Satellite image feature extraction processes, incorporating machine learning algorithms, are leading edge technologies being used to create geospatial data and map our planet. To date, these technologies have successfully extracted the attributes of over 8.8 million buildings across the Australian continent from satellite imagery for PSMA’s Geoscape®.

Geoscape is an analytics-ready, 3D built environment dataset providing location-based insights for every address in Australia. It captures building footprints and heights, rooftop materials, solar panels, swimming pools and more and links them to a geocoded address. The data can be visualised using mapping platforms and 3D modelling tools. It’s the first time location-based information and data analytics capability have been made available in this way on a national level.

PSMA has collaborated with DigitalGlobe, the leading provider of broad coverage, very high-resolution satellite imagery, to create Geoscape. DigitalGlobe’s ability to image Earth and extract features of the built environment delivers key components of our Geoscape dataset.

PSMA has been the vanguard of what is now materialising globally. DigitalGlobe will use the feature extraction processes that have allowed us to build Geoscape for Australia to create a new 2D building footprints product, initially for North America and extending globally over time. As part of this, PSMA will permit DigitalGlobe to on-licence internationally the Australian 2D building footprints that constitute a part of our Geoscape, thereby extending our partnership.

PSMA commends DigitalGlobe for pursuing an international building footprints dataset and building on the vision we had for Australia. Extracting building features from satellite imagery covering Australia’s 7.6 million square kilometres for Geoscape has demonstrated that continent-wide 2D building footprints is achievable.

DigitalGlobe’s satellite constellation

Header image: Dan Paull, CEO PSMA Australia (right) and Kevin Lausten, Senior Director DigitalGlobe

Comments are closed.