Insight4eo: A Customizable Product Line for Low-Latency on-Board Processing of EO Data

Robert Hinz, Juan Ignacio Bravo Pérez-Villar, Álvaro Morón, Enrique Sepúlveda Jorcano, André Batista De Oliveira, José Antonio Pulido Pavón, Rohaan Ahmed

Abstract: Onboard spacecraft processing of Earth Observation (EO) products introduces novel application possibilities. Techniques for data prioritization can be implemented to eliminate irrelevant data, conserving bandwidth and power. Additionally, novel emergency management solutions can be utilized to receive timely alerts on the ground before the complete dataset becomes available. However, incorporating such solutions on-board the spacecraft poses two primary challenges. Firstly, EO products often require L1/L2 data, which is not inherently accessible on board. Secondly, there is a need for an efficient and user-friendly execution environment capable of obtaining EO products with minimal latency. Deimos Space’s AI-enabled Insight4EO line of software products addresses these challenges by offering an optimized L1/L2 product designed for on-board processing of both Synthetic Aperture Radar (SAR) and optical payloads, along with a versatile multi-application processing environment based on Multi-Processor and FPGA technology. The capabilities of Insight4EO are highlighted through two on-board AI powered use cases: on-board cloud mask generation and extreme weather detection.

External IDs:dblp:conf/igarss/HinzPMJOPA24