Projection trouble

[Gayathri Swaminathan]

Chris,
We had the exact same problem when dealing with WMS on a raster
layer...back in 2003 end sometime

A white offset appeared when projecting raster layer over our shape
layer...and we tracked it down to this file mapwmslayer.c at this
function, "msDrawWMSLayerLow" function (used to be just
msDrawWMSLayer....I think..)

The resolution on the float .12f on the map cell size when
reprojecting was not working for us.,. we had to increase
resolution..and then recompile mapserver...and the white bar went
away...

Gayathri


On Tue, 4 Jan 2005 20:09:13 +0100, Arnulf Christl
<arnulf.christl at ccgis.de> wrote:
> Hi guys,
> this one is still open. The answers until now didn't help, everything
> suggested works perfectly fine. No one out there with an idea?
>
> Its the same data with the same projection in the same client with the
> same request, only one is a mapserv401, the other a mapserv441. In
> EPSG:4326 both maps fit on top of each other perfectly, in EPSG:31466
> they show an offset. You can have a look at it here:
>
> http://www.mapbender.org/testbed.html
>
> Zoom in to see the post code areas, you'll understand what I mean. Find
> the tech details of the installation at the end of this mail.
>
> I suspect PROJ4 of doing things differently in 401 and 441.
>
> Many thanx,
> Arnulf.
>
> Bart van den Eijnden wrote:
> > Hi Arnulf,
> >
> > just checking a thing which maybe could cause displacement.
> >
> > Does your Mapbender client use the exact same ratio of pixel coordinates
> > (width versus height) to world coordinates (maxx-minx versus maxy-miny).
> > Ofcourse there were changes in Mapserver WMS between 4.0.1 and 4.4.1 which
> > can cause non square pixels if the ratio differs, which maybe could cause
> > this ....
> >
> > Just a thought.
> >
> > Best regards,
> > Bart
> >
> > On Sun, 2 Jan 2005 13:17:03 +0100, Arnulf Christl
> > <arnulf.christl at CCGIS.DE> wrote:
> >
> >> Hi,
> >> sorry I postet a wrong link. The problem can be seen here:
> >> http://www.mapbender.org/testbed.html
> >> I just added a box where the projection can be changed on the fly. Zoom
> >> in to below 1:50000 and switch to WGS84 and everything looks fine, the
> >> geometries from both servers overlap perfectly. Switch to GK3
> >> (EPSG:31467) and there is that offset.
> >>
> >> Arnulf.
> >>
> >> Arnulf Christl wrote:
> >>
> >>> Hi,
> >>> we experience a strange offset when overlaying the same data as WMS maps
> >>> with MapServer versions 4.0.1 and 4.4.1.
> >>>
> >>> *4.0.1 server*
> >>> Windows 2000: SP4
> >>> IIS: 5
> >>> proj: came with 4.0.1
> >>> gd: came with 4.0.1
> >>> gdal: came with 4.0.1
> >>> Cap URL:
> >>> http://wms.ccgis.de/umn/bin/mapserv.exe?map=d:/umn/germany.map&SERVICE=WMS&REQUEST=GetCapabilities&VERSION=1.1.0
> >>>
> >>>
> >>> Data: local shape files in EPSG:4326 projected to EPSG:31467
> >>>
> >>> *4.4.1 server*
> >>> FreeBSD: 4.10
> >>> Apache: 1.3.33
> >>> proj: 4.4.9
> >>> gd: 2.0.33
> >>> gdal: 1.2.1
> >>> Cap URL:
> >>> http://wms1.ccgis.de/cgi-bin/mapserv?map=/data/umn/germany_demo/germany_demo.map&VERSION=1.1.1&REQUEST=GetCapabilities&SERVICE=WMS
> >>>
> >>>
> >>> Data: the same shape files as above (WGS84) but dumped to
> >>> PostgreSQL/PostGIS (and also projected to EPSG:31467).
> >>>
> >>> The problem does not seem to result from MapServer but from proj?
> >>>
> >>> You can have a look at the problem here:
> >>> http://www.mapbender.org/demoserver.html
> >>>
> >>> Zoom in to a scale of 1:100,000 or less to see the post code areas. They
> >>> overlap with an x-offset (north):
> >>> in northern Germany 190m
> >>> in central Germany 145m
> >>> in southern Germany 100m
> >>>
> >>> The y-offset (east) is:
> >>> in western Germany 45m
> >>> in eastern Germany 125m
> >>>
> >>> That resembles a lot to a datum shift which we stumbled accross (years
> >>> ago) when reading GPS EPSG:4326 WGS84 data directly into the database
> >>> and overlaying it with Gauss-Kruger EPSG:31467 (Bessel projection).
> >>>
> >>> What has been changed and what impact does it have on projects which
> >>> overlay other data?
> >>>
> >>> Best regards, Arnulf.
> >>>
> >>
> >
> >
> >
> > --
> > Using Opera's revolutionary e-mail client: http://www.opera.com/m2/
> >
>
> --
> ------------------------
> Arnulf B. Christl
> ------------------------
> http://www.ccgis.org
> http://www.mapbender.org
> ------------------------
>