>>>>> Hamish <[EMAIL PROTECTED]> writes:
[...]
> an idea: write a little script to make a vector coversheet index.
> maybe output that with the HTMLMAP driver for an interactive web brower
> pull up.
> I am not sure, but I think v.patch does not clean topology so overlap
> may be ok.
Wouldn't it be better to save non-overlapping parts and an
intersection as separate polygons? A separate attribute will be
needed to store the list of IDs of the covering images.
[...]
> this is just the idea. non functional...
> # store current region
> g.region save=old_region
It makes me wonder each time I see such a fragile construct like
this, will GRASS ever support overriding the region
``temporarily''? (E. g., via a command line argument, or an
environment variable.)
> # create new empty map
> v.in.ascii -e out=map_index
> v.db.addtable map_index columns="image varchar(128)"
> i=0
> for MAP in `g.mlist patt='ortho_*'` ; do
> i=`expr $i + 1`
> g.region rast="$MAP"
> v.in.region out="${MAP}_box"
Oh, perhaps `v.in.region' could be patched to print the category
of the vector feature it creates (if requested)? I guess, this
usage pattern is quite common for `v.in.region'?
And, could there be something a bit more clever than storing the
region of the image as a vector feature? A combination of
downsampling, `r.mapcalc' (to obtain the fill mask), and
`r.contour', may be?
> ? v.edit to reset the region box centroid's cat to $i ?
> # or a series of v.category del then add steps
> # merge into master index
> g.remove vect=map_index_old
> g.rename vect=map_index,map_index_old
> v.patch in=map_index_old,"${MAP}_box" out=map_index
May there `v.patch -a' be used instead? Like:
v.patch -a in="${MAP}_box" output=map_index
> # write map name to DB entry
> v.db.update map_index column=image value="'$MAP'" where="cat = $i"
> done
> #reset to original region
> g.region old_region
[...]
_______________________________________________
grass-user mailing list
[email protected]
http://lists.osgeo.org/mailman/listinfo/grass-user
