Lubumbashi Building Footprints from 30 cm Satellite Imagery

Lubumbashi, Democratic Republic of the Congo, is covered by this dataset of 8,498 image tiles and 148,459 building footprint labels. It was produced for the ramp project by MLHUB and last updated in January 2023. The dataset includes high-resolution 30 cm imagery from Maxar ODP and corresponding building vector labels in tile/label pairs.

Use Cases

• Train building segmentation models based on high-resolution 30 cm satellite imagery.
• Analyze urban and peri-urban development patterns based on building footprint vector labels.
• Benchmark geospatial machine learning models on a dataset containing 148,459 individual buildings.

Strengths

• Contains 148,459 individually labeled building footprints.
• Imagery has a high spatial resolution of 30 cm.
• Includes 8,498 pre-chipped 256 x 256 pixel tile/label pairs for training.

Limitations

• Dataset is a Tier 2 ramp dataset, meaning it has not been thoroughly reviewed and improved.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

MLHUB for the ramp project, using Maxar ODP satellite imagery.

Collection Method

Likely created by applying machine learning models and/or manual annotation to satellite imagery.

Freshness

Last updated 2023-01-01 00:00:00; freshness should be verified.

Geography

Lubumbashi, Democratic Republic of the Congo, covering urban, peri-urban, and rural areas.