OpenDatacube implementation

Defaults

• sentinel1_metadata.yaml : sentinel1 metadata.

• Sentinel1_product_definition.yaml : sentinel1 product definition.

Utility scripts

• dhusget : script to download whole datasets from scihub.copernicus.eu.

• s1prepare : script to prepare preproccessed images to be indexable through the datacube api

Setup - Implementation so far

1. First configure the datacube Datacube installation
2. Define the product and the metadata through the use of the datacube api
3. Test the database with a simple jupyter notebook.

Activation

• Run the script under /PATH_OF_YOUR_DATACUBE_INSTALLATION/datacube_env/bin/activate to get access to the virtualenv

Problems - Observations

• The path of the images-bands is incorrectly annotated if the rasters are in a subfolder.

• Final yaml is saved as agdc-metadata.yaml on all inputs

• The preparation script requires rasters to be on a folder called SENTINEL_1A (configurable through the script).

TODO

• Implement recursive folder search to correct annotate the rasters' path.

• Provide example queries for simple operations.

Usage - Scenario

Adding product definitions

datacube product add <path-to-product-definition.yaml>

Examples of product definition can be found here

Adding metadata for the desired product

datacube metadata add <path-to-metadata-definition.yaml>

Prepare the preproccessed .tif files

python <path-to-prepare-script.py> <path-to-folder-containing-tif-files>

Add the dataset generated by the prepare script

datacube dataset add <path-to-yaml-generated-from-script>