Western Australia Building Footprints from Satellite Imagery

Western Australia building footprints derived from 2018-2019 satellite imagery using deep learning segmentation and classification. The dataset results from a collaboration between the Department of Primary Industries and Regional Development and Landgate. Data is available in multiple geospatial formats including GeoJSON and SHP.

Use Cases

• Analyze building footprint density and distribution across Western Australia for urban planning.
• Validate and improve deep learning segmentation models for building detection from satellite imagery.
• Integrate building footprint data with agricultural or land-use datasets via spatial joins.
• Serve building geometry via WFS or WMS for web mapping applications in the region.

Strengths

• Derived from Vivid 2.0 satellite imagery captured from 2018 to 2019.
• Processed with a two-stage deep learning pipeline for segmentation and classification.
• Available in 8 geospatial formats including GeoJSON, SHP, and Geopackage.
• Published under a Creative Commons Attribution (CC-BY) license.

Limitations

• Imagery source is mostly from 2018-2019, potentially missing newer constructions.
• Coarse spatial boundaries may lack fine architectural detail for individual buildings.
• Unknown total feature count and specific attribute columns limit analytical scope.

Provenance

Source

Department of Primary Industries and Regional Development (DPIRD) and Landgate.

Collection Method

Deep learning image segmentation of satellite data, with further filtering by a binary classifier.

Time Range

Imagery mostly captured from 2018 to 2019.

Freshness

Last updated March 2026.

Geography

Western Australia.