i.landsat.toar not respecting region?

[CarlosGrohmann]

As the title says, I noticed that i.landsat.toar is not respecting the region.

Is there a reason behind this? Or is it a missing feature? (I'd love to know how to implement it...)

Running GRASS 8.0 dev (26976e3) on Ubunt 21.04

cheers

[marisn]

It seems to be intentional since inception of the module. It should be included in the documentation (open a bug report). @YannChemin or @metzm could explain the reasoning of such approach.

grass/imagery/i.landsat.toar/main.c

Line 403 in 07bd00f

Rast_get_cellhd(band_in, "", &cellhd);

[veroandreo]

My guess is that the reason for using the whole raster region instead of the computational region has to do with DOS (dark object substraction); i.e., to have the largest possible sample of pixels from where to get the darkest one of the scene and perform the correction.

In any case, if you don't need the full scene, you could set the region in advance and only import that subset. All of r.in.gdal, r.import and i.landsat.import allow that.

I do agree this behavior should be mentioned in the manual page. Would you mind suggesting a text snippet in a PR, @CarlosGrohmann ?

[CarlosGrohmann]

Sure, I'll prepare this

[CarlosGrohmann]

To which branch should I submit the PR?

[veroandreo]

main, I would say. Then we backport

[CarlosGrohmann]

Ok, done

[YannChemin]

Dear all, Right, the DOS requires the full image, but of course you can create your region of interest, import the band in it and then run the module (rightly said Vero!) Initially we had the module as a script auto import system, take everything and run pre-processing (i.landsat.toar, i.atcorr, then all the i.akbedo, i.ndvi, i.evapo* etc) on the full tiles, when reaching Level 2 of processing, then see what you have and run your specific model for 10 years of meteo and RS data. So of course, a single image subset area is far from the initial idea. I think there are still some scripts example of that in pygrass wiki for ref. So yes, unless someone interested in coding the region change, use Vero’s point of view. Best regards, Avec mes meilleurs sentiments, Dr. Yann Chemin Scientific Officer European Commission Joint Research Centre Directorate D – Sustainable Resources Via E. Fermi 2749, TP 272 I-21027 Ispra (VA) - Italy Tel: +39 0332 78 6698 Work (not pgp): ***@***.*** Encrypted (safe): ***@***.*** Encrypted Audio/Video (safe): jami.net yannchemin

…

On Sep 10, 2021, at 11:34 AM, Veronica Andreo ***@***.***> wrote: My guess is that the reason for using the whole raster region instead of the computational region has to do with DOS (dark object substraction); i.e., to have the largest possible sample of pixels from where to get the darkest one of the scene and perform the correction. In any case, if you don't need the full scene, you could set the region in advance and only import that subset. All of r.in.gdal, r.import and i.landsat.import allow that. I do agree this behavior should be mentioned in the manual page. Would you mind suggesting a text snippet in a PR? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#1853 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABM224DCDV3GGNALQHXFIKDUBHGITANCNFSM5DYH6PSA>. Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

[CarlosGrohmann]

Thanks for the answer. I see the reason behind it. I was just a bit confused because it wasn't mentioned in the manual (will write something about it).