On 5/22/07 10:40 PM, "Yann" <yann.chemin@gmail.com> wrote:
After importing a multispectral image with r.in.gdal, a group is created
holding all these layers.
I don't follow you on this. What kind of group? AFAIK, r.in.gdal just
creates a set of separate raster files with identical prefixed names. Does
it build an image group?
Could we have a button that would show/select/load any existing group in the
GUI (it could have a tree inside for all the bands available).
Once loaded we could select one or any combination of 3 bands for display.
Putting the maps in separate layers would simply overlay one on top of the
other, not recombine them into an RGB image again. This would have to take
place within d.rgb.
Following your idea, an enhanced d.rgb would be able to load a group, in
addtion to just any individual layer, making layer selection more convenient
for bands of a multiband image. I'd add that it should also have a couple of
options for doing automatic equalization stretches (linear, log, and
gaussian if I want to imagine a really nice module) on the bands before it
displays them.
Michael
__________________________________________
Michael Barton, Professor of Anthropology
School of Human Evolution & Social Change
Center for Social Dynamics & Complexity
Arizona State University
Yann wrote:
> After importing a multispectral image with r.in.gdal, a group is
> created holding all these layers.
Michael:
I don't follow you on this. What kind of group? AFAIK, r.in.gdal just
creates a set of separate raster files with identical prefixed names.
Does it build an image group?
Yes, r.in.gdal puts all imported bands into a group. "g.list group"
Get map names in the group with "i.group -g $GROUPNAME". (or "-l")
> Could we have a button that would show/select/load any existing
> group in the GUI (it could have a tree inside for all the bands
> available).
>
> Once loaded we could select one or any combination of 3 bands for
> display.
Putting the maps in separate layers would simply overlay one on top of
the other, not recombine them into an RGB image again. This would have
to take place within d.rgb.
Following your idea, an enhanced d.rgb would be able to load a group,
in addtion to just any individual layer, making layer selection more
convenient for bands of a multiband image. I'd add that it should also
have a couple of options for doing automatic equalization stretches
(linear, log, and gaussian if I want to imagine a really nice module)
on the bands before it displays them.
"d.rgb group=" or rgbhis.tcl browse button.
The group list is not ordered, so you would have to assume that Red was
first, Green second, and Blue third, and ignore any other bands.
Maybe ok for hinting map names in rgbhis.tcl.
Maybe putting the first 3 maps in a group into d.rgb would be handy. I was
thinking of just having d.rgb searching the group if a group was loaded
rather than searching the entire mapset for files when you browse. Not sure
how this would be handled though.
On 5/23/07 12:23 AM, "Hamish" <hamish_nospam@yahoo.com> wrote:
"d.rgb group=" or rgbhis.tcl browse button.
The group list is not ordered, so you would have to assume that Red was
first, Green second, and Blue third, and ignore any other bands.
Maybe ok for hinting map names in rgbhis.tcl.
__________________________________________
Michael Barton, Professor of Anthropology
School of Human Evolution & Social Change
Center for Social Dynamics & Complexity
Arizona State University
Maybe putting the first 3 maps in a group into d.rgb would be handy. I
was thinking of just having d.rgb searching the group if a group was
loaded rather than searching the entire mapset for files when you
browse. Not sure how this would be handled though.
As it relies on a fuzzy guess, I'd prefer that it happen in rgbhis.tcl
to provide prefill answers, and not in the "lower-level" d.rgb at runtime.
On 5/23/07 6:23 PM, "Hamish" <hamish_nospam@yahoo.com> wrote:
Michael Barton wrote:
Maybe putting the first 3 maps in a group into d.rgb would be handy. I
was thinking of just having d.rgb searching the group if a group was
loaded rather than searching the entire mapset for files when you
browse. Not sure how this would be handled though.
As it relies on a fuzzy guess, I'd prefer that it happen in rgbhis.tcl
to provide prefill answers, and not in the "lower-level" d.rgb at runtime.
Hamish
__________________________________________
Michael Barton, Professor of Anthropology
School of Human Evolution & Social Change
Center for Social Dynamics & Complexity
Arizona State University