Crop Field Boundary Labels for Africa from 2017 to 2023

A multi-year collection of crop field boundaries digitized in Planet satellite imagery across Africa between 2017 and 2023. It was developed by Farmerline, Spatial Collective, and the Agricultural Impacts Research Group at Clark University with support from the Lacuna Fund. The dataset includes supplemental labels primarily from 2018, which have been used to train and validate a CNN-based cropland mapping model.

Use Cases

• Train computer vision models for cropland detection based on digitized field boundaries.
• Validate satellite-based land cover classification models using the provided labels.
• Analyze agricultural field patterns and sizes across different African regions.
• Develop geospatial models for agricultural impact assessment using the multi-year coverage.

Strengths

• Covers a multi-year time range from 2017 to 2023.
• Includes labels used for training and validating a CNN-based model, as referenced.
• Developed through a collaboration of multiple organizations and research groups.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and dataset size are unknown, which may limit suitability assessment.
• Last update date is unknown; freshness unverified.

Provenance

Source

Farmerline, Spatial Collective, and the Agricultural Impacts Research Group at Clark University.

Collection Method

Boundaries digitized in Planet satellite imagery.

Time Range

2017 to 2023

Freshness

Last updated date is unknown.

Geography

Africa