[Mapserver-users] rubber sheeting

[Jan Hartmann]

The problem is that AFAIK all these rectifying routines are computed 
globally. You can use of course a higher order polynomial, but even so 
the placement of every control point will affect the whole map. 
Moreover, you can't use very high polynomials and too many control 
points. What I am looking for is something that will do *local* 
transformations based on nearby control points, and stretch the results 
together some way. Those old maps are generally correct on small 
subregions, but the subregions fit in no way together. There is 
certainly no numerical soluton for it.

I used GRASS to rectify the old map I showed at the MUM (online next 
week). When you use one set of control points, you get an unbelievably 
correct map for one part of the town, and a distorted one for other 
parts. If you put your control points in some other part, that one will 
get georeferenced correctly, and the first one is off. If you put 
control points everywhere, distortions will be averaged out, no matter 
what kind of polynomial you use. That is not what I am looking for. 
Rubber sheeting is a dirty way of doing the job; it will never give a 
perfect match for the whole map, but it gives something to work with.

ArcInfo does something like this, but I don't know how they do it, and I 
would like to implement this for local history research here in 
Amsterdam. Many people would love to look up matching points on old maps 
as a hobby, and would be able and willing to contribute to a 
reconstruction of the whole map in a geometrically correct way. It is 
just too bad that we can't afford ArcInfo in every household :-) . 
That's why I'm looking for a public domain solution. Again, it is a 
difficult problem and probably requires more fundamental research. I 
could use some pointers on this from more mathematically oriented people.

Jan

Lowell Filak wrote:
> Being you already know GRASS does that mean i.rectify won't do the job?
> Lowell
> 
> The following message was sent by Jan Hartmann <jhart at frw.uva.nl> on
> Fri, 18 Jul 2003 15:06:59 +0200.
> 
> 
>>Hi list,
>>
>>I'm looking for some public domain rubber sheeting transformation 
>>programs or algorithms. I have some historical maps that are distorted 
>>differently in different parts of the map. I can find lots of control 
>>points in contemporary exact maps, and I would like to transform the old 
>>ones as accurately as possible into the modern coordinate system by 
>>rubber sheeting. This can be done in ArcInfo, but is there something 
>>available in the public domain? An algorithmic description would be a 
>>second best solution; I could probably program something around it (and 
>>perhaps put it into MapServer, Proj, or Gdal).
>>
>>Thanks,
>>
>>Jan
>>
>>_______________________________________________
>>Mapserver-users mailing list
>>Mapserver-users at lists.gis.umn.edu
>>http://lists.gis.umn.edu/mailman/listinfo/mapserver-users
> 
> 
>