IGNF/FLAIR-HUB_LC-G_utae

🌐 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.	🛩️ Historical
LC-A	✓		✓
LC-D	✓		✓			✓
LC-F	✓		✓			✓	✓
LC-G	✓					✓
LC-I	✓				✓
LC-L	✓		✓	✓	✓	✓	✓
LPIS-A		✓	✓
LPIS-F		✓				✓
LPIS-I		✓			✓	✓	✓
LPIS-J		✓	✓		✓	✓	✓

🔍 Model: FLAIR-HUB_LC-G_utae

• Encoder: UTAE
• Decoder: UTAE
• Metrics:

mIoU	O.A.	F-score	Precision	Recall
34.24%	57.83%	47.30%	48.12%	47.59%

• Params.: 0.9

---

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: UTAE
- 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:
    - SENTINEL2_TS : [1,2,3,4,5,6,7,8,9,10]

---

Training Data

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

---

Training Logging

---

Metrics

Metric	Value
mIoU	34.24%
Overall Accuracy	57.83%
F-score	47.30%
Precision	48.12%
Recall	47.59%

Class	IoU (%)	F-score (%)	Precision (%)	Recall (%)
building	34.91	51.75	54.37	49.37
greenhouse	0.00	0.00	0.00	0.00
swimming pool	0.00	0.00	0.00	0.00
impervious surface	38.27	55.35	51.43	59.92
pervious surface	27.43	43.05	51.24	37.12
bare soil	33.59	50.29	56.33	45.42
water	65.32	79.02	71.19	88.79
snow	67.54	80.63	69.71	95.61
herbaceous vegetation	34.44	51.23	51.81	50.66
agricultural land	42.08	59.24	57.01	61.65
plowed land	10.23	18.56	19.29	17.88
vineyard	41.10	58.26	67.59	51.20
deciduous	55.99	71.79	67.97	76.06
coniferous	48.22	65.06	77.39	56.12
brushwood	14.46	25.27	26.54	24.12
Selection deleted

---

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