Orthorectified Sentinel-1 using eo-learn

Hi All,
I’m trying to retrieve orthorectified and radiometrically terrain corrected Sentinel-1 data using the eo-learn package. I’m referring to the processing options here. My understanding is that the default backCoeff would be GAMMA0_ELLIPSOID, so here is what I use in eo-learn for my input task:

input_task = SentinelHubInputTask(
    data_collection = DataCollection.SENTINEL1_IW,
    bands= ['VV', 'VH],
    bands_feature = (FeatureType.DATA, 'GRD_Data'),
    resolution = 10,
    time_difference = datetime.timedelta(hours=1),
    config = config(),
    max_threads = 1,
    aux_request_args = {

I successfully retrieve the data, but I’m not sure if it’s indeed radiometrically terrain corrected as if I don’t pass the aux_request_args the results are exactly the same. If I change the backCoeff to gamma0_ellipsoid and set orthorectify to False, I still get the same results.

Any guidance on how to ensure I get the right processing options passed to the API would be appreciated.


Hi @hamedalemo

The correct way to pass the additional processing settings is like so:

for Hamed about eo-learn:

aux_request_args={'processing': card4l_processing},

card4l_processing = {
    "backCoeff": "GAMMA0_TERRAIN",
    "orthorectify": True,
    "demInstance": "COPERNICUS",
    "downsampling": "BILINEAR",
    "upsampling": "BILINEAR"


PS: the example has all the parameters to request CARD4L compliant S-1 data.

Hi @hamedalemo ,

An easy way to check if the request format is correct is using Requests Builder. From the Request Preview window you’ll have the correct format of your desired request, then you can compare the payload created from _create_sh_request, which returns SentinelHubRequest that has payload attribute.

More info:

Thanks Matej. This is great, and worked smoothly.

Thank you @chung.horng . I didn’t know about the Requests Builder. This is a great resource.