,-60,320,0,0,0,0]],PRIMEM["Greenwich",0],UNIT["degree",0.0174532925199433]],PROJECTION["Lambert_Conformal_Conic_2SP"],PARAMETER["standard_parallel_1",48.5985227778],PARAMETER["standard_parallel_2",50.3959116667],PARAMETER["latitude_of_origin",49.5],PARAMETER["central_meridian",2.3372291667],PARAMETER["false_easting",600000],PARAMETER["false_northing",200000],UNIT["Meter",1],AUTHORITY["EPSG","12345"]]
- click on 'Finalize' and 'Accept'. You'll see that gvSIG refers to the new CRS as 'USR:1' or similar.
- click 'Accept' again and open the view
- add the file montreuil_27561.shp without any additional option because we know it's the same CRS of the view.
- go to the 'Add file' option again, choose the file montreuil_2154.shp and before clicking on 'Accept', click on the button to indicate the CRS of the file that is being added. Search for EPSG:2154 (this one has no problem, you can use it) and select it, then you need to choose the 'Manual transformation' option in the transformation combo box. The seven parameters are:
168
60
-320
0
0
0
0
- click 'Accept' on all the following dialogs, and both shapefiles should match. There is a small error of about 1.5 m, but I don't think that's a problem for you.
Does this process work for you?
The CRS you have defined should be available from now on, but you will have to type the seven parameters each time from the 'Manual transformation' option.
Those seven parameters are valid if the view is in 'NTF Lambert Zone 1 France 2SP'. If the view is in EPSG:2154 and you add a shapefile in 'NTF Lambert Zone 1 France 2SP', you'll have to reverse the sign of the parameters.
If you don't want to deal with user CRS and parameters inside gvSIG, you can reproject your shapefiles before starting gvSIG. There is a nice free program that does that:
http://fwtools.maptools.org
The command would be something like:
ogr2ogr -f "ESRI Shapefile" -s_srs epsg:2154 -t_srs epsg:27561 new_file.shp old_file_2154.shp
(but again, we should check the meaning of EPSG:27561 in that case)
Regards,
Juan Lucas Domínguez Rubio
---
Prodevelop SL, Valencia (España)
Tlf.: 96.351.06.12 -- Fax: 96.351.09.68
http://www.prodevelop.es
---
De: gvsig_internacional-bounces at listserv.gva.es en nombre de Bouillot Michel
Enviado el: jue 15/07/2010 14:22
Para: Users and Developers mailing list
Asunto: Re: [Gvsig_english] trouble with EPSG 2154 and 27561
Hello,
I think that you are right.
I join shapefiles in both EPSG.
Thanks for the tests.
Regards,
Michel Bouillot
----- Mail Original -----
De: "Juan Lucas Dominguez Rubio" <jldominguez at prodevelop.es>
À: "Users and Developers mailing list" <gvsig_internacional at listserv.gva.es>, "Users and Developers mailing list" <gvsig_internacional at listserv.gva.es>
Envoyé: Jeudi 15 Juillet 2010 13:29:45
Objet: Re: [Gvsig_english] trouble with EPSG 2154 and 27561
#yiv701327610 p {margin:0;}
Hello, I think there's a misunderstood.
gvSIG is able to do on-the-fly reprojection (this is what Michel needs, I think).
But I'm afraid that, even if he follows the right steps to do it, the result will be wrong because (I think) gvSIG does not know the right parameters to perform that conversion (between EPSG:2154 and EPSG:27561).
There have been problems with the French projections in the past, and I'm not sure if they were caused by a bug in gvSIG or in the libraries used by gvSIG (especially Proj4).
Michel: can you please send to the list two little shapefiles (one in EPSG:2154 and the other in EPSG:27561) that should match but do not match in gvSIG? So we can also do some tests.
Regards,
Juan Lucas Domínguez Rubio
---
Prodevelop SL, Valencia (España)
Tlf.: 96.351.06.12 -- Fax: 96.351.09.68
http://www.prodevelop.es
---
De: gvsig_internacional-bounces at listserv.gva.es en nombre de Bouillot Michel
Enviado el: jue 15/07/2010 13:02
Para: Users and Developers mailing list
Asunto: Re: [Gvsig_english] trouble with EPSG 2154 and 27561
Hi,
Thanks for this explanation.
I find GVSIG very good but in France we have just changed of RS. We begin to have data in the 2 RS.
The reprojection in the view would be very pactical !
Is there a hope for a such function in the futur ?
Regards
Michel Bouillot
----- Mail Original -----
De: "Guillermo Tamburini" <tamburini_gui at gva.es>
À: "Users and Developers mailing list" <gvsig_internacional at listserv.gva.es>
Envoyé: Jeudi 15 Juillet 2010 12:43:50
Objet: Re: [Gvsig_english] trouble with EPSG 2154 and 27561
Hi,
In fact, you can't correctly represent data in different reference systems in the same View (without reporojecting, it have no sense).
There are some software applications which applies automatically the reprojection, but it is not the case in gvSIG with vector layers.
First of all, you have to select the correct RS for your View, and then load the data (the target RS you are going to work).
Otherwise, you will see your data but in a incorrect situation. In other utm spindle, or displaced...
If your data is in different RS, you will have to reproject it. Goprocessing tools->Data Conversion->Reproject; the layer in the not current projection to the current projection.
Regards
Bouillot Michel wrote:
#yiv701327610 p {margin:0;}
#yiv701327610 p {margin:0;}
#yiv701327610 p {margin:0;}
Hello,
I try to open in the same view one shp in EPSG 2154 and one SHP in EPSG 27561 but it don't match correctly. There is a problem with the X like a translation witch have a value who seems to be the difference between Greenwich and Paris meridian.
How to correct this trouble ?
regards
Michel Bouillot
Chef de projet SIG
01 48 70 63 23 / michel.bouillot at montreuil.fr
Ville de Montreuil / Direction Urbanisme et Habitat
Service Études Développement Urbain
3, rue de Rosny / 93100 Montreuil
_______________________________________________
Gvsig_internacional mailing list
Gvsig_internacional at listserv.gva.es
http://listserv.gva.es/cgi-bin/mailman/listinfo/gvsig_internacional
--
Guillermo Tamburini Beliveau
Proyecto gvSIG
Consellería de Infraestructuras y Transportes (GVA)
Valencia (Spain)
_______________________________________________
Gvsig_internacional mailing list
Gvsig_internacional at listserv.gva.es
http://listserv.gva.es/cgi-bin/mailman/listinfo/gvsig_internacional
_______________________________________________
Gvsig_internacional mailing list
Gvsig_internacional at listserv.gva.es
http://listserv.gva.es/cgi-bin/mailman/listinfo/gvsig_internacional
_______________________________________________
Gvsig_internacional mailing list
Gvsig_internacional at listserv.gva.es
http://listserv.gva.es/cgi-bin/mailman/listinfo/gvsig_internacional
_______________________________________________
Gvsig_internacional mailing list
Gvsig_internacional at listserv.gva.es
http://listserv.gva.es/cgi-bin/mailman/listinfo/gvsig_internacional
-----Inline Attachment Follows-----
_______________________________________________
Gvsig_internacional mailing list
Gvsig_internacional at listserv.gva.es
http://listserv.gva.es/cgi-bin/mailman/listinfo/gvsig_internacional
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://listserv.gva.es/pipermail/gvsig_internacional/attachments/20100717/aebbcdd4/attachment.htm
More information about the Gvsig_internacional
mailing list