farmingpy.eo API

farmingpy.eo API#

BioPhysS2tbx

Implementation of Sentinel 2 Toolbox Neural Network for predicting biophysical parameters from Sentinel 2 data.
class BioPhysS2tbx(product='LAI', resolution=20, stbx_version='3.0')[source]#

Implementation of Sentinel 2 Toolbox Neural Network for predicting biophysical parameters from Sentinel 2 data. Uses weights from s2tbx-biophysical: senbox-org/s2tbx

The model does input and output range validation, but omits convexity checks.

Based on: Weiss, M., Baret, F., Jay, S., 2020. S2ToolBox level 2 products LAI, FAPAR, FCOVER. EMMAH-CAPTE, INRAe Avignon. http://step.esa.int/docs/extra/ATBD_S2ToolBox_V2.0.pdf

__call__(ds, *args, **kwargs)[source]#

Run the model on dataset.

Parameters:

xarray.DataArray (ds) – Sentinel 2 data. The format needs to match the data retrieved using twinyields.eo.S2SentinelHub class.

Returns:

Vegetation index

Return type:

xarray.DataArray

__init__(product='LAI', resolution=20, stbx_version='3.0')[source]#
Parameters:

  • product (str, optional) – Output product, “LAI”, “FAPAR”, “FCOVER”, “LAI_Cab” or “LAI_cw”. Defaults to “LAI”.

  • resolution (int, optional) – Model resolution to use 10 or 20. Defaults to 20.

  • stbx_version (str, optional) – Toolbox version for weights 2.1 or 3.0. Defaults to “3.0”.

read_planet(img_file, apply_mask=True, confidence=60, clip=None)[source]#

Read planetscope AnalyticMS_SR_8b .tif and apply mask.

Parameters:

  • img_file (str) – Image file path.

  • apply_mask (bool, optional) – Apply udm1 mask and remove points with low mask confidence if True. Defaults to True.

  • confidence (int, optional) – Mask confidence threshold. Defaults to 60.

  • clip (GeoDataFrame, optional) – GeoDataFrame used to clip the image.

Returns:

Planetscope image xarray.DataArray: Mask

Return type:

xarray.DataArray

planet_to_S2_dataset(ds)[source]#

Convert PlanetScope 8 band image to fake Sentinel2 dataset to be used with twinyields.eo.BioPhysStbx models.

Parameters:

ds (xarray.DataArray) – Planet image loaded using read_planet.

Returns:

Dataset with bands that can be used with twinyields.eo.BioPhysStbx 10m models.

Return type:

xarray.DataArray