[GRASS-windows] Mosaic and georectify commands in GRASS GIS 6.3.0

[Glynn Clements]

Peta.Wright at csiro.au wrote:

> I am a new user of GRASS GIS 6.3.0, and have been experiencing a few
> problems with some of the commands. Specifically with the mosaic and
> georectification tool. I have been trying to mosaic 2 rasters together
> however I keep receiving the following error message:
> 
> 
> Mosaicing two images...
> syntax error, unexpected '@'<mailto:'@'>, expecting '='

You should check whether your mail program has an option to disable
that <mailto:...> "feature". It's could be problematic on the GRASS
lists, given the way that GRASS uses "@".

> cat: P:/CSE/_SustainableRange/Hymenachne/Peta's: No such file or directory
> cat: folder/Sample_Images//colr/2518_2006CK at PERMANENT: No such file or directory

Several of the shell scripts (and a small number of binary modules)
fail if either the GRASS database directory (aka GISDBASE) or the
GRASS installation directory (aka GISBASE) contain spaces. Actually,
the single quote (apostrophe) character might also cause problems.

Many of these have been fixed since 6.3.0 was released (6.3.0 is the
first version which has been widely used on Windows). You can obtain
more recent versions of the scripts from the Subversion repository:

http://trac.osgeo.org/grass/browser/grass/branches/develbranch_6

For shell scripts, you can just obtain the updated script, e.g.:

http://trac.osgeo.org/grass/browser/grass/branches/develbranch_6/scripts/i.image.mosaic/i.image.mosaic?format=raw

and replace the existing script with it.

Alternatively, you can avoid this class of problem entirely by
choosing directories whose names don't contain "awkward" characters
(although this may require a certain level of access privilege).

> My other query is regarding the georectification command and how I
> should interpret the forward and backward RMS error? With other GIS
> programs, the total RMS error has been given as one number.

In most cases, the forward error is the figure which you are likely to
be interested in. That indicates roughly how far a point will end up
from its "true" position.

> When I try
> to georectify the image furthermore, the program freezes. So I am also
> wondering whether I need to compress the images before rectification?
> I have been using approx. 50 MB images.

I have no idea what the problem is here, but compression is unlikely
to help. Compression only affects the size on disk, not the memory
consumption or processing time (other than the fact that compressed
images are slightly slower to read/write, due to the need for
(de)compression).

-- 
Glynn Clements <glynn at gclements.plus.com>