New AI models integrate petabytes of Earth observation data to produce a unified data representation that revolutionizes global mapping and monitoring.
Every day, satellites capture information-rich images and measurements, providing scientists and experts with near real-time information about the Earth. This data is extremely impactful, but its complexity, multimodality, and refresh rate create new challenges in connecting disparate datasets and effectively leveraging them all.
Today we’re introducing AlphaEarth Foundations, an artificial intelligence (AI) model that acts like a virtual satellite. It accurately and efficiently characterizes the entire Earth’s landmass and coastal waters by integrating vast amounts of Earth observation data into a unified digital representation (“embedding”) that can be easily processed by computer systems. This allows the model to provide scientists with a more complete and consistent picture of Earth’s evolution, helping them make more informed decisions about important issues such as food security, deforestation, urban expansion, and water resources.
To accelerate research and uncover use cases, we are now releasing AlphaEarth Foundations’ annual collection of embeddings as a Google Earth Engine satellite embedding dataset. Over the past year, we’ve worked with over 50 organizations to test this dataset in real-world applications.
Our partners are already realizing significant benefits by using data to better classify unmapped ecosystems, understand agricultural and environmental change, and significantly improve the accuracy and speed of mapping efforts. In this blog, we’re excited to highlight some of their feedback and show you the tangible impact of this new technology.
