SBS Boston Hackathon

About

Info about this hackathon project.

Goals:

• CO2.Storage & Bacalhau compute
  • Demonstrate efficient processing of Sentinel satellite imagery over CO2.Storage assets
  • Allow networking in Bacalhau docker image for efficient querying of COG data
• Interoperability in ecological assets
  • Process Sentinel data for any CO2.Storage asset that has Geojson data
  • Provide NDVI results back on CO2.Storage in standard format
• Tracking deforestation
  • Use Sentinel imagery to compute NDVI summaries over long time periods

Docker

docker pull ghcr.io/cambiumteam/sentinel-processing:latest

The docker image contains a file process.py which handles processing of Sentinel data. Run this script with the following parameters:

• -s, --start: Start date.
• -e, --end: End date.
• p, --process: Process type: ndvi-summary, rgb. Defaults to ndvi-summary.

Configure input and output directories. A docker volume can be mounted if not running in Bacalhau:

• -i, --input: Input directory. Directory with geojson files, or a single input file. Defaults to /inputs.
• -o, --output: Output directory to write results to. Defaults to /outputs.

Bacalhau

Create a Bacalhau job that runs ghcr.io/cambiumteam/sentinel-processing:latest.

The following inputs are supported:

• CID with GeoJSON content --inputs $CID
• GeoJSON URL --input-urls $URL

bacalhau --api-host $HOST \
  docker run --network full --inputs $CID \
  ghcr.io/cambiumteam/sentinel-processing:latest -- \
  python process.py -s 2021-01-01 -e 2021-06-01`

Local containers

Run local-containers/start.sh to configure a local environment:

• ipfs_host: IPFS node connected to CO2.Storage IPFS swarm.
• bacalhau: Bacalhau requester & compute node. Configured to allow full network access for jobs.