[Qgis-developer] Fwd: Discrepancy between value tool, spectral profile and Point sampling tool for large data sets (2)

[Agustin Lobo]

FYI
Illustration:
https://sites.google.com/site/openfiles2/home/valuetoolproblem2.jpeg
Test data:
https://sites.google.com/site/openfiles2/home/samplingtoolproblem.zip

Agus

---------- Forwarded message ----------
From: Agustin Lobo <Agustin.Lobo at ija.csic.es>
Date: 2010/12/2
Subject: Discrepancy between value tool, spectral profile and Point
sampling tool  for large data sets (2)
To: Ghislain Picard <ghislain.picard at lgge.obs.ujf-grenoble.fr>,
fidel at centrogeo.org.mx, borys at wolf.most.org.pl, alobolistas at gmail.com


> Hi!
>
> I've found that value tool reports values in a different order than spectral profile and Point Sampling tool
> for very large data sets (this one 72 layers). See in the attachment
> the value of the first band, for example, in Value tool (118),  while spectral profile and the values for that point extracted by Point Sampling Tool are different.
> The sequence 228, 225, 248 ... in Value tool starts in layer 50
>
> I've checked in R (where the file is made) and Value Tool is right.
>
> The most confusing is that a 3-band raster made with the first 3 bands
> works the same for all 3 tools: 118, 243, 236
>
> I'm trying to make a little 3x3x72 raster for testing and will report to
> qgis lists as well.
>
> Agus