Closed palmoreck closed 4 years ago
This issue is solved.
I used:
/LUSTRE/MADMEX/tasks/2020/4_landcover_manglares_check/3_checking_s2_model_fit_predict_and_db_to_raster.ipynb
and again:
/LUSTRE/MADMEX/tasks/2020/4_landcover_manglares_check/2_checking_s2_recipes_20m.ipynb
to see what was the problem.
The problem was related to not using the correct recipe sentinel2. I needed to use the recipe that resamples to 10m the 20m product such as:
antares apply_recipe -recipe s2_20m_resampled_10m_001 -b 2018-01-01 -e 2019-12-31 \
-region area_estudio_manglares_Qroo --name recipe_mex_s2_20m_resampled_to_10m_2018_2019 \
--resolution -10 10 --tilesize 50020 50020 --origin 2426720 977160 \
--proj4 '+proj=lcc +lat_1=17.5 +lat_2=29.5 +lat_0=12 +lon_0=-102 +x_0=2500000 +y_0=0 +a=6378137 +b=6378136.027241431 +units=m +no_defs' \
-sc /shared_volume/scheduler.json
I also generate some training points for sentinel2 madmex product of Quintana Roo quintana_roo_s2_2018_madmex_17_clases
and conabio_manglares
products refered in this issue.
With this training points i got:
[x] Use
quintana_roo_s2_2018_madmex_17_clases
already registered in DB as training data for model fit process and check results -> maybe directly launching predict and db to raster processes via API (as in docu) and visualize result?[x] Use
conabio_manglares
already registered in DB as training data for model fit process and check results -> maybe directly launching predict and db to raster processes via API (as in docu) and visualize result?