[Proj] using an NTv2 grid to apply a slight shift to coordinates in a metric CS

[Andre Joost]

Hi,

the nadgrid shift is usually done with lat/lon coordinates. You 
transform your cartesian coordiantes to lat/lon, apply the shift, and 
transform into another grid.

The syntax of the proj string is:
+proj=tmerc +lat_0=0 +lon_0=9 +x_0=3500000 +y_0=0 +k=1.000000 
+ellps=bessel +units=m +nadgrids=path/to/BETA2007.gsb +wktext

The grid file has to be in binary format, which can be transformed from 
an ASCII file.

For the German grid, information can be found here:
<http://crs.bkg.bund.de/crseu/crs/descrtrans/BeTA/de_dhdn2etrs_beta.php>

Most other users of NTv2 should have published similar documentation.

HTH,
André Joost



Am 03.01.2013 15:03, schrieb Maciej Sieczka:
> Hello,
>
> I have a script which takes an x,y pair of metric coordinates in
> EPSG:2174 (http://www.spatialreference.org/ref/epsg/2174/proj4) and
> returns them slightly shifted, in the same coordinate system.
>
> The difference between the input and output coordinates varies from 0
> to about 60 cm, non-systematicaly.
>
> I would like to be able to "reproject" my GIS data from EPSG:2174 to
> EPSG:2714 + the shift described above, and the other way round; using
> GDAL/OGR, GRASS, QGIS reprojection tools.
>
> Can anyone tell me if using a shift grid for that would make sense? I
> have a general idea how to create one from my scripts output, but there
> is something missing in my thinking. Let me put it this way - should my
> grid provide shifts between:
>
> a) the original EPSG:2174 and the EPSG:2174+shift
>
> or rather
>
> b) between WGS84 and the EPSG:2174+shift?
>
> And what would be the proj4 string for the new system in either case?
>
> Another thing is if it would be possible to utilise a similar grid to
> reproject between EPSG:2714+shift and other coordinate systems covering
> the same territory (e.g. EPSG:2180), again using GDAL etc. tools.
>
> Ideally, I would like to achieve both goals with just single shift
> grid.
>
> I'm afraid I might be asking my questions wrong way, but hopefully you
> get what I'm aiming at. I'd be grateful if you can correct my errors and
> give me some guidance here.
>
> Regards,
> Maciek
>