EODAG EnMAP Data Downloader

[jakimowb]

Compared to portals from, e.g. USGS https://earthexplorer.usgs.gov/ or Planet.com, finding and downloading suitable EnMAP data is complicated.

1. https://explore.eo-lab.org provides a nice, state-of-the-art UI to browse and search available EnMAP observations. However, it is unclear how to get registered right to download EnMAP data as well.

2. The DLR GeoService https://geoservice.dlr.de/web/ lists EnMAP data at https://download.geoservice.dlr.de/ENMAP/files/. They provide a STAC https://geoservice.dlr.de/eoc/ogc/stac/v1/search. However, most STAC browsers miss a proper support for spatial filtering - which is essentially for products with global coverage https://radiantearth.github.io/stac-browser/#/external/geoservice.dlr.de/eoc/ogc/stac/v1

It is recommended to easy the way to obtain EnMAP L2 data

• use EODAG, as it already supports various other data providers (USGS, ESA, ..)
• query EnMAP Observations by:
  • date and time
  • cloud cover
  • spatial regions
  • other metadata values
• download EnMAP L2 observation from STAC https://geoservice.dlr.de/eoc/ogc/stac/v1/search.
• run EnMAP-Box import processing algorithm to create VRTs that ensure the EnMAP-Box can read wavelength information

[jakimowb]

EnMAP provider for EODAG

# geoservice.dlr.de/eoc/ogc/stac/v1/collections/ENMAP_HSI_L2A
# https://geoservice.dlr.de/eoc/ogc/stac/v1/search
# EODAG Provider to download EnMAP data from geoservice.dlr.de/eoc/ogc/stac/v1/collections
enmap:
   search:
      type: StacSearch
      api_endpoint: https://geoservice.dlr.de/eoc/ogc/stac/v1/search
      need_auth: true
   products:
      GENERIC_PRODUCT_TYPE:
         productType: '{productType}'
   download:
      type: HTTPDownload
#   auth:
#      type: GenericAuth
#      username: ${EODAG_ENMAP_USERNAME}
#      password: ${EODAG_ENMAP_PASSWORD}

How to download EnMAP data with EODAG. Access credentials can be set using the environmental vars: EODAG_ENMAP_USERNAME and EODAG_ENMAP_PASSWORD

import os
from pathlib import Path

from eodag import EODataAccessGateway
from eodag.plugins.crunch.filter_property import FilterProperty
from osgeo import gdal, ogr, osr
from osgeo.osr import CoordinateTransformationOptions

gdal.UseExceptions()

path_output = Path(__file__).parent / "downloads"
os.makedirs(path_output, exist_ok=True)

max_cloud_cover = 25

path_enmapprovider = Path(__file__).parent / "enmapprovider.yml"
path_geom = Path(__file__).parent / 'geojson/enmap_aoi_utm.geojson'

# convert AOI to EPSG:4326 in x,y (!) order

assert path_geom.is_file()
ds: gdal.Dataset = ogr.Open(path_geom.as_posix())
lyr: ogr.Layer = ds.GetLayerByIndex(0)
srsDst: osr.SpatialReference = osr.SpatialReference()
srsDst.ImportFromEPSG(4326)
srsSrc = lyr.GetSpatialRef()
wkt = None
for feature in lyr:
    feature: ogr.Feature
    g = feature.GetGeometryRef()
    wktSrc = g.ExportToWkt()
    if srsSrc.ExportToWkt() != srsDst.ExportToWkt():
        srsSrc.SetAxisMappingStrategy(osr.OAMS_TRADITIONAL_GIS_ORDER)
        srsDst.SetAxisMappingStrategy(osr.OAMS_TRADITIONAL_GIS_ORDER)
        options = CoordinateTransformationOptions()
        trans = osr.CoordinateTransformation(srsSrc, srsDst)
        assert ogr.OGRERR_NONE == g.Transform(trans)
    wkt = wktDst = g.ExportToWkt()
    break
del ds

# Initialize EODAG
dag = EODataAccessGateway()
# add EnMAP provider (generic STAC)
dag.update_providers_config(yaml_conf=path_enmapprovider.read_text())


# Define search parameters
search_criteria = {
    "productType": "ENMAP_HSI_L2A",
    "start": "2024-01-01",
    "end": "2024-12-31",
    # 'geom': [7.08, 50.73, 7.15, 50.78] ,
    'geom': wkt
    #'location': str(path_aoi),
}

download_criteria = {
    # a regular expression to filter assets to download
    'asset': 'SPECTRAL_IMAGE_COG.TIF|METADATA.*|QL_PIXELMASK.*'
}


print("Searching for EnMAP HSI_L2A data...")
products = dag.search(**search_criteria, provider="enmap")
print(f"Found {len(products)} products")

filtered_products = products.crunch(
    FilterProperty(dict(cloudCover=25, operator="lt"))
)

print(f"{len(filtered_products)} with cloud cover <= {max_cloud_cover}")

# Display product information
for i, product in enumerate(filtered_products):
    print(f"\nProduct {i}:")
    print(f"  ID: {product.properties.get('id')}")
    print(f"  Date: {product.properties.get('start_datetime')}")
    print(f"  Cloud Cover: {product.properties.get('start_datetime')}")

for i, product in enumerate(filtered_products):
    if i in [0,1]:
        continue
    try:
        print(f"\nDownloading product {i + 1}/{len(products)}")
        print(f"  Product ID: {product.properties.get('id')}")

        path = dag.download(product,
                            output_dir=str(path_output),
                            **download_criteria)
        print(f"  Downloaded to: {path}")

    except Exception as e:
        print(f"  Failed to download product {i + 1}: {e}")

[danschef]

@jakimowb So would the goal of this issue be to integrate the download functionality for EnMAP data into the EnMAP-Box?

[jakimowb]

Yes. To use an existing download tool and provide a GUI for interactive selection and adjustment of search queries, and finally to download the EnMAP data.

[janzandr]

Very nice!