[Manifold-l] Custom datums

[Martin Roseveare]

To provide a bit of context to this discussion we are usually dealing with 
projects with dimensions of perhaps only a few hundred metres, at 
resolutions in decimetres and yes, we are surveyors. The problem arises when 
GPS long lat is used in conjunction with OSGB36 which results in errors of 
several metres relative to digital mapping from Ordance Survey who use the 
more accurate OSTN02 transformation.

The big difference between Manifold's transformation and OSTN02 is that the 
latter goes via ETRS89 and adds an offset that varies in size accordingly to 
your location within the grid. This offset is essentially a measure of the 
local error in the original OS mapping and allows 'true' co-ordinates from 
GPS to be mapped into OSGB36.

Martin Roseveare
Senior Geophysicist

Phone: +44 (0) 1989 730 564
Fax: +44 (0) 7050 369 790
Web: www.archaeophysica.co.uk

This e-mail and any attachment may contain proprietary or confidential 
material and be subject to legal privilege. If you have received it in error 
please promptly erase it from your system and inform the sender. It should 
not be copied, disclosed to, retained or used in any way by any other party.

Nothing in this e-mail should be regarded as representing a binding contract 
unless explicitly stated otherwise. All opinions are those of the sender and 
not necessarily ArchaeoPhysica Ltd.
----- Original Message ----- 
From: "F Whiteley" <franciswhiteley at btconnect.com>
To: "'Martin Roseveare'" <m.roseveare at archaeophysica.co.uk>; 
<adamw at manifold.net>; <manifold-l at lists.directionsmag.com>
Sent: Thursday, September 21, 2006 11:49 AM
Subject: RE: [Manifold-l] Custom datums


>>> centimetre resolution...
>>> Is there a wider need in the UK community for this?
>
> Amongst surveyors yes, but amongst GIS community unlikely.
>
> I certainly have had no use so far, however, in some arguments over
> coordinate transformation this could become an issue of client confidence
> rather that any necessity.
>
> Our objects are between 100m and 10km apart but quoted in GPS to various
> decimal places, UTM or GK generally rounded to 1m but sometimes 1cm.
> Accuracy has been varied and unverifiable.  We generally only re-project 
> to
> put things into context.
>
> So far Manifold has always won any transformation arguments, but we
> generally have worked at 25m resolution which matches sat images we have
> used in the past.
>
> Now that 1m resolution sat and aerial images appear to be available via
> Google and other servers, clients may expect their data to match exactly 
> and
> sub 10m accuracy may become important.
>
> I doubt that the error that Adam estimates is as big as this - I am sure
> someone working with CAD/GPS/OSGB/UTM/GK and differing agencies would have
> noticed earlier, but then again if you work with only one datum then you
> won't be aware of the problem.
>
> There are free transformation programs available so you can check (see 
> other
> threads).
>
> I have assumed that height above datum is not important at 1m accuracy.
>
> Francis
>
>
>
>