[GRASS5] [bug #859] (grass) raster data at lower resolution: no resampling...

[Glynn Clements]

Markus Neteler wrote:

> > > > (sorry to be so chatty)
> > > >
> > > > from the GRASS prog's manual I understand that raster data
> > > > are resampled on the fly when looking at them in lower resolution.
> > >
> > > Yes. Note: "resampled" does not imply "interpolated"; you can't
> > > meaningfully interpolate category values.
> > >
> > 
> > I have just send an email to Markus explaining just that. It was too long
> > for the list. If raster represents a continuous field, one has to
> > reinterpolate, if it is an area or line map with categories (what GRASS
> > originally was designed for) the resampling was acceptable.
> 
> ... maybe I am too tired right now, but how shall GRASS detect if
> the raster represents a continuous field (in INT) or a categories map?
> Currently GRASS seems to silently do something.

I don't think that it can "detect" it. It needs to be told.

And if it's to interpolate, then how? There are lots of possible
interpolation techniques, and none of them is the "one true way".

Consider the following (one dimensional) example:

	x	y
	0.0	0.0
	1.0	1.0
	2.0	1.0
	3.0	0.0

Many interpolation techniques (e.g. finding a polynomial which passes
through those four points) would produce a maximum at x=1.5, and the
maximum would be greater than 1. If the actual values are guaranteed
to lie in the [0,1] range, this is clearly wrong.

> Please be patient :-)

Sure.

-- 
Glynn Clements <glynn.clements at virgin.net>