Satlas: Given enough training examples, modern deep learning methods can extract data like the positions of wind turbines from satellite imagery as accurately as humans can.

The Allen Institute for AI has unveiled Satlas, a new platform harnessing AI to analyze global geospatial data derived from satellite imagery. Satlas presently provides three monthly-updated data products:

• Marine Infrastructure: This includes the positions of offshore wind turbines and platforms.
• Renewable Energy Infrastructure: It offers positions of solar farms and onshore wind turbines.
• Tree Cover: This data reveals the tree canopy coverage percentage within 100 m² cells.

The data is accessible under an open license (ODC-BY) and can be visualized on the Satlas Map or downloaded for offline analysis. The platform aims to expand its geospatial data offerings over time.

Access to timely geospatial data is vital for informed decision-making in areas like emissions reduction, disaster relief, and urban planning. Existing global geospatial data sources are often fragmented and outdated, making manual curation challenging.

Satlas overcomes this by utilizing AI and deep learning to analyze satellite images, including those from Landsat and Sentinel-2, which cover most of the Earth regularly. Modern deep learning techniques allow for accurate extraction of data like wind turbine positions. Satlas employs high-accuracy deep learning models to produce updated global geospatial data snapshots each month.

The potential applications of Satlas’ geospatial data are diverse, including environmental monitoring and planetary analysis. Researchers are exploring uses such as vessel movement trajectory classification.

Related Story:  Israeli Startups ASTERRA, Albo Climate Use Satellites And AI To Hunt For Lithium And Carbon