Closed Pratichhya closed 7 months ago
from openeo.processes import log
I would not do an from-import like that. "log" is an too-overloaded word to be a globally imported like that. I would just do import openeo.processes
and use openeo.processes.log(x)
where you need it
filter_window = (np.ones([601, 601])).tolist()
the tolist()
is not necessary I think, the python client does this automatically
factor = 1 / 361201
I would calculate the factor directly from the filter_window numpy array (e.g. 1/filter_window.sum()
) instead of hardcoding 361201
Let us plot the result
I'm confused by the image. As far as I understand the notebook, oil_spill
is a binary image because if the thresholding, but the image looks like grayscale
nice! You could also add raster_to_vector to try and get vector outline of the spill, if it produces something decent of course.
@soxofaan , Thank you for all the suggestions, changes have been made and indeed, yes, it was supposed to be a binary image.
@jdries I tried to add the raster_to_vector
process too; however, it fails with a different error when trying to execute the process with either CSV or NetCDF format, and JSON doesn't give interpretative result(or I wasn't able to understand it).
I am not sure if I need to include additional processes to get the vector data.
resolved merge conflict
@soxofaan, as discussed earlier, I will have to share fixed links for a couple of use cases, so I will merge them at this point after first reviewing the version and then changing them when needed. However, it seems like I would need at least one approval to merge 😅
Sure @jdries thank you.
Also, if possible, can you allow for #20 for the initial merge?
In this notebook, we want to showcase a basic methodology for mapping oil spills using openEO.
The methodology adopted in the notebook was shown here.
Additionally, the following references were equally helpful when preparing this usecase:
https://forum.step.esa.int/t/steps-for-oil-spill-detection/7428/6 https://ieeexplore.ieee.org/document/1370264