add an option to exclude bad bands from spectral profile plots

[Andreas Janz]

Requested by @Agustin Lobo .

Add a simple checkbox “Exclude bad bands“ somewhere to the vis. settings:

‌

[Andreas Janz]

Hi @Benjamin Jakimow you can now use the PRISMA l2 import algo to create a raster layer with bad bands defined for testing. See #910 for details.

[Benjamin Jakimow]

resolves #498, resolves #903

→ <<cset d87060a4865f>>

[Andreas Rabe]

Currently not concidering bad band information stored at band-level, as decribed in the Glossary: https://enmap-box.readthedocs.io/en/latest/general/glossary.html#term-bad-band-multiplier

We're planning to have a discussion on that.

[Andreas Rabe]

Also, improve label and fix typo here:

I guess “Band Bands” should be “Bad Bands”.

I would suggest to use “Show Bad Bands” or “Hide Bad Bands”, depending on what “false/true” means.

[Benjamin Jakimow]

seems that I made the “band bands” typo is several places…

[Benjamin Jakimow]

update qps, resolves #903

→ <<cset b3cd9fe29804>>

[Andreas Rabe]

As far as I can see, it is still not concidering bad band information stored at band-level, as decribed in the Glossary: https://enmap-box.readthedocs.io/en/latest/general/glossary.html#term-bad-band-multiplier

[Andreas Rabe]

Use this code snipped to set first 100 bands of the enmap testdata as bad bands:

from osgeo import gdal

from enmapbox import EnMAPBox, initAll
from enmapbox.exampledata import enmap
from enmapbox.testing import start_app
from enmapboxprocessing.rasterwriter import RasterWriter
from qgis._core import QgsRasterLayer

# create raster with a bad band
filename = 'enmap.vrt'
ds = gdal.Translate(filename, enmap)
writer = RasterWriter(ds)
for bandNo in range(1, 101):
    writer.setBadBandMultiplier(0, bandNo)
del writer, ds
layer = QgsRasterLayer(filename)

qgsApp = start_app()
initAll()
enmapBox = EnMAPBox(None)
enmapBox.onDataDropped([layer])
qgsApp.exec_()

‌

[Andreas Rabe]

‌

[Agustin Lobo]

Any planned date for adding this feature?

[Agustin Lobo]

Bad bands are not ignored by Version 3.10.3.20220824T155109.master using envi format with bbl in the hdr.

[Andreas Rabe]

Hi @Agustin Lobo , is this the same dataset you shared recently?

‌

If not, please also share this dataset.

[Agustin Lobo]

I have shared 2 datasets. The 1st one with GTiff + hdr (the one you are showing),

the 2nd one

https://www.dropbox.com/s/m2fkj70z74v4j1n/PRISMA_DESTRIPPED_AOIvAL.7z?dl=0

“in plain envi” as noted in #1348, is the one used to test here with Version 3.10.3.20220824T155109.master

‌

‌

[Andreas Rabe]

thanks

[Agustin Lobo]

I think this current issue can be closed. The problem was that I was interpreting that I needed to

set Bad bands as TRUE to get them out of the spectrum, while this option must be set to FALSE

(I would interpret “Bad bands: false” as “not to consider Bad bands list”. Perhaps this option should be clarified, eg

“Bad bands: included” and Bad Bands: excluded”, as we refer to the plot).

I would have bad bands excluded by default.

[Andreas Rabe]

Thanks Agus, I agree, a better name for the option would be “exclude bad bands” = True/False

‌