African Building Footprints from Aerial Drone Imagery

Drone imagery from 10 different cities and regions across Africa supports mapping for disaster risk management. The dataset was developed for the Open Cities AI Challenge by MLHUB to accelerate open-source AI model development. It was last updated in 2023.

Use Cases

• Train semantic segmentation models to identify building footprints from high-resolution aerial imagery.
• Benchmark model performance across diverse urban landscapes in 10 distinct African regions.
• Develop disaster risk assessment tools by analyzing building density and spatial distribution from geospatial labels.

Strengths

• Covers 10 distinct cities and regions across Africa, providing geographic diversity.
• High-resolution drone imagery suitable for detailed building footprint analysis.

Limitations

• Unknown dataset size, row count, and specific image resolution.
• Potential class imbalance if non-building areas significantly outnumber building pixels.
• Geographic scope is limited to specific African cities, not continent-wide.

Provenance

Source

MLHUB, developed for the DrivenData Open Cities AI Challenge.

Collection Method

Collected via drone aerial imagery.

Freshness

Last updated in 2023.

Geography

10 cities and regions across Africa.