[Openstreetmap-dev] Re: [Openstreetmap] Landsat update

Nick Whitelegg nick at hogweed.org
Sat Apr 30 19:18:40 BST 2005


On Saturday 30 Apr 2005 18:44, you wrote:
> * On 30-Apr-2005 at 10:32AM PDT, Nick Whitelegg said:
> > > no, apart from the usual GPS error factor it shouldn't make a
> > > difference.
> > >
> > > > i.e. it uses Airy coordinates even though the GPS uses WGS84.
> > > > to screen display, the error becomes worse!
>
> I knew this sounded strange, since JPL's Landsat tiles have in my
> experience mostly been georectified quite well. So it still sounds
> like you have a projection problem. Have a look at this:
>
>     $ proj +proj=utm +zone=30 +ellps=airy
>     0 51.5
>     708196.62       5709273.47
>     $ proj +proj=utm +zone=30 +ellps=WGS84
>     0 51.5
>     708216.87       5709696.98
>
> So, according to PROJ.4, 51.5N, 0E differs when projected into UTM on
> the WGS84 ellipsoid versus the Airy ellipsoid by about 424 meters.
> That would account for the displacement you observe. My sincerest
> recommendation is to ditch the OSGB coordinates entirely, since this
> is where your problem is coming from. If that's undesirable or
> infeasible, I would be glad to offer help integrating your app with
> PROJ.4.

Thanks, will have a look at PROJ.4 to see how well the API would fit in with 
my existing code.

BTW One other experiment I've just tried is converting the GPX coordinates to 
Airy as well as the transformation from lat/lon to screen - so everything is 
consistently in Airy. I get exactly the same offset error across the board as 
I did when I converted everything to WGS84.

In other words the erroneous mixture of WGS84 and Airy gave me the least bad 
error while pure WGS and pure Airy both give me the same (worse) error.

Nick




More information about the dev mailing list