Search papers, labs, and topics across Lattice.
This paper introduces OrbitALIF, a federated learning framework designed for efficient cloud removal in low-earth-orbit satellites, addressing the limitations of traditional ground-based processing methods. By leveraging a compact spiking neural network with advanced attention mechanisms, OrbitALIF enables onboard training and inference, significantly reducing energy consumption during cloud removal tasks. The results demonstrate that OrbitALIF achieves competitive performance while consuming only 0.287 mJ per inference, representing a remarkable 98.6% energy reduction compared to conventional artificial neural networks.
Achieving a 98.6% energy reduction in cloud removal for LEO satellites could revolutionize Earth observation capabilities.
Low-earth-orbit (LEO) satellites enable high-resolution, large-scale Earth observation for applications such as disaster monitoring and environmental surveillance. However, cloud coverage often obscures the Earth's surface, and conventional cloud-removal pipelines that download cloudy images to ground stations for processing suffer from limited contact windows, constrained satellite-to-ground bandwidth, and high latency. In this work, we propose a novel satellite federated learning framework for cloud removal across LEO constellations, named orbital attention leaky integrate-and-fire (OrbitALIF). OrbitALIF performs both onboard training and inference using a compact 2.30,M-parameter spiking neural network (SNN) backbone with an adaptive gated fusion module (AGFM) and a spectral-spatial hybrid attention module (SHAM), combined with a decentralized federated learning strategy that shares model weights via inter-satellite links. Our experiments show that OrbitALIF achieves competitive cloud removal quality while consuming only 0.287,mJ per inference on neuromorphic hardware, a 72.3 times (98.6%) energy reduction versus an equivalent artificial neural network (ANN).
