EarthSight: A Distributed Framework for Low-Latency Satellite Intelligence

2025-11-13

AI Summary

Traditional satellite remote sensing relies on full-image downlink before ground-based analysis, suffering from bandwidth constraints that induce delays of hours to days. Existing onboard machine learning (ML) approaches typically perform isolated inference per satellite, resulting in cross-satellite and cross-task redundancy that exacerbates power consumption and computational overhead. This paper proposes the first distributed collaborative inference framework tailored for satellite constellations, reframing intelligent image analysis as a “space–ground cooperative decision-making” problem. Our approach innovatively integrates a multi-task shared backbone network, a ground-station query-driven dynamic filtering and ranking mechanism, and a resource-aware scheduling strategy—enabling inter-satellite computational load sharing, global priority prediction, and early-stage redundancy elimination. In simulation, the average per-image computation time is reduced by 1.9×, and the end-to-end latency (from first contact to delivery) 90th-percentile drops from 51 to 21 minutes, significantly enhancing timeliness-critical applications such as disaster response.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Computer Vision: Remote Sensing / Geospatial AI
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines

Abstract

Low-latency delivery of satellite imagery is essential for time-critical applications such as disaster response, intelligence, and infrastructure monitoring. However, traditional pipelines rely on downlinking all captured images before analysis, introducing delays of hours to days due to restricted communication bandwidth. To address these bottlenecks, emerging systems perform onboard machine learning to prioritize which images to transmit. However, these solutions typically treat each satellite as an isolated compute node, limiting scalability and efficiency. Redundant inference across satellites and tasks further strains onboard power and compute costs, constraining mission scope and responsiveness. We present EarthSight, a distributed runtime framework that redefines satellite image intelligence as a distributed decision problem between orbit and ground. EarthSight introduces three core innovations: (1) multi-task inference on satellites using shared backbones to amortize computation across multiple vision tasks; (2) a ground-station query scheduler that aggregates user requests, predicts priorities, and assigns compute budgets to incoming imagery; and (3) dynamic filter ordering, which integrates model selectivity, accuracy, and execution cost to reject low-value images early and conserve resources. EarthSight leverages global context from ground stations and resource-aware adaptive decisions in orbit to enable constellations to perform scalable, low-latency image analysis within strict downlink bandwidth and onboard power budgets. Evaluations using a prior established satellite simulator show that EarthSight reduces average compute time per image by 1.9x and lowers 90th percentile end-to-end latency from first contact to delivery from 51 to 21 minutes compared to the state-of-the-art baseline.

Problem

• Reduces satellite image delivery delays from hours to minutes
• Optimizes onboard computation across multiple vision tasks
• Manages limited downlink bandwidth and onboard power constraints

Innovation

• Multi-task inference on satellites with shared backbones
• Ground-station query scheduler for request aggregation and prioritization
• Dynamic filter ordering for early rejection of low-value images