agriculture cog labeled land cover machine learning satellite imagery
Crop field boundaries digitized in Planet imagery collected across Africa between 2017 and 2023, developed by Farmerline, Spatial Collective, and the Agricultural Impacts Research Group at Clark University, with support from the Lacuna Fund (Estes et al, 2024; Wussah et al. (2023)). This dataset has been further supplemented by additional labels collected primarily for for 2018 over a subset of countries, which provide an example of their application in training and validating a CNN-based cropland mapping model (Khallaghi et al. 2025).
Updated versions of the dataset are added as they are developed
Planet NICFI participant license agreement
Information on the primary dataset can be found here. Documentation for added labels is available here.
The Agricultural Impacts Research Group
See all datasets managed by The Agricultural Impacts Research Group.
A region-wide, multi-year set of crop field boundary labels for Africa was accessed on DATE
from https://registry.opendata.aws/africa-field-boundary-labels. Primary dataset: Estes et al. (2024). A region-wide, multi-year set of crop field boundary labels for Africa. arXiv:2412.18483. Additional labels: Khallaghi et al. (2025). Generalization enhancement strategies to enable cross-year cropland mapping with convolutional neural networks trained using historical samples. Remote Sensing, 17(3), 474.
arn:aws:s3:::africa-field-boundary-labels
us-west-2
aws s3 ls --no-sign-request s3://africa-field-boundary-labels/
arn:aws:s3:::africa-field-boundary-labels/extra
us-west-2
aws s3 ls --no-sign-request s3://africa-field-boundary-labels/extra/