[GRASSLIST:2128] Re: New Location

[Markus Neteler]

On Fri, Jan 02, 2004 at 10:10:40AM +0000, Paul Kelly wrote:
> On Fri, 2 Jan 2004, Markus Neteler wrote:
> 
> > Most unreferenced maps are imported with *negative* XY coordinates
> > when using r.in.gdal. This behaviour should be discussed
> > and probably changed. IMHO positive XY coordinates should be
> > written except for AVHRR (as far as I remember, maybe also SAR).
> 
> Are both the X and Y co-ordinates negative? 

Yes, but for some formats I have already fixed it and changed to
positive XY coordinates. The r.in.gdal code needs a cleanup IMHO,
at least it is fairly complex now to read.

> If so it doesn't seem to make
> sense. It would make sense to me if only the Y co-ordinates were negative,
> as then the absolute y co-ordinate of the centre of each pixel would be
> the same as its row number (in the conventional way of numbering pixels in
> image processing). This is maybe more familiar to people like me who
> currently do most image processing in MATLAB.

For me simply positive coordinates were fine (no MATLAB user).
> 
> But I don't really know how this works in practice and if it's hard to get
> your head around while actually doing image processing in GRASS. If you
> were just temporarily importing into the XY location before rectifying
> then I don't see that the XY co-ordinate system would make much difference
> at all.
> 
> Probably there is another disadvantage that I have missed?

As far as I understand the problem it's a matter of convention.
So changing the behaviour should be ok (and avoid the many user mails
here where they import and don't find the image later).
Maybe Frank has a comment for us?

Markus