GeoscienceAustralia / TSCloudMask

1 stars 1 forks source link

Consider the production application of the algorithm #4

Open alexgleith opened 4 years ago

alexgleith commented 4 years ago

When we run this algorithm in production, we'll be wanting to produces a per-scene product.

I wonder if we can frame up the processing to take a dataset ID and from that, go into the past or future of observations and build a new QA band that matches the original data footprint.

Also, output should be Cloud Optimised GeoTIFF and be accompanied with at least YAML metadata, including lineage, and preferably also a STAC conversion of the YAML.