Kenneth P. Bowman writes:
> There goes my chance for fame. :-(
Well, hang on. There's still trouble in River City,
and it rhymes with "p", and it stands for "projection".
I got a little too excited before. :-)
It's pretty good. Not bad. But not bang on right, either.
If I open a GeoTIFF file like this AVHRR file here:
http://www.dfanning.com/map_tips/tiffoverlay.html
Then I can find the UV bounding box by using the tie-points,
the scales, and the size of the image:
s = Size(image, /Dimensions)
xscale = geotag.ModelPixelScaleTag[0]
yscale = geotag.ModelPixelScaleTag[1]
tp = geotag.ModelTiePointTag[3:4]
xOrigin = tp[0]
yOrigin = tp[1] - (yscale * s[1])
xEnd = xOrigin + (xscale * s[0])
uv_box = [xOrigin, yOrigin, xEnd, yEnd]
This assumes that the image is "squared-up" with the UV
coordinate system.
Unfortunately, my image is NOT "squared-up" with the UV
coordinate system. It is tilted a little bit. Here are
the four corners of my warped image in UV space, after
I warped the lat and lon arrays and converted the corner
points to UV space. (Starts in lower-left and goes clockwise.)
-2361218.9 -1869566.0
-2373277.4 1841973.4
2360269.4 1842014.4
2348276.6 -1869524.8
Thus, when I try to treat it as if it were the AVHRR image
above, I *almost* get it right, but not quite. I think I need
some way to "square up" this image with the UV coordinate system.
Unfortunately, it is not a simple rotation, because the image
itself is not a rectangle in this warped space. It's a
squarish blob. :-(
I'm looking for drugs or ideas, either one will do.
Cheers,
David
--
David Fanning, Ph.D.
Fanning Software Consulting, Inc.
Coyote's Guide to IDL Programming: http://www.dfanning.com/
Se****e ma de ni thui. ("Perhaps thou speakest truth.")
|
