L7 Irish: Landsat 7 ETM+ Cloud Cover Assessment Validation Scenes (cleaned, COG)
- Provider
- USGS (Irish et al.; Scaramuzza et al.), repackaged by Stewart et al. (TorchGeo)
- License
- CC0-1.0 (source)
- Size
- 20 files
- Stored at
- Hugging Face (https://huggingface.co/datasets/torchgeo/l7irish)
- Access
- Free and public. No account with the data provider is needed.
- DOI
- 10.48550/arXiv.2306.09424
- Category
- Earth Observation & Geospatial
- Source
- https://huggingface.co/datasets/torchgeo/l7irish
About L7 Irish: Landsat 7 ETM+ Cloud Cover Assessment Validation Scenes (cleaned, COG)
Cloud and cloud-shadow segmentation dataset of 206 Landsat 7 ETM+ Level-1G scenes with manually generated cloud masks, spread over nine latitude zones (austral, boreal, mid-latitude north and south, polar north and south, subtropical north and south, tropical). Each scene is a 9-band image of roughly 8000x8000 pixels at 30 m (15 m pan); masks have five classes: fill, cloud shadow, clear, thin cloud and cloud. The scenes were originally selected by Irish et al. and digitised by Scaramuzza et al. for USGS cloud-cover algorithm validation. This copy is the cleaned version from the SSL4EO-L paper: corrupted thermal files and missing masks were replaced from the alternate USGS release, filenames made consistent, bands stacked and written as compressed Cloud-Optimized GeoTIFFs. Each zone is one tar.gz, with a _raw variant alongside (about 228 GB in total). Load with torchgeo.datasets.L7Irish. Source data are USGS public domain and the repackaged dataset is released under CC0 1.0 (Stewart et al. 2023); cite Scaramuzza et al. 2012 and Stewart et al. 2023.
How to load this dataset in Python
from huggingface_hub import hf_hub_download
path = hf_hub_download(repo_id="torchgeo/l7irish", filename="austral.tar.gz", repo_type="dataset")
print(path)Data structure
Files
Example file: austral.tar.gz
Analyze this dataset without downloading it
On Scigantic, L7 Irish: Landsat 7 ETM+ Cloud Cover Assessment Validation Scenes (cleaned, COG) opens in a hosted JupyterLab notebook with the data already mounted, so you can read files where they are stored instead of copying them first. Each notebook includes an AI assistant that can see the mounted files and write and run analysis code. Open the dataset page to launch a notebook, or browse the full catalog.
This page is also available as Markdown and as JSON.
Related Earth Observation & Geospatial datasets
- RapidCrops: A pan-European label dataset for large-scale crop classification
- Global Fields of The World (FTW) 2nd Edition: Field Boundary Predictions 2017-2025
- Geospatial ML
- Data to Science Catalog
- DEA Sentinel-2B MSI Analysis Ready Data, Collection 3 (Australia)
- NEXRAD Level II real-time chunks (Unidata, AWS)
- AMX Project 2024-04 PMTiles
- USGS US Topo 7.5-Minute Quadrangle Maps (Geospatial PDF and GeoTIFF)