On-board AI chain reduces Earth data using wildfire segmentation
A Generic On-Board AI Processing Chain for Earth Observation: From Application Development to Embedded Deployment and Mission-Level Orchestration
Artificial IntelligenceHardware ArchitectureDistributed, Parallel, and Cluster Computing
Summary
Satellites gather more images than they can send back to Earth, so it helps to process the data onboard to only send important results. The authors created a flexible AI processing system that runs directly on the satellite hardware to spot things like wildfires in images. They made sure the AI works well after being simplified for embedded chips and built a workflow to combine tile-based AI results into full images. This system also organizes and runs the entire process in parallel on the satellite, making onboard AI more practical for real space missions.
What this means in practice
- •For satellite mission teams: Deploy complete AI processing chains on satellites to reduce data sent to Earth by detecting wildfires and other features onboard.
- •For embedded ai developers: Use a modular workflow and unikernel orchestration approach to run efficient AI inference at scale on embedded heterogeneous hardware.
Authors
Michel-Fran{\c c}ois Foulon, Liz Ramos Medina, Thomas Goudemant, Clotilde Szywala, Virgile Robles, Cl{é}ment Petitjean, Ad{è}le Hankache, Omar Hlimi
Abstract
Earth observation satellites acquire more data than they can downlink. Extracting information on board with artificial intelligence (AI) turns the satellite into a data-reduction stage that downlinks compact products, such as masks or detections, instead of full scenes; deploying it in a mission, however, involves the whole processing chain and not only the neural network. We present a generic on-board AI processing chain along three axes, with wildfire segmentation as the worked example among ship detection and cloud segmentation. (i) The AI model is tuned (spectral-band selection, input normalization), quantized to 8-bit integers (int8) and ported to an AMD Versal VCK190, with its excellent segmentation performance preserved on the board. (ii) A modular tiler-inference-mosaicer workflow, reused unchanged across applications, brings the tile-level model to scene level, turning a full acquisition into a scene-level mask or detection list. (iii) Each step is packaged as a unikernel and orchestrated on board by ORCHIDE, a cloud-native orchestrator. The chain is characterized end to end, from the standalone execution on the target to its execution under orchestration with parallel workflows on a representative test bench. Beyond AI inference on individual tiles, this approach aims at increasing the maturity of on-board AI deployment by addressing a complete processing chain, from the acquired scene to the mission-level execution.