I've had some problems in this respect, where multiple images have been
tiled. I did some on-screen digitising where two images overlapped and found
that dpending upon which image I was looking at, the location of the points
varied. It was very difficult to get the two images matching exactly, and I
had to live with the resulting discrepancy in the vector locations taken
from them.
A question that has bothered me however; if vectors are stretched to fit the
raster, what impact does this have on distance and area functions. If I
measure distance off a map that has a raster image underlying it, I presume
the vector has probably been streched, and the resulting measurement is
likely to be incorrect? Would this also apply to functions like buffer; if I
buffer a point that is underlain by a rester image, will the buffer distance
be calculated correctly?
Regards
Peter Walsh
Dames & Moore
email: [EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>
Ph. +61 (2) 9955 7772
> -----Original Message-----
> From: Gary [mailto:[EMAIL PROTECTED]]
> Sent: Friday, 14 May 1999 12:32
> To: Gerald Kornelson
> Cc: [EMAIL PROTECTED]
> Subject: Re: MI re Bitmap
>
>
> Actually, when you register an image, you are definining an
> affine transformation that is applied to all of the
> vector layers so they match up with the raster image. So,
> what you've experienced is "correct", at least
> according to the way MI Pro works. What I'd like to know is
> how MI Pro handles multiple images with different
> registrations, say for tiling purposes. It sure would make
> more sense for the transformation to be applied to the
> raster, not the vector layers.
>
> Gary.
>
> Gerald Kornelson wrote:
>
> > I recently received an air photo of our entire city in .jpg
> format. It was rotated about 20 degrees off from
> > tru north-south.
> >
> > I opened it in MI, registered it so that the error
> indicators seemed very close, and saved it.
> >
> > If I open it and any other MI table (ie. image as a
> "backdrop") the whole bunch is off by the same +/- 20
> > degrees. The vector map is rotated to match the correct
> location on the raster instead of the
> > raster being rotated into correct UTM coordinates. Doesn't
> matter which I open first, vector or raster.
> > Isn't the first table supposed to control the projection of the map?
> >
> > I took the image into Photo Editor and rotated it about 20
> degrees. It looks OK, but I cannot be sure
> > that the actual rotation is correct.
> >
> > Isn't registering a raster supposed to "re-orient" the
> raster so it displays the raster pixels in the location
> > that were assigned during the registration process?
> >
> > Any ideas?
> >
> > Gerald Kornelson
> > City of Winnipeg
> >
> ----------------------------------------------------------------------
> > To unsubscribe from this list, send e-mail to
> [EMAIL PROTECTED] and put
> > "unsubscribe MAPINFO-L" in the message body, or contact
> [EMAIL PROTECTED]
>
> ----------------------------------------------------------------------
> To unsubscribe from this list, send e-mail to
> [EMAIL PROTECTED] and put
> "unsubscribe MAPINFO-L" in the message body, or contact [EMAIL PROTECTED]
>
----------------------------------------------------------------------
To unsubscribe from this list, send e-mail to [EMAIL PROTECTED] and put
"unsubscribe MAPINFO-L" in the message body, or contact [EMAIL PROTECTED]
