[Proj] Proj 4.9.1RC3 Released
Even Rouault
even.rouault at spatialys.com
Sat Feb 21 15:49:41 EST 2015
Le samedi 21 février 2015 20:43:45, Sebastiaan Couwenberg a écrit :
> On 02/21/2015 08:22 PM, Hermann Peifer wrote:
> > On 2015-02-21 18:31, Howard Butler wrote:
> >> All,
> >>
> >> Proj4 4.9.1RC3 is now released for testing. Should no significant issues
> >> be identified, ...
> >
> > (...)
> >
> >>> o Regenerate nad/epsg with GDAL r28536 to avoid precision loss in
> >>> TOWGS84
> >>>
> >>> parameters, e.g. on Amersfoort / RD EPSG:4289 (#260)
> >
> > No sure if this is significant, but instead of a precision loss for
> > TOWGS84's X,Y,Z-offsets, there is now a "precision gain" for the
> > rotation values in several cases, e.g. for the mentioned Amersfoort / RD
> > EPSG:4289. I wonder if this is as it should be.
>
> The Dutch Kadaster raised the issue about the imprecise +towsg84 values
> for EPSG:28992 on the dutch OSGeo list. See the link included in #260.
>
> https://trac.osgeo.org/proj/ticket/260
> http://lists.osgeo.org/pipermail/dutch/2014-October/000980.html
>
> The correct values specified by Kadaster use 6 decimals max:
>
> +towgs84=565.4171,50.3319,465.5524,-0.398957,0.343988,-1.877402,4.0725
>
> This used to be:
>
> +towgs84=565.417,50.3319,465.552,-0.398957,0.343988,-1.8774,4.0725
>
> And is now:
>
> +towgs84=565.4171,50.3319,465.5524,-0.398957388243134,0.343987817378283,-1.
> 87740163998045,4.0725
>
> The "precision gain" is better than what we had before, but it may also
> make sense to limit the number of decimals in GDAL a bit.
GDAL nows takes the TOWGS84 components as the exact strings found in the CSV
file produced by the libgeotiff process.
So if truncation should be done it would be in the importer from the EPSG
database in libgeotiff. The many decimals likely come from the conversion from
milliradians (since it seems that the unit used by the EPSG database) to
whatever unit is used in towgs84. They are likely not harmful.
>
> Kind Regards,
>
> Bas
--
Spatialys - Geospatial professional services
http://www.spatialys.com
More information about the Proj
mailing list