orthorectification/warping w/in MI or MI Tools would be a blessing beyond
ability to reward.....

As all processes to complete achieve this goal using MI are long and
complicated utilizing additional software such as ER Mapper or CHIPS and/or
Photoshop.

Regards
Roy

----- Original Message -----
From: "Greg Phillips" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Wednesday, October 29, 2003 3:40 AM
Subject: Re: MI-L Distortion of scanned map


Hi everyone - as this is my first post to MI-L I should apologize for the
imminent ramble.

I have several years experience warping raster historic maps (pre 1936 OS
plus other individual surveys) for use in various GIS, converting their map
projections to the current British National Grid.  Having started in MapInfo
and for the reasons outlined by Jaromir found it was not achievable, so as
an alternative I used ER Mapper to georeference and warp the raster images
for use in MapInfo.

Having moved employer I was again required to carry out the same tasks in
ArcView, and used a very good (and possibly free?) extension called
Imagewarp to georeference custom historic map tiles.  You can then composed
an image mosaic of the historic map in ER Mapper.

I would recommend this as the easiest method in my experience as you can
revisit the mosaic as and when you get time to add new tiles to it, and all
the georeferencing is carried out in one application (ArcView).  Obviously
there are many different capable software applications capable of this kind
of operation - some GIS and other Image Processing packages.

Whether warping images for use in MapInfo or ArcView, ER Mapper has the
advantage that if you are working with large maps (one I am thinking of was
a double A0 size, hand painted original from 1829AD) and have a smallish
scanner you can still scan tiles and georeference them separately, ensuring
relatively good registration accuracy per tile, then you can seamlessly
mosaic these images together, drawing a clipping polygon on each tile to
remove edge matching problems with null pixels overlying adjoining tiles,
and then write a compressed .ecw file out for use in either MapInfo or
ArcView.

Its a pretty complex multi step process but if you need any help or advice
please feel free to email me off list.

At a MUG UK demo of Vertical Mapper recently I raised the question as to
whether orthorectification and rubber sheeting might be added to VM as a
future development, as it is the logical Raster extension to MapInfo.  In my
opinion that added functionality would really extend the tools available to
exclusive MapInfo users, and it seems that having acquired Vertical Mapper
that MapInfo are in the perfect position to enhance this product.

/ramble OFF

Greg Phillips


--------------------------------------------------------------
Greg Phillips BA PGDip MIFA
Environment GIS Consultant

Built and Natural Environment
Northamptonshire County Council
PO Box 163
County Hall
Northampton
NN1 1AX

Tel: 01604 237433
Mob: 07799 832584
--------------------------------------------------------------

>>> Jaromir Svasta <[EMAIL PROTECTED]> 10/29/03 06:13am >>>
MapInfo only adjusts the coordinate system of vector layers to fit the
raster. It applies one transformation formula (I think bilinear) all over
the area, so if the raster is distorted it cannot guarantee that control
points will match the raster, doesn't matter how many of them there are.
You need to rubber-sheet the raster. MapInfo won't help you :-( Arc Info,
Grass and some other "heavy" GIS can make it.

Jaromir

On Tue, 28 Oct 2003 23:50:35 +0100, Joep Orbons <[EMAIL PROTECTED]> wrote:

> Dear readers,
>
> I process historical maps and fit them to todays topography. The problem
> with these historical maps is the distortion of the maps. This
> distortion is not regular in X and/or Y direction, it can be very much
> rotated and scaled different in seperate parts of the maps.
> These maps always have some points I recognize and that I can mark on
> todays topography. Sometimes many points, sometimes only a few.
> How can I un-distort these maps, scanned or digitised, to fit the known
> points? I don't think it is possible in MapInfo. Maybe I need a seperate
> program with two layers: One layer with todays topography and known
> points and a second layer with the scan. Then drag points of the scan to
> the modern topography where the rest of the map follows. The more
> points, the more accurate.
> Any suggestions?
>
> Thanks,
>
> Joep Orbons
> Souterrains: [EMAIL PROTECTED]




----------------------------------------------------------------------------
----


> ---------------------------------------------------------------------
> List hosting provided by Directions Magazine | www.directionsmag.com |
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
> Message number: 8932


---------------------------------------------------------------------
List hosting provided by Directions Magazine | www.directionsmag.com |
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Message number: 8936

Reply via email to