Update README.md

c00a250 verified about 17 hours ago

26.3 kB

license: etalab-2.0
pipeline_tag: image-segmentation
tags:
  - semantic segmentation
  - pytorch
  - landcover
library_name: pytorch
model-index:
  - name: FLAIR-HUB_LC-B_RGB_swinbase-upernet
    results:
      - task:
          type: semantic-segmentation
        dataset:
          name: IGNF/FLAIR-HUB/
          type: earth-observation-dataset
        metrics:
          - name: mIoU
            type: mIoU
            value: 64.491
          - name: Overall Accuracy
            type: OA
            value: 77.437
          - name: IoU building
            type: IoU
            value: 85.126
          - name: IoU greenhouse
            type: IoU
            value: 78
          - name: IoU swimming pool
            type: IoU
            value: 61.59
          - name: IoU impervious surface
            type: IoU
            value: 76.064
          - name: IoU pervious surface
            type: IoU
            value: 58.228
          - name: IoU bare soil
            type: IoU
            value: 65.714
          - name: IoU water
            type: IoU
            value: 90.421
          - name: IoU snow
            type: IoU
            value: 63.742
          - name: IoU herbaceous vegetation
            type: IoU
            value: 53.162
          - name: IoU agricultural land
            type: IoU
            value: 58.385
          - name: IoU plowed land
            type: IoU
            value: 38.039
          - name: IoU vineyard
            type: IoU
            value: 78.754
          - name: IoU deciduous
            type: IoU
            value: 70.644
          - name: IoU coniferous
            type: IoU
            value: 60.549
          - name: IoU brushwood
            type: IoU
            value: 30.993

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

🔍 Model: FLAIR-HUB_LC-B_RGB_swinbase-upernet

Encoder: swin_base_patch4_window12_384
Decoder: upernet
Metrics:

mIoU	O.A.	F-score	Precision	Recall
64.49%	77.44%	77.22%	77.58%	77.46%

Params.: 181.4

General Informations

Contact: flair@ign.fr
Code repository: https://github.com/IGNF/FLAIR-HUB
Paper: https://arxiv.org/abs/2506.07080
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 : [1,2,3]
- Input normalization (custom):
    - AERIAL_RGBI:
        mean: [105.66, 111.35, 102.18]
        std:  [52.23, 45.62, 44.30]

Training Data

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

Training Logging

Metrics

Metric	Value
mIoU	64.49%
Overall Accuracy	77.44%
F-score	77.22%
Precision	77.58%
Recall	77.46%

Class	IoU (%)	F-score (%)	Precision (%)	Recall (%)
building	85.13	91.97	91.83	92.10
greenhouse	78.00	87.64	83.87	91.76
swimming pool	59.54	74.64	72.11	77.36
impervious surface	76.06	86.40	86.86	85.96
pervious surface	58.23	73.60	72.37	74.87
bare soil	65.71	79.31	74.84	84.35
water	90.42	94.97	94.55	95.39
snow	63.74	77.86	96.94	65.05
herbaceous vegetation	53.16	69.42	73.08	66.11
agricultural land	58.38	73.73	68.86	79.34
plowed land	38.04	55.11	54.26	56.00
vineyard	78.75	88.11	85.15	91.29
deciduous	70.64	82.80	81.92	83.69
coniferous	60.55	75.43	78.56	72.53
brushwood	30.99	47.32	48.55	46.15

Inference

Aerial ROI

Inference ROI

Cite

BibTeX:

@article{GARIOUD2026271,
title = {FLAIR-HUB: Large-scale multimodal dataset for land cover and crop mapping},
author = {Anatol Garioud and Sébastien Giordano and Nicolas David and Nicolas Gonthier},
journal = {ISPRS Journal of Photogrammetry and Remote Sensing},
volume = {237},
pages = {271-300},
year = {2026},
issn = {0924-2716},
doi = {https://doi.org/10.1016/j.isprsjprs.2026.04.017},
url = {https://www.sciencedirect.com/science/article/pii/S0924271626001899},
}

APA:

Anatol Garioud, Sébastien Giordano, Nicolas David, Nicolas Gonthier. 
FLAIR-HUB: Large-scale Multimodal Dataset for Land Cover and Crop Mapping.
ISPRS Journal of Photogrammetry and Remote Sensing, Volume 237, 2026.
DOI: https://doi.org/10.1016/j.isprsjprs.2026.04.017