After successfully completing its nine-month commissioning phase, the miniature satellite Φsat-2 has begun delivering scientific data, officially starting its operational mission. This technological gem, launched in August 2024, represents a milestone in the way we observe and analyze our planet. Placed in orbit at an altitude of 510 kilometers, Φsat-2 uses advanced artificial intelligence algorithms directly in space to optimize the collection and processing of Earth images, opening the door to a new era of smart and efficient satellite monitoring. Its ability to independently decide which data to send back to Earth dramatically accelerates the availability of key information for scientists, crisis management agencies, and numerous other users.

The transition from the testing phase to the scientific phase, concluded in the second quarter of 2025, marks the beginning of the regular delivery of data that has the potential to transform numerous industries and scientific disciplines. The first image was delivered just four days after launch, testifying to the system's exceptional efficiency and reliability. Now, with fully calibrated instruments, the satellite is ready to fulfill its primary role – providing detailed insights into the state of the environment, from monitoring glacier melting to identifying illegal activities at sea.

A revolution in Earth observation: AI in space

What makes Φsat-2 truly special is its ability to process data on the satellite itself, known as "on-board processing". At the heart of this innovation is a powerful AI processor, the Ubotica CogniSAT, which runs a set of specialized applications. One of the fundamental tasks of the artificial intelligence on Φsat-2 is the automatic detection and rejection of images that are covered by clouds. Given that clouds often cover large parts of the Earth's surface, this functionality drastically reduces the amount of useless data that would otherwise be sent to Earth. This saves precious communication bandwidth and ground station resources, delivering only clean and usable images to users.

But its capabilities go far beyond cloud filtering. The artificial intelligence algorithms are trained to detect and analyze disaster-stricken areas. For example, the PhiFire AI application, developed by Thales Alenia Space, can identify forest fires in real time. The system not only detects the hotspot but also analyzes the surrounding area and classifies it into safe zones, already burnt areas, and water bodies, providing emergency response teams with crucial information for planning rescue and firefighting operations. Similarly, the satellite can quickly analyze the aftermath of earthquakes or floods, identifying passable access routes for rescue teams.

Maritime security is another key area of application. The vessel detection application, developed by the Portuguese center CEiiA, uses machine learning to identify and classify ships in multispectral images. This technology enables the monitoring of maritime traffic, as well as the fight against illegal activities such as illegal fishing or smuggling. The satellite can spot vessels even if they have turned off their automatic identification systems (AIS), thus providing authorities with a tool for monitoring remote and sensitive marine areas. In addition, specialized algorithms can detect marine pollution, such as oil slicks or harmful algal blooms, enabling rapid response and remediation.

The technical heart of the Φsat-2 mission

The Φsat-2 satellite belongs to the class of so-called "CubeSat" satellites. This is a standardized format of miniature satellites whose basic unit is a 10x10x10 centimeter cube. Φsat-2 is a "6U" CubeSat, meaning it is composed of six such units, and its total dimensions are just 22 x 10 x 33 cm with a mass of 8.9 kilograms. This modular and compact design allows for a significant reduction in development and launch costs compared to traditional, large satellites. CubeSats often "hitch a ride" into space as a secondary payload on rockets carrying larger satellites, which further cheapens the mission. It was launched on a SpaceX Falcon 9 rocket from the Vandenberg base in California.

It orbits the Earth at an altitude of 510 kilometers in a Sun-synchronous orbit (SSO). This type of orbit is nearly polar and is carefully calculated so that the satellite passes over any given point on Earth at the same local solar time. This ensures nearly identical lighting conditions for each overpass, which is crucial for comparing images taken at different time periods and for monitoring environmental changes, such as urban sprawl, deforestation, or ice melt.

The main instrument on the satellite is the Simera Space MultiScape100 multispectral imager, which generates images using seven spectral bands, from the visible to the near-infrared part of the spectrum, and one high-resolution panchromatic band. The different spectral bands allow for the analysis of various properties of the imaged surface – for example, the near-infrared band is extremely useful for assessing vegetation health. The panchromatic band provides sharp, detailed black-and-white images with a spatial resolution (GSD - ground sampling distance) of approximately 5 meters per pixel. This combination makes the instrument extremely versatile and suitable for a wide range of applications, including environmental monitoring, land management, agriculture, and cartography.

Images that change everything: Five views of our planet

To mark the beginning of the mission's scientific phase, five representative images were selected to demonstrate the satellite's wide range of capabilities in different applications and over diverse terrains. All images are shown in true color, using the red, green, and blue spectral bands, with the exception of the Bahia Blanca estuary image, which is shown in false color to highlight specific characteristics.

• Clavering Øer glaciers, Greenland: This image shows the vast ice masses on the east coast of Greenland. Data collected over polar regions is key to monitoring the impact of climate change, the dynamics of glacier melting, and changes in the ice sheet. Continuous monitoring of these sensitive ecosystems helps scientists to better understand and predict sea-level rise.


• Bahia Blanca estuary, Argentina: Displayed in false color using the near-infrared band, this image reveals a complex network of waterways and wetlands. The near-infrared spectrum is particularly sensitive to vegetation and water content, so this technique allows for a detailed analysis of vegetation health, sediment distribution, and water quality in the estuary, which is an important habitat for many plant and animal species.


• City of Innsbruck, Austria: The image of the alpine city of Innsbruck shows the satellite's utility for urban planning and management. Detailed high-resolution images can be used to monitor urban sprawl, analyze green spaces, identify urban "heat islands," and supervise infrastructure projects.


• Sediment in the Gulf of Tunis, Tunisia: This image clearly shows sediment plumes flowing from the Medjerda River into the Mediterranean Sea. Monitoring sediment transport is important for understanding coastal erosion, ocean currents, and the impact of agricultural activities and wastewater on the marine ecosystem.


• Ships in Port Said, Egypt: This image of one of the world's busiest ports at the northern entrance to the Suez Canal perfectly demonstrates the capability of the AI vessel detection application. The system can automatically identify and count ships, track their movement and traffic density, which is crucial for port management and ensuring maritime security.
  
  Source: European Space Agency