{"success":true,"data":{"author":"LINKS Foundation","bucketAccessType":"public","bucketName":"links-ads/spada-dataset","bucketPrefix":"","category":"Earth Observation & Geospatial","cloudProvider":"huggingface","createdAt":"2026-11-16T14:11:25.967Z","description":"SPADA is the dataset from Land Cover Segmentation with Sparse Annotations from Sentinel-2 Imagery (IGARSS 2023), built to train land cover and fuel-type segmentation models from weak, sparse labels over European areas of interest. Inputs are Sentinel-2 L2A images from 2018 with 12 bands in reflectance x 10,000 (GeoTIFF), obtained from the Microsoft Planetary Computer. Two kinds of label are provided: scribble annotations derived from CORINE Land Cover 2018 polygons after a scribbling process that keeps only sparse strokes inside each class region, and point-wise labels from the 2018 LUCAS in-situ survey rasterised in 5x5 pixel blocks. The training split also provides overlapping 512x512 tiles, while validation and test are cut into non-overlapping 2048x2048 sections. The data are stored as segmented tar.gz parts under data/ that must be concatenated and extracted (about 165 GB); training code is in github.com/links-ads/igarss-spada. Released by LINKS Foundation under CC BY 4.0; cite the IGARSS 2023 SPADA paper.","displayTitle":"SPADA: Sentinel-2 Land Cover Segmentation with Sparse CORINE Scribbles and LUCAS Point Labels","id":"23f4aa39-5545-520a-b01a-efd96562133a","license":"CC-BY-4.0","licenseEvidenceUrl":"https://huggingface.co/datasets/links-ads/spada-dataset","objectCountEstimate":158,"probeStatus":"ok","referenceUrl":"https://arxiv.org/abs/2306.16252","region":"us","schemaCard":{"accessHints":[{"language":"python-datasets","snippet":"from datasets import load_dataset\nds = load_dataset(\"links-ads/spada-dataset\")"},{"language":"python-hub","snippet":"from huggingface_hub import snapshot_download\nsnapshot_download(repo_id=\"links-ads/spada-dataset\", repo_type=\"dataset\")"}],"caveats":["HuggingFace metadata: author=links-ads, license:cc-by-4.0, size_categories:10M<n<100M, downloads=443"],"format":"mixed","generatedAt":"2026-11-16T14:11:30.274Z","layout":{"example":"assets/sample.png","pattern":"{repoId}/{filename}"},"mountPath":"/mnt/archive","sampleHeaders":[{"bytes":"89504e470d0a1a0a0000000d4948445200000e60000004df0806000000af3dd98d000000017352474200aece1ce90000000467414d410000b18f0bfc61050000000970485973000016250000162501495224f00000ffa549444154785eecdd07785bd5fdc6712710ca2aa52d6d192d23b1b5bc12c7899dc51e85d252fe8c522850661865070a94bd43062164400661efbd77d8237bd8ce626f0884ec9dbcfff3d3208a726d4bb6645bf2f7f33cef9361e9eaea6af85cfbbc3a79020000000000000000000000000000000000000000c07a286002000000000000000000000000000000000000000024a08009000000000000000000000000000000000000000090800226000000000000000000000000000000000000000040020a9800000000000000000000000000000000000000000009286002000000000000000000000000000000000000000024a08009000000000000000000000000000000000000000090800226000000000000000000000000000000000000000040020a9800000000000000000000000000000000000000000009286002000000000000000000000000000000000000000024a08009000000000000000000000000000000000000000090800226000000000000000000000000000000000000000040020a9800000000000000000000000000000000000000000009286002000000000000000000","encoding":"hex","key":"assets/sample.png","truncatedAt":512}],"sampleObjects":["assets/sample.png","data/test/test.tar.0000.gz.part","data/test/test.tar.0001.gz.part","data/test/test.tar.0002.gz.part","data/test/test.tar.0003.gz.part"],"sidecars":[{"bytes":"---\nlicense: cc-by-4.0\ntask_categories:\n- image-segmentation\ntags:\n- sparse-labels\n- segmentation\n- eo\npretty_name: SPADA Dataset\nsize_categories:\n- 10M<n<100M\n---\n\n# SPADA Dataset\n\nThis dataset contains images and sparse labels used in the paper [Land Cover Segmentation with Sparse Annotations from Sentinel-2 Imagery\n](https://arxiv.org/abs/2306.16252), published at IGARSS 2023.