Φsat-2, a 6U CubeSat developed by the European Space Agency (ESA), is demonstrating the integration of artificial intelligence (AI) in Earth observation. Launched on August 16, 2024, Φsat-2 is equipped with a multispectral camera and onboard AI applications designed to process and analyze satellite imagery directly in orbit, enhancing data efficiency and decision-making capabilities.
The satellite's AI applications include:
Cloud Detection: Identifies and discards images obscured by cloud cover, ensuring that only clear, usable data is transmitted to Earth.
Street Map Generation: Converts satellite imagery into street maps to assist emergency response teams during disasters by identifying accessible routes.
Maritime Vessel Detection: Utilizes machine learning to detect and classify vessels, supporting maritime monitoring and environmental conservation efforts.
Onboard Image Compression and Reconstruction: Compresses images onboard to increase data download efficiency, with a focus on detecting buildings upon reconstruction.
Marine Anomaly Detection: Spots anomalies in marine ecosystems, such as oil spills and harmful algal blooms, in real-time.
Wildfire Detection: Provides critical real-time information for wildfire response teams, tracking fire spread and identifying potential hazards.
These capabilities are expected to significantly enhance disaster response, maritime monitoring, and environmental protection efforts by providing timely and actionable insights directly from space.