IGNF/FLAIR-HUB_LPIS-A_swinbase-upernet

🌐 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_LPIS-A_swinbase-upernet

• Encoder: swin_base_patch4_window12_384
• Decoder: upernet
• Metrics:

mIoU	O.A.	F-score	Precision	Recall
22.30%	86.63%	31.21%	37.26%	31.06%

• 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
- 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	22.30%
Overall Accuracy	86.63%
F-score	31.21%
Precision	37.26%
Recall	31.06%

Class	IoU (%)	F-score (%)	Precision (%)	Recall (%)
grasses	49.37	66.10	72.82	60.53
wheat	34.23	51.00	41.11	67.15
barley	13.13	23.21	40.73	16.23
maize	60.50	75.39	77.30	73.57
other cereals	3.49	6.74	8.51	5.57
rice	0.00	0.00	0.00	0.00
flax/hemp/tobacco	2.71	5.27	63.81	2.75
sunflower	12.59	22.36	17.40	31.26
rapeseed	37.98	55.05	61.15	50.06
other oilseed crops	0.00	0.00	0.00	0.00
soy	0.00	0.00	0.00	0.00
other protein crops	3.05	5.93	6.82	5.24
fodder legumes	13.26	23.41	33.03	18.14
beetroots	53.90	70.04	64.80	76.20
potatoes	7.48	13.92	11.05	18.81
other arable crops	19.74	32.97	33.93	32.07
vineyard	43.42	60.55	55.72	66.29
olive groves	13.55	23.87	42.01	16.67
fruits orchards	36.82	53.82	51.31	56.60
nut orchards	2.87	5.59	10.36	3.83
other permanent crops	14.78	25.75	66.07	15.99
mixed crops	1.49	2.93	6.75	1.87
background	88.61	93.96	92.41	95.56

---

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