High-resolution satellite imagery, crowdsourcing, machine learning and big data processing have been combined to create a digital representation of the built environment across the entire Australian continent’s 7.6 million square kilometres.
Electromagnetic spectral analysis and classification processes use algorithms to extract features from satellite-captured imagery, which are linked to a geocoded address.
The data captured includes: building footprints, heights, area, elevation, roof construction, solar panels, swimming pools and more, for every building with a roof area greater than nine (9) square metres — more than 15 million of them, as well as land and tree cover for the entire continent.
The resulting dataset has uses across many industries including telecommunications, emergency services and infrastructure planning. The presentation will describe the technology behind the data capture and case studies of its applications in:
- Fixed wireless telecommunications network design.
- Flood mapping and disaster planning in the Banana Shire, Queensland.
- The Greater Launceston Transformation Project — a Smart Cities and Suburbs Funded Program.