Charles,

By pixel coordinates, one should understand the pixel positions (row, column
from the upper left corner). This technique assumes that registration points
keep the same (geographic) coordinates but refer to new positions in the
pixel space through a "scaling" factor (in your example 0.5).

One should be aware that the results (i.e. the adjustment of the image to
the map, measured by the average errorof the fit) may be less good that
before because of the roundings that may occur in the scaling operation of
the pixel coordinates. And that does take into account the added lack of
precision of the image itself, resulting from identical rounding/averaging
operations in the change of definition.

All together, it is not a recommandable operation if anyone is concerned
with precision. But if it is just for having a less bulky map background,
why not.

Jacques Paris
============================================================
e-mail        [EMAIL PROTECTED]
   alternate   [EMAIL PROTECTED]

gis activity (MapInfo mainly)
      http://www.total.net/~rparis/gisproducts.htm
============================================================

> -----Original Message-----
> From: [EMAIL PROTECTED]
> [mailto:[EMAIL PROTECTED]]On Behalf Of Charles Huyck
> Sent: July 19, 1999 3:06 PM
> To: [EMAIL PROTECTED]
> Subject: RE: MI Decreasing Raster Image Resolution
>
>
>
>
> If you decrease the size of an image in a separate program by, say, 50%,
> you might be able to open the .tab in a text editor and multiply all the
> pixel coordinates by 0.5. Of course, make back-ups and check your work in
> the registration dialog.
>
>
> ----------------------------------------------------------------------
> 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]

Reply via email to