IGNF/FLAIR-HUB_LC-A_IR_convnextv2base-unet

🌐 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-A_convnextv2base-unet

• Encoder: convnextv2_base
• Decoder: unet
• Metrics:

mIoU	O.A.	F-score	Precision	Recall
64.16%	77.17%	76.92%	77.58%	76.59%

• Params.: 92.8

---

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: convnextv2_base-unet
- 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	84.15	91.39	91.15	91.64
greenhouse	76.22	86.50	84.11	89.04
swimming pool	60.03	75.02	76.08	73.99
impervious surface	75.24	85.87	86.75	85.01
pervious surface	56.17	71.94	69.87	74.14
bare soil	63.02	77.31	74.19	80.71
water	88.96	94.16	94.98	93.35
snow	72.54	84.08	97.77	73.76
herbaceous vegetation	54.22	70.31	71.67	69.01
agricultural land	57.09	72.68	69.75	75.87
plowed land	36.27	53.23	52.71	53.77
vineyard	77.47	87.30	85.34	89.36
deciduous	71.33	83.26	81.90	84.67
coniferous	60.43	75.33	80.13	71.08
brushwood	29.30	45.33	47.34	43.48

---

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