[Proj] Dutch correction grid
j.l.h.hartmann at uva.nl
Thu Apr 7 04:57:24 EST 2011
Ah, that's where they hid it! Almost all information on that site has to
be bought, they only give pointers to what they have and where you can
buy it. Even the grid coordinates themselves are not public, I believe.
They probably thought the correction grid data were so useless they
could put them safely on the Internet without losing any mony on it.
Well, that's what Open Source people live from, don't you think so? :-)
BTW This indirectly solves one of the major problem I have been
struggling with for years, the exact georeferencing of the Topographical
Maps of the Netherlands from 1850 to 1930. They are in a different
projection (Bonne), and when you use the official transformation
formulas, the resulting maps are about 60 meters off. No one seems to
have noticed it before; it has to do with the slightly smaller ellipsoid
the used for that projection, and you can't see it really on maps at
scale 1:50000, unless you scan them at the very high resolutions
available only since a few years. The problem is that the latitude value
of the zero-coordinate (near Breda) (51.30 degrees north, the middle of
the Kingdom of the Netherlands as it then was, together with Belgium and
Luxemburg) was computed not from stellar measurements on the spot, but
from measurements taken from the Westertoren in Amsterdam (52°22472).
From that point they computed the sperical distance to 51.30 degrees,
but using a small ellipsoid, expecially computed for the Netherlands by
an engineer called Ackermans. This was before Bessel computed his
ellipsoid for Germany that was afterwards also used for the new Dutch
projection. So they got about 60 meters too far to the south, and the
PROJ-string for the Bonne projection has to be adapted with a slightly
different central latitude.
This takes away most of the errors, but not all of them. There remains a
small residual error of about 10 meters, some of it random, but some of
it systematic in the east-west direction: the farther east or west from
the zero-point, the greater the error. I haven't found the reason for
this (did they use a slightly smaller meter, specially adapted for the
Netherlands? Dutch bureaucracies have been known to think up stranger
things). It isn't difficult to correct this with some GCP-based rubber
sheeting with gdalwarp, since I do have the coordinates for about 1000
trinagulation points (mostly church towers) both in the old and the new
coordinate systems in 1850 and 1929, but it remains a makeshift solution
(both literally and figurative).
But there remained a major difficulty: to get an exact georeferenced old
map, e.g. over Google Satellite, I had to take a few steps that I
couldn't base on a mathematical or systematic rule. I just used two
rules of hand (translate 60 meters and stretch a bit). Sixty meters is a
big distance when you try to locate old parcel boundaries or building
foundations, and the problem is that there is a lot of money going round
in this country on that scale (protected buildings, environmental
protection, heritage subsidies, agrarian reconstruction etc). When those
old boundaries are not reconstructed on a completely objective basis,
and without completely reproducible means, you can expect lots and lots
of judicial procedures, and I would not like to be caught in the middle
of that mess, being, as far as I know, the only one to have some
expertise in the problem.
Now, however, I can create a correction grid file from the two series of
church towers coordinates, and use that with the official transformation
formula to correct the 60 meters error. Perhaps there will be
discussions even about that (church towers that have been translated a
few meters or so, etc.), but the method itself is unambiguous, and can
be understood be judges and lawyers alike, without the need to explain
much about ellipsoids and PROJ parameters. Takes really a load off my mind.
So in the long term it pays to give some attention to an Open Source
mailing list, even if the subjects are not directly relevant to what you
doing or paid for at the moment.
On 7-4-2011 11:03, Bart van den Eijnden (OSGIS) wrote:
> Hi Jan,
> great! Thanks in advance.
> I got the spreadsheet from:
> Best regards,
>> If someone could give me some pointers to the grid shift format and the
>> way it is used in PROJ, I will have a go at producing a grid shift file
>> from this spreadsheet. Where did you get it from, Bart?
>> On 7-4-2011 9:18, Bart van den Eijnden (OSGIS) wrote:
>>> did anybody ever attempt to transform the information from the Dutch
>>> correction grid into a format for a datum shift grid file that PROJ.4
>>> Best regards,
>>> Proj mailing list
>>> Proj at lists.maptools.org
More information about the Proj