metadata

license: etalab-2.0
pipeline_tag: image-segmentation
tags:
  - semantic segmentation
  - pytorch
  - landcover
model-index:
  - name: FLAIR-HUB_LC-A_swinbase-upernet
    results:
      - task:
          type: semantic-segmentation
        dataset:
          name: IGNF/FLAIR-HUB/
          type: earth-observation-dataset
        metrics:
          - type: mIoU
            value: 64.131
            name: mIoU
          - type: OA
            value: 77.453
            name: Overall Accuracy
          - type: IoU
            value: 83.859
            name: IoU building
          - type: IoU
            value: 78.385
            name: IoU greenhouse
          - type: IoU
            value: 61.59
            name: IoU swimming pool
          - type: IoU
            value: 75.696
            name: IoU impervious surface
          - type: IoU
            value: 57.171
            name: IoU pervious surface
          - type: IoU
            value: 62.939
            name: IoU bare soil
          - type: IoU
            value: 90.35
            name: IoU water
          - type: IoU
            value: 63.381
            name: IoU snow
          - type: IoU
            value: 54.34
            name: IoU herbaceous vegetation
          - type: IoU
            value: 57.135
            name: IoU agricultural land
          - type: IoU
            value: 34.85
            name: IoU plowed land
          - type: IoU
            value: 77.743
            name: IoU vineyard
          - type: IoU
            value: 71.732
            name: IoU deciduous
          - type: IoU
            value: 62.603
            name: IoU coniferous
          - type: IoU
            value: 30.189
            name: IoU brushwood
library_name: pytorch

🌐 FLAIR-HUB Model Collection

Trained on: FLAIR-HUB dataset 🔗
Available modalities: Aerial images, SPOT images, Topographic info, Sentinel-2 yearly time-series, Sentinel-1 yearly time-series, Historical aerial images
Encoders: ConvNeXTV2, Swin (Tiny, Small, Base, Large)
Decoders: UNet, UPerNet
Tasks: Land-cover mapping (LC), Crop-type mapping (LPIS)
Class nomenclature: 15 classes for LC, 23 classes for LPIS

🆔 Model ID	🗺️ Land-cover	🌾 Crop-types	🛩️ Aerial	⛰️ Elevation	🛰️ SPOT	🛰️ S2 t.s.	🛰️ S1 t.s.
LC-A	✓		✓
LC-D	✓		✓			✓
LC-F	✓		✓			✓	✓
LC-G	✓					✓
LC-I	✓				✓
LC-L	✓		✓	✓	✓	✓	✓
LPIS-A		✓	✓
LPIS-F		✓				✓
LPIS-I		✓			✓	✓	✓
LPIS-J		✓	✓		✓	✓	✓

🔍 Model: FLAIR-HUB_LC-A_swinbase-upernet

Encoder: swin_base_patch4_window12_384
Decoder: upernet
Metrics:

mIoU	O.A.	F-score	Precision	Recall
64.13%	77.45%	76.88%	77.36%	76.89%

Params.: 89.4

General Informations

Contact: [email protected]
Code repository: https://github.com/IGNF/FLAIR-HUB
Paper: https://arxiv.org/abs/2506.07080
Project page: https://ignf.github.io/FLAIR/FLAIR-HUB/flairhub
Developed by: IGN
Compute infrastructure:
- software: python, pytorch-lightning
- hardware: HPC/AI resources provided by GENCI-IDRIS
License: Etalab 2.0

Training Config Hyperparameters

- Model architecture: swin_base_patch4_window12_384-upernet
- Optimizer: AdamW (betas=[0.9, 0.999], weight_decay=0.01)
- Learning rate: 5e-5
- Scheduler: one_cycle_lr (warmup_fraction=0.2)
- Epochs: 150
- Batch size: 5
- Seed: 2025
- Early stopping: patience 20, monitor val_miou (mode=max)
- Class weights:
    - default: 1.0
    - masked classes: [clear cut, ligneous, mixed, other] → weight = 0
- Input channels:
    - AERIAL_RGBI : [4,1,2]
- Input normalization (custom):
    - AERIAL_RGBI:
        mean: [106.59, 105.66, 111.35]
        std:  [39.78, 52.23, 45.62]

Training Data

- Train patches: 152225
- Validation patches: 38175
- Test patches: 50700

Training Logging

Metrics

Metric	Value
mIoU	64.13%
Overall Accuracy	77.45%
F-score	76.88%
Precision	77.36%
Recall	76.89%

Class	IoU (%)	F-score (%)	Precision (%)	Recall (%)
building	83.86	91.22	91.21	91.23
greenhouse	78.38	87.88	83.69	92.52
swimming pool	61.59	76.23	76.84	75.63
impervious surface	75.70	86.17	86.35	85.99
pervious surface	57.17	72.75	70.97	74.62
bare soil	62.94	77.25	73.75	81.11
water	90.35	94.93	95.95	93.93
snow	63.38	77.59	94.83	65.65
herbaceous vegetation	54.34	70.42	73.05	67.96
agricultural land	57.14	72.72	69.80	75.90
plowed land	34.85	51.69	49.20	54.43
vineyard	77.74	87.48	84.99	90.11
deciduous	71.73	83.54	82.25	84.87
coniferous	62.60	77.00	79.23	74.90
brushwood	30.19	46.38	48.34	44.57

Inference

Aerial ROI

Inference ROI

Cite

BibTeX:

@article{ign2025flairhub,
  doi = {10.48550/arXiv.2506.07080},
  url = {https://arxiv.org/abs/2506.07080},
  author = {Garioud, Anatol and Giordano, Sébastien and David, Nicolas and Gonthier, Nicolas},
  title = {FLAIR-HUB: Large-scale Multimodal Dataset for Land Cover and Crop Mapping},
  publisher = {arXiv},
  year = {2025}
}

APA:

Anatol Garioud, Sébastien Giordano, Nicolas David, Nicolas Gonthier. 
FLAIR-HUB: Large-scale Multimodal Dataset for Land Cover and Crop Mapping. (2025). 
DOI: https://doi.org/10.48550/arXiv.2506.07080