Innovation Portfolio
Applying the field-boundary approach to a broader set of physical features visible from space — training models capable of detecting roads, buildings, solar panels, trees, and more at global scale.
A global, regularly updated map of built and natural features gives governments, development organizations, and businesses a new window into how the physical landscape is changing. Applications range from tracking renewable energy deployment and disaster damage assessment to monitoring deforestation, land-use change, and urban growth in rapidly developing regions — all from a single, open platform.
Features of The World applies the same approach used for Fields of The World to a much broader set of physical features visible from satellite imagery.
Labeled data drawn from publicly available map sources, paired with satellite imagery from multiple providers to cover an initial set of built and natural features.
AI models trained to detect and map roads, buildings, solar panels, trees, and additional features.
Models flexible enough to identify new feature types without full retraining — enabling the dataset to expand beyond features anticipated at launch.
Methods for estimating feature coverage efficiently by drawing on neural world models developed in Fields of The World.