Resource-efficient In-orbit Detection of Earth Objects

TL;DR

This work tackles the dual bottlenecks of downlink bandwidth and onboard compute for EO counting by introducing TargetFuse, a satellite-ground collaboration that runs a shallow counter in space and a deeper counter on the ground. It combines adaptive image tiling, clustering-based data deduplication, and bandwidth-aware downlinking throttling to minimize counting errors under energy and bandwidth constraints. Across diverse datasets and energy budgets, TargetFuse reduces CMAE by about $3.4\times$ on average and achieves up to $9.6\times$ improvements in bandwidth efficiency under constrained downlink. The approach demonstrates the practical viability of pragmatic space-ground analytics with COTS hardware, enabling more accurate EO analytics while respecting the harsh orbital environment.

Abstract

With the rapid proliferation of large Low Earth Orbit (LEO) satellite constellations, a huge amount of in-orbit data is generated and needs to be transmitted to the ground for processing. However, traditional LEO satellite constellations, which downlink raw data to the ground, are significantly restricted in transmission capability. Orbital edge computing (OEC), which exploits the computation capacities of LEO satellites and processes the raw data in orbit, is envisioned as a promising solution to relieve the downlink burden. Yet, with OEC, the bottleneck is shifted to the inelastic computation capacities. The computational bottleneck arises from two primary challenges that existing satellite systems have not adequately addressed: the inability to process all captured images and the limited energy supply available for satellite operations. In this work, we seek to fully exploit the scarce satellite computation and communication resources to achieve satellite-ground collaboration and present a satellite-ground collaborative system named TargetFuse for onboard object detection. TargetFuse incorporates a combination of techniques to minimize detection errors under energy and bandwidth constraints. Extensive experiments show that TargetFuse can reduce detection errors by 3.4 times on average, compared to onboard computing. TargetFuse achieves a 9.6 times improvement in bandwidth efficiency compared to the vanilla baseline under the limited bandwidth budget constraint.

Figures (12)

• Figure 1: A satellite periodically captures Earth images. Images are pre-processed into tiles before applying DNN models.
• Figure 2: The real-world data, including energy and computing power, is collected from Baoyun satellite.
• Figure 3: The overall workflow of TargetFuse.
• Figure 4: Effect of varied tile size on mAP accuracy and execution time of yoloV3-tiny model. mAP 0.5:0.95 is the average mAP over IoU threshold lin2014microsoft.
• Figure 5: The tiles generated through image tiling in the DOTA 9560031 dataset we utilize contain several images with similarities.