\n\n---------\n- **Repository:** https://github.com/links-ads/igarss-spada\n- **Paper:** https://paperswithcode.com/paper/land-cover-segmentation-with-sparse\n---------\n![Dataset sample](assets/sample.png)\n\n## Dataset Preparation\n\nThe dataset has been compressed into segmented tarballs for ease of use within Git LFS (that is, tar > gzip > split).\nTo revert the process into files and directories, follow these steps:\n\n```console\n$ git clone https://huggingface.co/datasets/links-ads/spada-dataset\n$ cd spada-dataset\n# revert the multipart compression: merge first, then untar\n$ cat data/train/train.tar.* | tar -xzvf - -i\n$ cat data/test/test.tar.* | tar -xzvf - -i\n$ cat data/val/val.tar.* | tar -xzvf - -i\n```\n\nIt is very likely that the extracted files will retain the internal directory structure, making the `train/val/test` directories useless.\nAdapt the output structure as you see fit, the original structure is shown below.\n\n## Dataset Structure\n\nThe main dataset used in the paper comprises the following inputs:\n\n| Folder  | Data Type          | Description                                                                               | Format                   |\n|---------|--------------------|-------------------------------------------------------------------------------------------|--------------------------|\n| Img     | Sentinel-2 Image   | L2A data from 2018, with 12 channels in reflectance/10k format                            | GeoTIFF (.tif)           |\n| CLC     | Scribble labels    | 2D labels derived from Corine Land Cover (CLC), obtained after the scribbling process     | GeoTIFF (.tif)           |\n| Fuel    | Pointwise labels   | LUCAS points rasterized in a 5x5 grid to obtain point-wise annotations                    | GeoTIFF (.tif)           |\n\nThe training set also provides overlapping 512x512 tiles, while validation and testing are divided into non-overlapped sections with size 2048x2048 for ease of use. \n\n### Source Data\n\n- Sentinel-2 L2A 2018 images were downloaded from Microsoft Planetary Computer, using predefined AoIs.\n- CLC and LUCAS data from 2018 were downloaded from the official Copernicus Land platform, at [https://land.copernicus.eu/en](https://land.copernicus.eu/en). \n- Urban Atlas (UA) can also be downloaded from there.\n\n### Licensing Information\n\nCC-BY-4.0 [https://creativecommons.org/licenses/by/4.0/](https://creativecommons.org/licenses/by/4.0/)\n\n### Citation Information\n\n```bibtex\n@inproceedings{galatola2023land,\n  title={Land Cover Segmentation with Sparse Annotations from Sentinel-2 Imagery},\n  author={Galatola, Marco and Arnaudo, Edoardo and Barco, Luca and Rossi, Claudio and Dominici, Fabrizio},\n  booktitle={IGARSS 2023-2023 IEEE International Geoscience and Remote Sensing Symposium},\n  pages={6952--6955},\n  year={2023},\n  organization={IEEE}\n}\n```\n\n### Contributions\n- Edoardo Arnaudo (edoardo.arnaudo@polito.it | linksfoundation.com)\n- Marco Galatola (marco.galatola@linksfoundation.com)\n- Luca Barco (luca.barco@linksfoundation.com)\n\n\n### Acknowledgements\n\nThis dataset was created in the context of the H2020 SAFERS project, under G.A. n. 869353, and HEU OVERWATCH project, under G.A. n.101082320, with the support of the EU Agency for the Space Programme (EUSPA).\nMore information:\n- [https://safers-project.eu/](https://safers-project.eu/)\n- [https://overwatchproject.eu/](https://overwatchproject.eu/)\n\n\n\n","encoding":"utf8","key":"README.md","name":"README.md","size":3774,"truncated":false}],"sourceObjectKey":"assets/sample.png","starterCell":"from huggingface_hub import hf_hub_download\npath = hf_hub_download(repo_id=\"links-ads/spada-dataset\", filename=\"assets/sample.png\", repo_type=\"dataset\")\nprint(path)","starterSamplePath":"assets/sample.png","totalObjects":158},"summary":"SPADA is the dataset from Land Cover Segmentation with Sparse Annotations from Sentinel-2 Imagery (IGARSS 2023), built to train land cover and fuel-type segmentation models from…","title":"SPADA: Sentinel-2 Land Cover Segmentation with Sparse CORINE Scribbles and LUCAS Point Labels","updatedAt":"2026-11-16T15:51:20.205Z","visibility":"public","websiteUrl":"https://huggingface.co/datasets/links-ads/spada-dataset"}}