PASTIS-R: Panoptic Segmentation of Agricultural Parcels from Sentinel-1 and Sentinel-2 Time Series (TACO)

Provider
Sainte Fare Garnot, Landrieu, Chehata (IGN France / LASTIG); TACO repackaging by ISP, Universitat de Valencia
License
CC-BY-4.0 (source)
Size
17 files
Format
PNG (in sampled files)
Stored at
Hugging Face (https://huggingface.co/datasets/isp-uv-es/pastis-r-taco)
Access
Free and public. No account with the data provider is needed.
DOI
10.1109/ICCV48922.2021.00483
Category
Agriculture & Environment
Source
https://huggingface.co/datasets/isp-uv-es/pastis-r-taco

About this dataset

PASTIS-R is the radar-augmented version of PASTIS, the reference benchmark for panoptic segmentation of agricultural parcels from satellite image time series (ICCV 2021; ISPRS 2022). It has 2,433 patches of 128x128 pixels at 10 m over four Sentinel-2 tiles in France, each with a full year of Sentinel-2 L2A (10 bands, 37 to 61 dates), Sentinel-1 ascending orbit (3 layers, 65 to 71 dates) and Sentinel-1 descending orbit (69 to 70 dates). Targets are an 18-crop semantic map built from the French Land Parcel Identification System, plus the parcel identifier of every pixel for instance and panoptic evaluation with metrics such as panoptic quality. The release defines five cross-validation folds; this copy uses folds 1-3, 4 and 5 as train, validation and test (1,455, 482, 496) and carries the fold so any 5-fold run can be rebuilt. Series keep their native variable length. About 53 GB in TACO format, read with tacoreader. Original on Zenodo under CC BY 4.0; cite Sainte Fare Garnot and Landrieu 2021.

How to load this dataset in Python

from huggingface_hub import hf_hub_download
path = hf_hub_download(repo_id="isp-uv-es/pastis-r-taco", filename="assets/logos.png", repo_type="dataset")
print(path)

Data structure

File types in sampled files: PNG

Files

Example file: assets/logos.png

Analyze this dataset without downloading it

On Scigantic, PASTIS-R: Panoptic Segmentation of Agricultural Parcels from Sentinel-1 and Sentinel-2 Time Series (TACO) 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 Agriculture & Environment datasets