BELDE: Building a Large-scale Earth-observation Land-cover Dataset for Europe

2026-06-18

AI Summary

Existing RGB remote sensing land cover segmentation datasets are limited in geographic coverage, scale, or public availability. This work introduces BELDE, a large-scale, publicly available dataset constructed from Sentinel-2 true-color imagery and ESA WorldCover labels, encompassing all of Europe (with over one million samples), South Korea, and the California–Nevada region of the United States, featuring seven land cover classes at 10-meter spatial resolution. BELDE enables, for the first time, cross-continental generalization studies in land cover mapping. Benchmark evaluations using multiple semantic segmentation models achieve an F1 score of 83.0% on the European test set, but performance drops significantly on out-of-domain subsets—66.4% on BELDE-CA-NV and 58.3% on BELDE-K—highlighting the substantial impact of geographic domain shift and establishing BELDE as a critical resource for advancing the robustness and transferability of remote sensing models.

Abstract

Earth observation imagery plays a critical role in environmental monitoring, urban planning, disaster assessment, and climate analysis. While multi-spectral sensors are increasingly available, true-color (RGB) imagery remains widely used due to the power, cost, and deployment constraints of many satellite and aerial platforms. However, existing land-cover segmentation datasets are often limited in geographic coverage, scale, or public accessibility. To bridge this gap, we introduce BELDE (Building a Large-scale Earth-observation Land-cover Dataset for Europe), a publicly available dataset tailored for RGB-based remote sensing semantic segmentation. Constructed from Sentinel-2 true-color images and ESA WorldCover data annotations, BELDE contains 1,088,385 curated image-segmentation map pairs spanning Europe with 7 land-cover classes at 10 m spatial resolution, making it one of the largest publicly available RGB land-cover segmentation datasets for Earth observation. To facilitate cross-region generalization studies, we additionally introduce BELDE-K (16,607 pairs) covering the Republic of Korea and BELDE-CA-NV (88,155 pairs) covering California and Nevada in the United States. We establish baseline results using multiple semantic segmentation architectures and evaluate both in-domain and cross-domain performance. Models trained on BELDE achieve an F1 score of 83.0% on the European test set, while performance decreases to 66.4% on BELDE-CA-NV and 58.3% on BELDE-K, highlighting the challenges posed by out-of-distribution geographic domain shift. By providing a continental-scale RGB segmentation and evaluation benchmark, BELDE supports the development of robust and transferable Earth observation models. The dataset and benchmark resources will be publicly released.

Problem

• land-cover segmentation
• RGB imagery
• Earth observation
• dataset scalability
• geographic domain shift

Innovation

• large-scale dataset
• RGB remote sensing
• semantic segmentation
• domain generalization
• land-cover mapping