GeoBench FieldsOfTheWorld

GeoBench FieldsOfTheWorld#

Intro#

Fields of The World (FoTW) (Kerner at al. 2024) is a comprehensive global dataset for agricultural field boundary delineation using multi-temporal satellite imagery. The dataset provides field boundary annotations across diverse agricultural landscapes worldwide, enabling automated field mapping for precision agriculture, crop monitoring, and food security applications.

Dataset Characteristics#

  • Modalities:

    • Multi-temporal Sentinel-2 optical imagery

  • Spatial Resolution: 10m (resampled from native Sentinel-2 bands)

  • Temporal Resolution: Multi-temporal composites across growing seasons

  • Spectral Bands:

    • S2: 10 bands (B02, B03, B04, B05, B06, B07, B08, B8A, B11, B12)

  • Image Dimensions: 256x256 pixels per patch

  • Labels: Agricultural field boundary segmentation

    • Binary segmentation (field boundaries vs. non-boundaries)

    • Precise boundary delineation for field mapping

  • Geographic Distribution: Global coverage across major agricultural regions

  • Temporal Coverage: Multiple growing seasons (2019-2021)

GeoBenchV2 Processing Pipeline#

Preprocessing Steps#

  1. Sample filtering:

    • Dataset was filtered to only contain data that is released under open CC-BY license, namely: “austria”, “brazil”, “corsica”, “denmark”, “estonia”, “finland”, “france”, “india”,”kenya”, “luxembourg”,”netherlands”, “rwanda”, “slovakia”, “spain”, “vietnam”

  2. Split Generation:

    • Use the existing train/val/test splits from the dataset

  3. Dataset Subsampling:

    • The final version consists of

      • 4,000 training samples

      • 1,000 validation samples

      • 2,000 test samples

References#

  1. Kerner, H., Chaudhari, S., Ghosh, A., Robinson, C., Ahmad, A., Choi, E., Jacobs, N., Holmes, C., Mohr, M., Dodhia, R. and Ferres, J.M.L., 2025, April. Fields of the world: A machine learning benchmark dataset for global agricultural field boundary segmentation. In Proceedings of the AAAI Conference on Artificial Intelligence (Vol. 39, No. 27, pp. 28151-28159).

  2. Fields of The World Website: https://fieldsofthe.world/

[1]:

import os
from pathlib import Path

import torch

from geobench_v2.datamodules import GeoBenchFieldsOfTheWorldDataModule
from geobench_v2.datasets import GeoBenchFieldsOfTheWorld
from geobench_v2.datasets.normalization import SatMAENormalizer, ZScoreNormalizer
from geobench_v2.datasets.visualization_util import (
    compare_normalization_methods,
    compute_batch_histograms,
    plot_batch_histograms,
    plot_channel_histograms,
    visualize_segmentation_target_statistics,
)

%load_ext autoreload
%autoreload 2

[2]:

DATA_ROOT = Path("../../data/fotw")

STATS_SATMAE_PATH = os.path.join(DATA_ROOT, "fotw_stats_satmae.json")
STATS_CLIP_RESCALE_PATH = os.path.join(DATA_ROOT, "fotw_stats_clip_rescale.json")

[3]:

band_order = GeoBenchFieldsOfTheWorld.band_default_order

datamodule = GeoBenchFieldsOfTheWorldDataModule(
    img_size=256,
    batch_size=4,
    num_workers=4,
    root=DATA_ROOT,
    band_order=band_order,
    data_normalizer=torch.nn.Identity(),  # we do custom normalization in the tutorial
)
datamodule.setup("fit")

Using provided pre-initialized normalizer instance: Identity
Using provided pre-initialized normalizer instance: Identity

[4]:

sample_dist_fig = datamodule.visualize_geospatial_distribution()
../_images/dataset_notebooks_fotw_4_0.png

Dataset Statistics#

Computed over the training dataset.

Image Statistics#

[5]:

fig = plot_channel_histograms(STATS_SATMAE_PATH)
../_images/dataset_notebooks_fotw_7_0.png
../_images/dataset_notebooks_fotw_7_1.png

Target Statistics#

[6]:

figs = visualize_segmentation_target_statistics(STATS_SATMAE_PATH, "FOTW")
../_images/dataset_notebooks_fotw_9_0.png

Raw Batch Statistics#

[7]:

# Get a batch of data from the dataloader
train_dataloader = datamodule.train_dataloader()
raw_batch = next(iter(train_dataloader))

raw_batch_stats = compute_batch_histograms(raw_batch, n_bins=100)


raw_figs = plot_batch_histograms(
    raw_batch_stats, band_order, title_suffix=" (Raw Data)"
)
raw_figs

[7]:

[<Figure size 1200x500 with 1 Axes>, <Figure size 1200x500 with 1 Axes>]
../_images/dataset_notebooks_fotw_11_1.png
../_images/dataset_notebooks_fotw_11_2.png

Effect of different Normalization Schemes#

[8]:

zscore_normalizer = ZScoreNormalizer(STATS_CLIP_RESCALE_PATH, band_order)
satmae_normalizer = SatMAENormalizer(STATS_SATMAE_PATH, band_order)

[9]:

norm_fig, normalized_batches = compare_normalization_methods(
    raw_batch, [zscore_normalizer, satmae_normalizer], datamodule
)
../_images/dataset_notebooks_fotw_14_0.png

Visualize Batch Data#

[10]:

fig, batch = datamodule.visualize_batch()
../_images/dataset_notebooks_fotw_16_0.png
Contents