Skip to content

Dataset · IGNF

Multimodal topography and crop segmentation dataset

IGNF/FLAIR-HUB

A large dataset for land cover and agricultural crop segmentation in France.

What's inside

Combines previous versions FLAIR#1 and FLAIR#2 into a single multisensory geospatial dataset covering over 2,500 square kilometers of French landscapes. Contains between 100k and 1M examples with ultra-high resolution labeling: 63 billion pixels manually annotated across 19 land cover classes and 23 crop types.

What you can do with it

  • Training satellite and aerial imagery segmentation models
  • Classifying soil, vegetation, and agricultural land types
  • Developing agricultural and environmental monitoring systems
Tasks
Image segmentation
Size
100K–1M rows
Downloads this month
13.6K
Likes
37

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("IGNF/FLAIR-HUB")