[GRASS-user] question on i.nightlights.intercalibration code

[Markus Metz]

On Fri, Jun 22, 2018 at 10:34 PM, Veronica Andreo <veroandreo at gmail.com>
wrote:
>
> Hi Gabriel,
>
> What you could do is import with r.in.gdal -a  that adjusts resolution
for lat long maps [0]

that will help to fix the resolution from 0.008333333300000 to
0.008333333333333, i.e. exactly 30 arc-seconds. The software used to create
the raster data has stored the resolution with limited precision.

> and then (before the intercalibration step), set the region to one of the
imported maps with g.region raster=yourmap

you will then get a message like
360 degree EW extent is exceeded by 1 cells

which is correct because the first and last column are duplicates. The cell
centers cover -180, 180, and the EW extents regarding cell borders are E:
180:00:15E, W: 180:00:15W, grown by half a cell, i.e. 15 arc-seconds.

Markus M

>
> HTH,
> Vero
>
> [0] https://grass.osgeo.org/grass74/manuals/r.in.gdal.html
>
> El vie., 22 jun. 2018 10:32, Gabriel Cotlier <gabiklm01 at gmail.com>
escribió:
>>
>> Dear Nikos and Veronica,
>>
>> Thanks a lot for your emails, I'm happy it worked!
>>
>> I will try to avoid of the extent message  "360 degree EW extent is
exceeded by 0.999827 cells" using:  g.region -a, should I enter this
command before importing the images - raster data layers ?
>>
>> Thanks a lot again for your guidance.
>>
>> Gabriel
>>
>>
>> On Fri, Jun 22, 2018 at 12:50 AM, Nikos Alexandris <
nik at nikosalexandris.net> wrote:
>>>
>>> * Gabriel Cotlier <gabiklm01 at gmail.com> [2018-06-21 18:24:38 -0300]:
>>>
>>>> Hello Veronica,
>>>>
>>>> Thanks a lot it worked!!
>>>
>>>
>>> Dear Gabriel,
>>>
>>> Thanks for posting details and great it works. Don't worry about the
>>> last image, there are no coefficients for F18 and 2013. We are forced
>>> to skip this image in any analysis.
>>>
>>>
>>> Details
>>>
>>> Veronica is right, and I am ignorant for so many things when it comes
to Windows.
>>>
>>> In Linux, one may use the syntax `some_command` or
>>> $(some_command). It's a smart way to feed-in longer list of items, such
>>> as map names in GRASS GIS. Note, the former is old-style, so best to
use the $()
>>> notation (if of interest, see
>>> https://unix.stackexchange.com/a/5782/13011).
>>>
>>> You might want to re-check the extent you have set, so as to
>>> get rid of the "360 degree EW extent is exceeded by 0.999827 cells"
>>> messages (?). Depending on how you have set your region, maybe the `-a`
>>> option for `g.region` would help (?).
>>>
>>> The last failing image is not a computational error. The error
>>> message
>>> --%<---
>>>>
>>>> ValueError: The selected model does not know about this
>>>> combination of Satellite + Year!
>>>
>>> --->%--
>>> means that there are no coefficients for satellite "F18" and for the
>>> year 2013. You may read this in
>>>
https://trac.osgeo.org/grass/browser/grass-addons/grass7/imagery/i.nightlights.intercalibration/intercalibration_coefficients.py#L99
>>> and the referenced papers, of course.
>>>
>>> For various reasons, I did not improve the script in
>>> handling this situations. I just let it emit an error message. I hardly
>>> remember having modified this message to say something like "There are
>>> given coefficients for satellite F?? and Year ????". But I can't trace
>>> anything in my local repository.
>>>
>>> (
>>> Note, there are many newer algorithms. It would be obviously great to
>>> integrate some of them in the existing module. If a new model bases upon
>>> a similar math formula, then it's only necessary to add:
>>>
>>> 1. a new set of coefficients in
https://trac.osgeo.org/grass/browser/grass-addons/grass7/imagery/i.nightlights.intercalibration/intercalibration_coefficients.py#L38
>>>
>>> 2. a new equation in
https://trac.osgeo.org/grass/browser/grass-addons/grass7/imagery/i.nightlights.intercalibration/intercalibration_equations.py#L15
>>>
>>> 3. a new subclass, for the new model, in
https://trac.osgeo.org/grass/browser/grass-addons/grass7/imagery/i.nightlights.intercalibration/intercalibration_models.py
>>>
>>> Maybe more work if there is a substantially different math logic.
>>> )
>>>
>>> Thanks, Nikos
>>
>>
>
> _______________________________________________
> grass-user mailing list
> grass-user at lists.osgeo.org
> https://lists.osgeo.org/mailman/listinfo/grass-user
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/grass-user/attachments/20180623/3f107398/attachment-0001.html>