Satellite models update efficiently with limited bandwidth using NE-LoRA

Resource-Aware Parameter-Efficient Model Adaptation for Onboard High-Dimensional Data

Computer Vision and Pattern RecognitionDistributed, Parallel, and Cluster ComputingNetworking and Internet Architecture

Summary

Updating onboard satellite models is difficult because sending large amounts of new data to space is slow and expensive. The authors propose a method called NE-LoRA that updates only small parts of big models, saving communication bandwidth while keeping good accuracy. NE-LoRA combines simple and complex updates together and uses a special training approach to improve performance. Tests show it works well on hyperspectral satellite data compared to older methods.

What this means in practice

  • For satellite operations teams: Reduce the size of model updates sent to satellites in LEO for hyperspectral imaging without losing accuracy, enabling more frequent onboard model improvements.
  • For remote sensing software engineers: Deploy efficient model update methods on satellite platforms that handle complex spectral-spatial data under strict communication limits.

Authors

Qiyang Zhang, Xinhao Li, Lei Shi, Zheng Lin, Jinfeng Wen, Ao Zhou, Shangguang Wang

Abstract

Onboard satellite models often require frequent updates, but the weights adapted to earlier data distributions can quickly become outdated. However, updating large-scale model parameters in orbit presents significant challenges due to the limited uplink bandwidth of Low Earth Orbit (LEO) satellite systems, particularly for hyperspectral satellite imagery, where high-dimensional spectral-spatial inputs lead to increased model size and update costs. Existing full fine-tuning methods are thus expensive to retrain and difficult to deploy under strict communication constraints. To address this challenge, we propose NE-LoRA, a parameter-efficient adaptation framework for bandwidth-constrained onboard hyperspectral model updates. NE-LoRA combines a primary low-rank branch with a nonlinear auxiliary branch to capture both global update trends and complex spectral-spatial variations. Additionally, we introduce a differentiated training strategy for multi-matrix adapters, motivated by the asymmetric initialization and gradient dynamics of different adapter matrices. Experiments on four hyperspectral datasets and three representative backbone models demonstrate that NE-LoRA consistently outperforms LoRA-based baselines and remains competitive with, and in several cases superior to, full fine-tuning. Across the evaluated settings, NE-LoRA updates only a small fraction of the total parameters on average while preserving low deployment overhead, offering a favorable accuracy-communication trade-off for onboard hyperspectral adaptation.