In-Orbit Commissioning Results of the ESA Φsat-2 Mission

The commissioning of the ESA Φsat-2 mission marks a significant milestone in the advancement of Earth observation technologies, particularly in the realm of artificial intelligence (AI) and edge computing. Launched on August 16, 2024, aboard the SpaceX Transporter 11 rideshare flight, the Φsat-2 mission aims to demonstrate the feasibility of running an entire data processing chain onboard, from Level 0 to Level 2, showcasing the latest advancements in edge computing and AI. The mission began with the successful separation of the Φsat-2 spacecraft from the launcher, followed by the establishment of communication 90 minutes post-launch over the Troll ground station. The Launch and Early Orbit Phase (LEOP) was completed within two days, transitioning the mission into the platform commissioning phase. This phase involved initial checks, including the activation of the payload and a series of image acquisitions over Portugal, Spain, and Greece. The platform commissioning phase, which lasted approximately four weeks, involved verifications on power generation, communications, and attitude determination and control, ensuring the spacecraft's readiness for the nominal mission scenario. Concurrently, the payload commissioning phase commenced, focusing on the complete chain and including the multispectral camera, payload pre-processor, and AI accelerator. The payload calibration campaign began during the second month of operations, moving the mission towards its operational phase. During the commissioning and calibration phase, over 100 GB of data was acquired, including Level 0 data over calibration sites, Level 1 onboard processed data to verify the correct application of the calibration parameters, the accuracy of the band co-registration and geolocation processor and finally Level 2 data to identify to compare the abilities of the AI applications to match the results obtained with synthetic data. Once in the operational phase the mission's primary objective is to fine-tune six onboard AI applications, which include deep compression, cloud detection, street detection, vessel detection, fire detection, and marine anomalies detection. The Φsat-2 mission is among the first to implement a complete data processing chain onboard, from Level 0 to Level 2, leveraging AI and edge computing to generate actionable information directly on the spacecraft. This capability allows for immediate availability of processed data to users upon download, significantly reducing the time between data acquisition and utilization. Looking ahead, the Φsat-2 mission will continue to focus on the fine-tuning of its AI applications before transitioning to routine acquisitions. The mission's full data catalogue, including Level 1 and Level 2 data, will be made freely available starting in February 2025. This open-access approach aims to foster innovation and collaboration within the Earth observation community, paving the way for future operational missions that leverage AI and edge computing technologies. This paper will provide an overview of the main results achieved so far, including an overview of the spacecraft performances and a focus on data quality obtained directly from onboard processing. In summary, the commissioning of the Φsat-2 mission represents a significant leap forward in Earth observation capabilities, demonstrating the potential of AI and edge computing to revolutionize data processing and utilization in space.