Toward Real-Time Earth System Intelligence Through Advancements in Onboard AI Processing
A review
Domenico Barretta (European Space Agency (ESA), Edilizia e Ambiente)
Luca Marini (TU Delft - Aerospace Engineering)
Francesca Razzano (Centro Direzionale Isola C4)
Giorgia Guerrisi (University of Rome Tor Vergata)
Sanna Sandberg (European Space Agency's (ESA's) Φ-Lab)
Roberto Del Prete (European Space Agency (ESA))
Valentina Zancan (Thales Alenia Space Italia)
Jian Guo (TU Delft - Aerospace Engineering)
Gabriele Meoni (TU Delft - ESA X, TU Delft - Aerospace Engineering)
More Authors (External organisation)
More Info
expand_more
Abstract
The Earth Observation (EO) sector is rapidly evolving into Earth Action (EA), with onboard Artificial Intelligence (AI) processing emerging as a key enabler. This technology offers strategic advantages through 1) enabling autonomous and low-latency EO missions with adaptive data processing capabilities that overcome the limitations of ground-based post-processing in handling the vast data volumes produced by growing satellite constellations and 2) supporting the evolution toward AI-driven and distributed EO mission architectures. This review explores the key technological advancements, mission architectures, and emerging paradigms that are shaping the next-generation EO systems. Pioneering EO missions are presented to showcase current capabilities, while the commercial, technical, and operational implications are analyzed alongside key challenges and ongoing research efforts (January 2020 to March 2025) aimed at enabling Real-Time (RT) Earth system intelligence.