-
Notifications
You must be signed in to change notification settings - Fork 379
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add BigEarthNet dataset #197
Add BigEarthNet dataset #197
Conversation
5e4b6e0
to
5b66e01
Compare
I see the following warnings in our tests:
Can you look into this? It seems like the test files are not actually GeoTIFFs but we're still using rasterio to load them. Are the actual data files GeoTIFFs? |
They are just dummy tiffs without georeferenced metadata. The actual files are geotiffs. |
Okay, so we just need to create real GeoTIFFs using https://github.com/microsoft/torchgeo/tree/main/tests/data#raster-data for these tests. Can you do that? |
Yes, I'll make a PR updating them. |
* add bigearthnet dataset * add dummy data for bigearthnet tests * add bigearthnet unit tests * updated bigearthnet dataset and tests with s1 imagery * add bigearthnet to docs * mypy fixes * updated docstrings
torchgeo.datasets.BigEarthNet
Notes:
BigEarthNet._load_paths
is really not elegant but it's due to the mapping between Sentinel-1 -> Sentinel-2 patches being in the individual json file for each Sentinel-1 patch. I'm computing the mapping on indexing and not on instantiation of the dataset because opening 590k json files is too time consuming for instantiating a dataset.Closes #63