Open Cosmos and ESA launch AI app-powered satellite

The launch took place on 16 August, 2024 at Vandenberg Space Force Base in California, with Phisat-2 launching alongside other satellites onboard SpaceX's Falcon 9.

The Phisat-2 satellite combines a multispectral camera with a computer that allows AI apps to run onboard, via the NanoSat MO Framework (NMF). This framework has been developed specifically for small satellites and allows the apps to be installed, updated and operated remotely from Earth.

ESA, who named Oxfordshire-based Open Cosmos as the mission’s prime contractor, said the satellite measures just 22 x 10 x 33cm and carries a multispectral instrument that images Earth in seven different bands (plus a panchromatic band) in the visible to near-infrared part of the electromagnetic spectrum.

While AI has already proved invaluable for analysing terabytes of satellite data, Open Cosmos said that most of this AI processing takes place on the ground, after the data has been downloaded; the Phisat-2 mission instead allows this to happen directly at source.

Instead of downlinking reams of raw data, the onboard apps can process and transmit only the most essential information back to Earth, giving tangible benefits in data transmission efficiency and resulting in faster decision-making.

The partners said that this will prove critical for disaster response efforts, maritime monitoring, environmental protection, cloud detection and more.

At launch, Phisat-2 ran four AI apps, including CGI’s Sat2Map which can send critical data to response teams during emergencies and disasters such as floods and earthquakes. It uses real-time satellite imagery to create maps of affected areas for use by teams on the ground. 

The Cloud Detection app, developed by KP Labs, can detect, identify and classify clouds to give real-time insights into cloud distribution. This helps users know when and where to gather satellite imagery, making the images more efficient and useful.

The Vessel Detection app, developed by CEiiA, uses machine learning to detect and classify vessels based on Phisat-2 multispectral images for real-time maritime monitoring, while GEO-K’s Deep Image Compression app compresses the images captured on board for more efficient transmission to Earth.

Two more apps are to be uploaded now the satellite is in orbit, including IRT Saint Exupéry Technical Research’s Anomalies in Marine Ecosystems app – which uses machine learning to scan satellite imagery and spot anomalies in marine ecosystems – and PhiFire AI, a wildfire detection app created by Thales Alenia Space, that can send early-warning signs to response teams.

In a statement, ‍Nicola Melega, ESA’s Φsat-2 technical officer said: “The successful launch of Phisat-2 once again showcases the power of collaboration between ESA and innovative companies like Open Cosmos. This mission highlights how advanced AI technology can transform our ability to monitor and respond to changes on Earth, making space data more actionable and impactful than ever before.”

Commenting on the launch, Harshbir Sangha, Missions and Capabilities Delivery director at the UK Space Agency, said:‍ “Congratulations to Open Cosmos on the successful launch of the Phisat-2 satellite. This is a great example of how UK companies are driving innovation by exploiting new technologies like AI to increase the speed and efficiency of data processing – using the unique vantage point of space to deliver valuable insights into life on Earth.

“The UK is leading member of the European Space Agency, and we’re playing a major role in the development of future Earth observation missions in partnership with industry and academia.”