[GRASS-user] Georectify problem

[John Stevenson]

Moritz Lennert wrote:
> On 30/07/08 02:00, Adam Dershowitz wrote:
>> The problem that I am running into is that only the middle half or so 
>> of the jpg is being rectified.  
>
>> Do I have to set some region or something to get it all to work?
>
> Yep, exactly.
>
> AFAICT, the georectify tool calls i.rectify with the -c flag (line 
> 1366 of gui/tcltk/gis.m/georect.tcl), which means "Use curr. region 
> settings in target location". So you need to set the correct location 
> in the target location.
>
> I'm not sure I understand why this is so. It would seem to be more 
> intuitive to run without that flag and thus have i.rectify "determine 
> the smallest window which covers the image", because otherwise the 
> user has to identify target region extents manually...
>
> Michael, any special reason for using -c ?
>
> Maybe this could be made into an option ?
I went through exactly the same problem this week on my first attempts 
on the georectify tool, and also wondered why it it chose those region 
settings.  I updated the wiki page last night to highlight the need for 
a suitable region.

Also, sometimes I found that it crashed when the target region was 
completely wrong because one of the intermediate stages was taking s=0 
w=0 and I was working in UTM with typical values of 400000 and 7100000 
and a resolution of 2 m.  The resulting map must have been huge.

In the end I calculated the region that I would need manually and worked 
out the resolution from the number of pixels in the scanned image.  Does 
i.rectify calculate these, or if not, is that something that could be 
easily added?

Later

John