[Gdal-dev] utility usage

[Matt Lynch]

Hi Frank,

Ok, so it looks like I have what I need for most of it, except for the
-mo option on gdal_translate.  What kind of key values can be put there?
Also with gdal_translate, it limits me to one band.  I am currently
looking at batching a process where a 3 band image is translated one
band at a time, then I use ImageMagick to recombine the bands and
finally imagewarp to add the worldfile information (ImageMagick strips
the headers on tiffs). Am I missing a gdal utility within the set that
can recombine the bands for me?

Thanks,

Matt

-----Original Message-----
From: gdal-dev-admin at remotesensing.org
[mailto:gdal-dev-admin at remotesensing.org] On Behalf Of Frank Warmerdam
Sent: Monday, April 14, 2003 2:08 PM
To: gdal-dev at remotesensing.org
Subject: Re: [Gdal-dev] utility usage


Matt Lynch wrote:
> Hi,
> 
> I am working on some tutorial style documentation on the gdal 
> utilities that I am making use of.

Cool!

 >  In the interests of being complete, I would
> like to demonstrate the usage of some the parameters that I am 
> unfamiliar with.  For example, in gdalwarp -order sets the order of 
> the polynom used for warping.  I have experimented with a few images 
> and am unable to find a distinction in the image produced.  Are there 
> specific circumstances in which it will be significant?

Note that this only comes into play if the input images uses GCPs
(ground control points) for the control information instead of an affine
transformation.  Otherwise there is no need to construct polynomials.

 > Similarly I am
> interested in cases where the error threshold (-et) would be set.

The algorithm can use a linear approximation of whatever the general
underlying transformation is, over local areas.  The way this works is
that the first, middle and last point on a line are transformed using
the underlying algorithm.  The difference between the computed middle
point and what it would have been if linearly interpolated is compared
to the "error threshold".  If less the linear approximation is used for
all the points otherwise the line is subdivided and the approximation
test is applied again.

So, the error threshold is intended to be used any time you want to
short circuit the high cost underlying transformation (such as a
reprojection) and you give it the amount of error (in source pixels I
think) that you are willing to accept.  I use it often, usually
specifying a smallish value like 0.1.

> Lastly -co allows you to set name=value pairs.  Could you point me to 
> documentation where I can find the possible settings.  I have also 
> starting using gdal_translate to crop and the -mo settings might be 
> where I can set world file information.  If anyone has some pointers, 
> I can wrap the tutorials up and submit to Frank for approval.

Each format driver will list the format specific format options that
apply to that form.  For instance TIFF has options like "TILED=[YES/NO]"
in it's format info at:

   http://www.remotesensing.org/gdal/frmt_gtiff.html

Best regards,

-- 
---------------------------------------+--------------------------------
---------------------------------------+------
I set the clouds in motion - turn up   | Frank Warmerdam,
warmerdam at pobox.com
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush    | Geospatial Programmer for Rent


_______________________________________________
Gdal-dev mailing list
Gdal-dev at remotesensing.org
http://remotesensing.org/mailman/listinfo/gdal-dev

---
Incoming mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.467 / Virus Database: 266 - Release Date: 4/1/2003
 

---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.467 / Virus Database: 266 - Release Date: 4/1/2003