New satellite aims to show how AI advances Earth observation

Synthetic intelligence applied sciences have achieved outstanding successes and proceed to indicate their worth as backbones in scientific analysis and real-world functions.

ESA’s new Φsat-2 mission, launching within the coming weeks, will push the boundaries of AI for Earth remark—demonstrating the transformative potential of AI for space technology.

Earth remark has, for many years, supplied a wealthy stream of actionable information for scientists, companies and policymakers. Because of new satellites and superior sensors, the size and high quality of obtainable Earth remark information have risen exponentially prior to now decade.

The combination of AI has considerably enhanced Earth remark. AI capabilities enable for extra information to be processed shortly and precisely, serving to to allow to rework huge quantities of uncooked information into actionable insights.

As a part of an initiative to advertise the event and implementation of revolutionary applied sciences onboard Earth remark missions, ESA launched Ф-sat-1 in 2020. It was ESA’s first experiment to show how synthetic intelligence can be utilized for Earth remark and paved the way in which for its successor: Φsat-2.

Φsat-2 is a devoted AI mission which can absolutely discover the advantages and capabilities of using prolonged onboard processing and additional show the advantages of utilizing AI for revolutionary Earth remark.

Measuring simply 22 x 10 x 33 cm, ESA’s Φsat-2 satellite is provided with a multispectral digital camera and highly effective AI laptop that analyzes and processes imagery in real-time—promising to ship smarter and extra environment friendly methods of monitoring our planet.

With six AI functions operating onboard, the satellite is designed to show photos into maps, detect clouds within the photos, classify them and supply perception into cloud distribution, detect and classify vessels, compress photos on board and reconstruct them within the floor decreasing the obtain time, spot anomalies in marine ecosystems and detect wildfires.

ESA’s Φsat-2 Technical Officer Nicola Melega, commented, “Φsat-2 will unlock a brand new period of real-time insights from space and can enable for customized AI apps to be simply developed, put in, and operated on the satellite even whereas in orbit. This adaptability maximizes the satellite’s worth for scientists, companies and governments.”

The Φsat-2 mission is a collaborative effort between ESA and Open Cosmos who serves because the prime contractor, supported by an industrial consortium together with Ubotica, GGI, CEiiA, GEO-Ok, KP-Labs, and SIMERA.

Φsat-2, which shares its trip into orbit with ESA’s Arctic Climate Satellite tv for pc, is scheduled to liftoff in July 2024 on a SpaceX Falcon 9 from the Vandenberg Air Power Base, California, within the U.S.

Φsat-2 carries a multispectral instrument that photos Earth in seven completely different bands and, via its AI functions, is able to many issues that may present actionable data on the bottom, together with:

Cloud detection

Not like conventional satellites that downlink all captured photos, together with these obscured by clouds, Φsat-2 processes these photos instantly in orbit, making certain that solely clear, usable photos are despatched again to Earth.

Developed by KP Labs, this software may also classify clouds and supply insights into cloud distribution. This offers customers extra flexibility when it’s time to determine whether or not a picture is usable or not.

Avenue map technology

The Sat2Map software, developed by CGI, converts satellite imagery into avenue maps. This functionality is especially useful for emergency response groups, enabling them to determine accessible roads throughout disasters corresponding to floods or earthquakes.

When the satellite orbits over the affected space and acquires photos, the photographs are handed to the onboard processer that may determine streets and generate a corresponding map.

Initially, this software might be demonstrated over Southeast Asia, showcasing its potential to help in disaster administration.

Maritime vessel detection

The maritime vessel detection software, developed by CEiiA, makes use of machine studying strategies to routinely detect and classify vessels in specified areas, facilitating the monitoring of actions like unlawful fishing. This software underscores the satellite’s position in supporting maritime safety and environmental conservation efforts.

On-board picture compression and reconstruction

Developed by GEO-Ok, this software is accountable for compressing photos on board. By considerably decreasing file sizes, this software will increase the quantity and velocity of information downloads. After being downlinked to the bottom, the photographs are reconstructed utilizing a devoted decoder. The primary demonstrations of this expertise will happen over Europe, specializing in the detection of buildings.

Φsat-2’s capabilities have been additional expanded with the incorporation of two extra AI functions that might be uploaded as soon as the satellite is in orbit.

These AI functions had been the successful entries within the OrbitalAI problem organized by ESA’s Φ-lab and was designed to offer firms the change to pioneer in-orbit Earth remark information processing. The successful functions are:

Marine anomaly detection

Developed by IRT Saint Exupery Technical Analysis, this software makes use of machine studying algorithm to identify anomalies in marine ecosystems—figuring out threats to the marine ecosystem corresponding to oil spills, dangerous algae blooms and heavy sediment discharges in real-time.

Wildfire detection

The wildfire detection system, developed by Thales Alenia Area, makes use of machine studying to offer important real-time data to response groups. The device gives a classification report that helps firefighters find wildfires, monitor hearth unfold and determine potential hazards.