[Geotiff] another problem, which I shold have brought up earlier

rudahl at rsgis.net rudahl at rsgis.net
Thu Feb 2 04:21:28 EST 2012


I mst apologize. When I first met this mailing list, I had an impression
that no active development was going on so I didn't forward any comments.
I should say I have been actively using GeoTIFF since I did some early 
testing with Mike Ruth, but have been away from it.

When I recently became more involved, I saw several apparent omissions,
at least from my point of view. These are, generally, the lack of tags
for identifying the sensor.

TIFF was set up for having either a single "band" (grey) or three (RGB,
or whatever). There isn't explicit support for using an arbitrary
number of bands nor is there any way that I can discover for figuring
out which band is which. Different manufacturers of remote sensing 
products do different ad-hoc things with multi-band files, so it seems one 
is left with no choice, in a GeoTIFF reader, but to try to identify
the producer and hope their new file is not different in layout from
the one you looked at last year.

It seems that GeoTIFF could, and should, rationalize at least this
uncertainty. It is easy, though, to see that this need also includes
needs for significantly more information about each band, separately.
What could be done, for example, with a Modis multiband file where
the different bands have different spatial properties?

Respectfully,

Kurt Rudahl
Goldin-Rudahl Systems, Inc.


More information about the Geotiff mailing list