Google Research has unveiled an enhanced version of its Open Buildings dataset that uses artificial intelligence to monitor urban development across the Global South, now including the ability to track changes in buildings over time.
The Open Buildings 2.5D Temporal Dataset, released last month, leverages AI technology to analyse low-resolution satellite imagery, mapping approximately 1.8 billion buildings across 58 million square kilometers in Africa, South and Southeast Asia, Latin America, and the Caribbean.
The blog post stresses the possibility of change. We could see an unprecedented urban migration of approximately 2.5 billion people over the next few decades, primarily in developing nations. This presents significant challenges and opportunities for organisations and governments to manage such dramatic urban expansion according to Google Research product manager Olivia Graham.
The project represents a significant advancement in mapping underserved regions, where commercial satellite providers often lack coverage due to limited profit potential. Google's research team, headquartered in Ghana with collaborators in Tel Aviv and Zurich, developed innovative AI techniques to overcome the challenges of working with low-resolution satellite data.
The dataset has already proven valuable for humanitarian efforts. In Uganda, nonprofit organisation Sunbird AI is using the data to help address electricity access issues, where approximately 73% of the population lacks power. The temporal aspect of the new dataset allows officials to track rapid urban growth and adjust resource planning accordingly.
The system can also monitor the impact of natural disasters. Following the September 2018 earthquake and tsunami in Sulawesi, Indonesia, the dataset clearly showed changes in coastal development patterns, providing valuable information for disaster response and recovery efforts.