Re: [Gimp-user] Putting together scanned parts of a map

2001-05-14 Thread Torben Linde


- Original Message - 
 I have scanned a map with my A4 scanner, but as the map is quite large,
 it did take several scans. Now I try to put the thing back together.
 The identification part is simple, because the map does contain a rectangular
 grid, so that cutting on a grid line is possible. I can also rotate, so
 that the grid is looking rectangular, but:
 * the scanner (a not very expensive Microscan model) is not working completely
   parallel, so that when the left and lower boundary are fine, the right
   boundary is changing by 10 pixels to the left (on 3000 for the height).
   How can I best deal with the distortion?
 * Once I have these rectangular pieces, how do I best put them together
   to one large map. By now I have scanned with 300dpi and the grid lines
   are at a disctance of 2cm, so I can calculate at which absolute positions
   I have to put the parts together, but I did not find a menu to put the
   data into...
   (I know that 300dpi is a lot, but I rather scan with high resolution and
   compact it later...)
 I do have a Duron 800 @ 256MB main memory, so that I can play around a bit
 with large graphics as needed.
 
 Any ideas?
 Lutz
---

Are you sure that the map-grid really is orthogonal ?

Any map of the earth is a projection of a globe onto a plane, and that is impossible 
to 
do accurately for simple geometric reasons, there will always be small deficiences in 
the grid requiring a bit of cheating along the edges.

Torben

___
Gimp-user mailing list
[EMAIL PROTECTED]
http://lists.xcf.berkeley.edu/mailman/listinfo/gimp-user



Re: [Gimp-user] Putting together scanned parts of a map

2001-05-14 Thread Lutz Jaenicke

On Mon, May 14, 2001 at 11:28:55PM +0200, Torben Linde wrote:
 Are you sure that the map-grid really is orthogonal ?
 
 Any map of the earth is a projection of a globe onto a plane, and that is impossible 
to 
 do accurately for simple geometric reasons, there will always be small deficiences 
in 
 the grid requiring a bit of cheating along the edges.

Yes, it is not the latitude/longitude grid. The swedish grid (based on
Gauß-Krüger) is similar to the UTM grid, see
  http://www.bacher.de/abc/g.htm#GaußKrügerGitter
  http://www.maptools.com/UsingUTM/index.html
  http://www.joe.mehaffey.com/utminfo.htm
Of course, such projection is only valid within certain limits:
  http://www.colorado.Edu/geography/gcraft/notes/coordsys/gif/utmzones.gif
(sorry if these links are off topic for graphics, but they are related
to my problem :-)

Therefore: yes, this thing _should_ be rectangular...

Best regards,
Lutz
-- 
Lutz Jaenicke [EMAIL PROTECTED]
BTU Cottbus   http://www.aet.TU-Cottbus.DE/personen/jaenicke/
Lehrstuhl Allgemeine Elektrotechnik  Tel. +49 355 69-4129
Universitaetsplatz 3-4, D-03044 Cottbus  Fax. +49 355 69-4153
___
Gimp-user mailing list
[EMAIL PROTECTED]
http://lists.xcf.berkeley.edu/mailman/listinfo/gimp-user