UDP for parcel delineation.

Author: manugv

openeo_udp/parceldelineation/README.md

@@ -0,0 +1,26 @@
+# Parcel delineation
+This is an [openEO](https://openeo.org/) example for delineating agricultural parcels based on a neural network, using Sentinel-2 input data.
+
+The example focuses on the inference step, using a trained model. It demonstrates data loading and preprocessing, 
+inference, and finally producing vector data as a result.
+
+The example serves as a technology demonstration for openEO, and is not intended for use in another context. 
+Please contact the authors in case you're interested in a field detector!
+
+## Authors
+
+- Kristof Van Tricht
+- Jeroen Dries
+- Victor Verhaert
+
+Tuning by:
+- Kasper Bonte
+- Bart Driessen
+
+[VITO Remote Sensing](https://remotesensing.vito.be)
+
+## Running the example
+
+Most openEO providers require an account to run the example. They offer various options for this, often providing you with some trial to test out the platform.
+You can find all known openEO providers on the [openEO hub](https://hub.openeo.org/).
+

openeo_udp/parceldelineation/generate.py

@@ -0,0 +1,81 @@
+import json
+from pathlib import Path
+import openeo
+from openeo.api.process import Parameter
+from openeo.rest.udp import build_process_dict
+
+
+# TODO investigate setting max cloud cover and kernel size as parameters as well
+def generate() -> dict:
+    # define spatial_extent
+    spatial_extent = Parameter.bounding_box(
+        name="spatial_extent",
+        default={"west": 5.0, "south": 51.2, "east": 5.1, "north": 51.3})
+
+    # define temporal_extent
+    temporal_extent = Parameter.temporal_interval(
+        name="temporal_extent", default=["2021-01-01", "2021-12-31"]
+    )
+
+    # backend to connect and load
+    backend_url = "openeo.dataspace.copernicus.eu/"
+    conn = openeo.connect(backend_url).authenticate_oidc()
+
+    # load the input data and filter based on cloud cover.
+    # max cloud cover should be less than 10%
+    s2_bands = conn.load_collection(
+        collection_id="SENTINEL2_L2A",
+        spatial_extent=spatial_extent,
+        temporal_extent=temporal_extent,
+        bands=["B04", "B08"],
+        max_cloud_cover=10,
+    )
+
+    # The delineation will be estimated based on the NDVI. The `ndvi` process can be used for these calculations.
+    ndviband = s2_bands.ndvi(red="B04", nir="B08")
+
+    # Apply ML algorithm
+    # We now apply a neural network, that requires 128x128 pixel 'chunks' input.
+    segment_udf = openeo.UDF.from_file("udf_segmentation.py")
+    # segment_udf = openeo.UDF.from_url("https://raw.githubusercontent.com/Open-EO/openeo-community-examples/main/python/ParcelDelineation/udf_segmentation.py")
+    segmentationband = ndviband.apply_neighborhood(
+        process=segment_udf,
+        size=[{"dimension": "x", "value": 64, "unit": "px"},
+              {"dimension": "y", "value": 64, "unit": "px"}],
+        overlap=[{"dimension": "x", "value": 32, "unit": "px"},
+                 {"dimension": "y", "value": 32, "unit": "px"}],
+    )
+    
+    # We postprocess the output from the neural network using a sobel filter and
+    # Felzenszwalb's algorithm, which are then merged. This time, we work on larger
+    # chunks, to reduce the need for stitching the vector output.
+    segment_postprocess_udf = openeo.UDF.from_file("udf_sobel_felzenszwalb.py")
+    # segment_postprocess_udf = openeo.UDF.from_url("https://raw.githubusercontent.com/Open-EO/openeo-community-examples/refs/heads/main/python/ParcelDelineation/udf_sobel_felzenszwalb.py")
+    sobel_felzenszwalb = segmentationband.apply_neighborhood(
+        process=segment_postprocess_udf,
+        size=[{"dimension": "x", "value": 2048, "unit": "px"},
+              {"dimension": "y", "value": 2048, "unit": "px"}],
+        overlap=[{"dimension": "x", "value": 0, "unit": "px"},
+                 {"dimension": "y", "value": 0, "unit": "px"}],
+    )
+
+    return build_process_dict(
+        process_graph=sobel_felzenszwalb,
+        process_id= "Parcel Delineation",
+        summary= "Parcel delineation using Sentinel-2 data retrived from the CDSE and processed on openEO.",
+        description= "Parcel delineation using Sentinel-2",
+        parameters= [spatial_extent, temporal_extent],
+        returns=None,  # TODO
+        categories=None,  # TODO
+    )
+
+
+if __name__ == "__main__":
+    # save the generated process to a file
+    output_path = Path(__file__).parent
+    print(output_path)
+    output_path.mkdir(parents=True, exist_ok=True)
+
+    # Save the generated process to a file
+    with open(output_path / "parcel_delineation.json", "w") as f:
+        json.dump(generate(), f, indent=2)