Some answers to your questions.
On 9/7/07, Martin Desruisseaux <[EMAIL PROTECTED]> wrote:
>
> In "Image I/O metadata for multidimensional coverage" document
>
>
> > 5. ImageIO MetaData for Multidimensional Coverages overview
> > -----------------------------------------------------------
> > In order to be able to describe a Multidimensional Coverage (or
> > ND-Coverage), we need to introduce the concept of coverage metadata,
> > which sits right in between the stream metadata concept and the image
> > metadata concept.
>
> I was expecting a 1-to-1 relationship between an Image and a Coverage
> (or at least a GridCoverage2D). The need for a Coverage intermediate
> level suggests that there is no such 1-to-1 relationship. What is the
> relationship then? Do we expect Coverage metadata to be returned by
> "smart" readers only, and not the "flat" ones? Do we expect a one-to-one
> or one-to-many relationship between "Coverage metadata" and "Image
> metadata"?
Basically, we could say that a coverage refers to a set of 2D slices
(images) each one identifying a restricted data portion. Supposing a simple
multidim coverage contains gridded data for the temperature for 3 different
temporal instants, at 5 different elevations (for each instant) for a fixed
X-Y region, we may identify 3x5=15 different 2D slices (one for instant T0,
elevation Z0, one other for T0,Z1, one other for T0,Z2 and so on... ).
ImageMetadata refers to a single 2D Slice.
Coverage metadata will be returned by smart reader.
"overview" in Figure 8
> ----------------------
> "Overviews" seems to be used there in a sense similar to "Thumbnails" in
> J2SE ImageReader. Is "Overviews" the term used in OGC specification? If
> not, should we consider using the same term than J2SE ImageReader?
>
>
> "bandKeys" in Figure 8
> ----------------------
> See my comment for stream metadata. I believe that "bandKeys" and
> "nullValues" should not be part of CRS axis, but rather part of
> SampleDimensions. Or maybe the whole "axis" node is about
> SampleDimensions, in which case I would suggest to rename it?
I provided you a brief answer about this in the other mail.
--
Best Regards,
Daniele Romagnoli
-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Geotools-devel mailing list
Geotools-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geotools-devel
