EuroCropsML

EuroCropsML provides a time series benchmark dataset for few-shot crop type classification in Europe, leveraging Sentinel-2 imagery and the open-source EUROCROPS collection to advance machine learning benchmarking and research comparability in transnational crop classification.

Key Features:

• Time series remote sensing data: Each data point features a time series of per-parcel median pixel values extracted from Sentinel-2 L1C data with precise geospatial coordinates.
• Large-scale multi-class labeling: The dataset comprises 706,683 multi-class labeled data points across 176 crop classes.
• Few-shot classification benchmark: It is the first time-resolved remote sensing dataset designed to benchmark transnational few-shot crop type classification algorithms.
• Open data availability: The dataset is publicly available on Zenodo.

Scientific Applications:

• Crop type classification: Enables machine learning benchmarking of time series crop type classification across European countries.
• Algorithmic development: Supports advancements in transnational few-shot crop type classification algorithms.
• Research comparability: Facilitates standardized comparison of machine learning methods for agricultural remote sensing.

Methodology:

Per-parcel median pixel values are extracted from Sentinel-2 L1C data, with each data point assigned precise geospatial coordinates and multi-class crop labels from the EUROCROPS collection.