[GRASS5] image3-lib, i.points3, i.rectify3

[Markus Neteler]

Bob,

(cc grass5)

thanks for sending the update. I have tested and updated again:

On Tue, Jan 08, 2002 at 02:01:18PM -0400, Bob Covill wrote:
> Markus,
> 
> I have attached tar balls of the directories needed to try i.rectify3.
> You first need to untar the image3.tar file in libes and build the new
> library. There are a couple of modified files there that convert
> selected coordinates back and forth to user coordinates.

You had fixed two things, I have updated the lib in CVS so that we have
the same code base (release_branch as before).

> I could not remember if there was anything changed in i.points3 so
> I have attached it anyway. One of the files has a line that prints the
> coordinates to stderr that I put in for debugging. This should help to
> make sure the coordinates are being converted correctly.  When using
> i.points3 remember to select the transformation type (1st, 2nd, ...).
> This initializes a file REF(?) which i.rectify3 looks for.

it is the TRANS file. I suggest that we implement POLY1 as default,
or. better, a question before entering i.points3. Otherwise most
users will forget this and wonder about the problem later.
I have fixed a few cc warnings. Latest code is in release branch (as
before).

For those fiddling with i.points/i.vpoints: The i.points3 is really
great and much more convenient!

> The last tar ball is the old i.rectify3 program. I seem to remember that
> I had trouble building both the cmd and inter versions. Eventually I got
> the the cmd version working and used that.  You don't need to build the
> inter version.

Well, I have updated both to the GRASS 5 raster management, according
to i.rectify2 (it differs a bit due to the new methods in i.rectify3).
Both CMD and INTER are functional now. I have updated in main as this
module is not in release_branch.

> Let me know if things work for you. It was several months ago that I
> looked at this, and I may have forgot somethings. I am interested to see
> what results i.rectify3 produces. According to the notes I found in the
> directories it was intended to replace i.rectify??
Yes. It produces reasonable results.

Next steps are:

 - re-activate/implement the missing functions like POLY2/3/PHOTO/LANDSAT
   in image3 library
 - update i.points3 accordingly to be able to select these methods
 - add query at beginning of i.points3 to have the method always 
   defined (avoids later i.rectify3 crash)
 - update i.rectify3 to allow other methods than POLY1

If done, it seems that we can generate Ortho-satellite images through
DTM correction. That will be really great (and needed, e.g. here where
I live).

Thanks for the latest code,

 Markus