AI in Earth observation: The European Space Agency (ESA) has selected the PhiFireAI project of the Italian Thales Alenia Space team, together with ‘Irma’ of the French team, to test their technologies in orbit onboard the Φsat-2 satellite. A microsatellite, to be precise, a CubeSat 6U, scheduled to be launched in June 2024, it is specified on the TAS website, with the aim of studying, evaluating and demonstrating the full potential of artificial intelligence integrated onboard Earth-observing satellites.

The PhiFireAI project aims to monitor the Earth’s Space in the event of a fire: ‘Thanks to the processing of satellite images by artificial intelligence, it will be possible to determine whether the current image acquisition establishes the presence of a fire,’ is explained in the mission data sheet. If a fire is detected, therefore, the solution will make it possible to analyze the entire image to understand the percentage of hectares affected, establish the presence of water zones and safe places, and identify burnt areas.

AI working alongside humans

The innovation lies in the ability to run the algorithm set on Space-qualified hardware that can provide important information to the end user in real-time. The application developed by the PhiFireAI project will have a strong added value for feeding existing fire surveillance databases. It will also be possible to detect a hazard autonomously, directly onboard the satellite. All tests concerning the performance and adaptability of the model were carried out in collaboration with UBOTICA and CGI. An AI that, in this case, will not replace human beings in their work; on the contrary, it will improve the levels of efficiency, precision, safety and quality of performance.

AI works alongside humans and other machines to increase precision, efficiency and safety. In an interview published on the ESA website, Rochelle Schneider, head of AI applications at the agency’s Φ-lab, said: “Artificial intelligence harnesses computers and machines to work for us in performing highly repetitive tasks or functions that can be automated. This allows us to work more efficiently and focus on those activities that require irreplaceable human roles, such as emotional intelligence, human relationships and intuition”.

Built-in AI processing

“Earth observation is helped enormously by artificial intelligence. Several satellites now have AI computers on board,” Schneider continued, “to filter and process data, and downstream there are many applications that use artificial intelligence to create critical information for end users. Computer vision amply illustrates the contribution of artificial intelligence. We extract many features from satellite images, such as vegetation or coal mines, and although computers have been doing this for some time, they traditionally use algorithms or fixed models to locate the required objects”.

The Φsat-2 has built-in AI processing of images from its multispectral camera. It already includes AI applications on the satellite that will remove clouds from its images, create road maps and detect navigation. Still, it will also allow commercial companies to upload and run their own applications in Space. Φsat-2 and other satellites with AI on board could generate several new business models and commercial opportunities, even for companies outside the Space arena.