roger at opq.se
Thu Jan 12 15:44:35 EST 2006
On Thu, 2006-01-12 at 17:19 +0000, Paul Kelly wrote:
> Hello Roger
> On Thu, 12 Jan 2006, Roger Oberholtzer wrote:
> > This time, I am testing OSGB36 (UK). My current method uses geotrans. I
> > am using a 7 parameter transform (in geotrans-speak). When I convert
> > WGS84 lat/longs into OSGB36 X/Y, the difference from the accepted
> > location is less than 10 meters (which is an issue for a future
> > posting). So, I thought I would see how proj4 compares. Once again, my
> > brain is nowhere in sight. I get location errors that are much greater.
> > Usually over 90 meters.
> > I am trying this:
> > cs2cs +proj=latlong +ellps=WGS84 +to +proj=tmerc +lat_0=49 +lon_0=-2
> > +k=0.9996012717 +x_0=400000 +y_0=-100000 +ellps=airy +units=m +no_defs
> > These settings are based on the esri file and look reasonable compared
> > to the values I use in geotrans.
> You need to do a datum transformation (WGS84 to OSGB36)
> With the following command line
> cs2cs +proj=latlong +ellps=WGS84 +towgs84=0,0,0 +nodefs +to +proj=tmerc
> +lat_0=49 +lon_0=-2 +k=0.9996012717 +x_0=400000 +y_0=-100000 +ellps=airy
> +towgs84=446.448,-125.157,542.060,0.1502,0.2470,0.8421,-20.4894 +units=m
> I get these results
> 620286.26 309466.38
> for your test point. Much closer.
This was an improvement. The difference for this point it 4.376 meters.
> That parameter set is used in GRASS and the comment with it indicates an
> accuracy of 5 metres. For anything better you're likely to need to use
> some of the gridshifting data available at http://www.gps.gov.uk/ (not yet
> in a suitable format for use with PROJ.4)
TRL (Transport Research Lab) point me to GridInQuest, which is a
Windows-only implementation of what you are refering to. I have
contacted them and they say they will not do a Linux version. I believe
it is work paid for by public funds. But the implementation is not
distributed. The application does run on Linux/Wine, but that does not
really help me.
Using geotrans, I get differences up to 15 meters, with 10 seeming more
the average. I think proj may be better. But, unfortunately, not as good
as I may need in the UK, which seems to be going to OSTN02. As you no
doubt already know.
More information about the Proj