Reading the documentation always helps. The GeomTransform takes the
upper left corner as the origin, not the lower left, as in my math
handbook. And the order of the coefficients is <c b a f e d> With some
twiddling of angles and signs I now can compute the affine
transformation of a scan based on two control points. That means that
with the vrt-file and its geomtransform it is possible to georeference a
scan in real time without warping, at least for a first order
polynomial. Marvelous!
Thanks for your help.
Jan
On 07/04/2013 09:58 PM, Dmitriy Baryshnikov wrote:
Hi Jan,
Do you mean this : http://www.gdal.org/gdal_datamodel.html (Affine
GeoTransform srection) or this http://www.gdal.org/gdal_tutorial.html
(Getting Dataset Information section)?
Also, see this:
http://www.gdal.org/classVRTDataset.html#a1531325b71e290b90c8d9cc88bce6c2b
Best regards,
Dmitriy
04.07.2013 20:12, Jan Hartmann пишет:
I'm struggling with the order of the Geotransform section in a vrt
file. I would like to rotate an image, and computed the six
parameters for the affine transformation. I can transform points
exactly, using the following formula:
x2 = a.x1 + b.y1 + c
y2 = d.x1 + e.y1 + f
Can anyone tell me what the order is in the GeoTransform section of
the VRT-file?
<GeoTransform> c a b f d e </GeoTransform>
results in an image somewhere in the neighbourhood, but mirrored and
translated. What is the order of the parameters a-f?
Jan
_______________________________________________
gdal-dev mailing list
[email protected]
http://lists.osgeo.org/mailman/listinfo/gdal-dev
_______________________________________________
gdal-dev mailing list
[email protected]
http://lists.osgeo.org/mailman/listinfo/gdal-dev
_______________________________________________
gdal-dev mailing list
[email protected]
http://lists.osgeo.org/mailman/listinfo/gdal-dev
