Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

An Edge AI Solution for Space Object Detection

Wenxuan Zhang, Peng Hu

TL;DR

The paper tackles real-time space object detection onboard LEO satellites to support collision avoidance, proposing a hybrid CNN–ViT framework with SE attention (GELAN-ViT-SE) and introducing the SODv2 dataset for realistic benchmarking. The method combines CNN and Vision Transformer feature extraction with SE-based channel recalibration to improve detection of small objects while maintaining feasible edge-level computation. Experimental results show GELAN-ViT-SE achieving higher AP than GELAN-t and GELAN-ViT on SODv2 and operating with acceptable latency and memory on Jetson Orin Nano, illustrating practical onboard applicability. Overall, the work advances edge AI capabilities for space sustainability by enabling accurate, low-latency onboard SOD for collision avoidance.

Abstract

Effective Edge AI for space object detection (SOD) tasks that can facilitate real-time collision assessment and avoidance is essential with the increasing space assets in near-Earth orbits. In SOD, low Earth orbit (LEO) satellites must detect other objects with high precision and minimal delay. We explore an Edge AI solution based on deep-learning-based vision sensing for SOD tasks and propose a deep learning model based on Squeeze-and-Excitation (SE) layers, Vision Transformers (ViT), and YOLOv9 framework. We evaluate the performance of these models across various realistic SOD scenarios, demonstrating their ability to detect multiple satellites with high accuracy and very low latency.

An Edge AI Solution for Space Object Detection

TL;DR

The paper tackles real-time space object detection onboard LEO satellites to support collision avoidance, proposing a hybrid CNN–ViT framework with SE attention (GELAN-ViT-SE) and introducing the SODv2 dataset for realistic benchmarking. The method combines CNN and Vision Transformer feature extraction with SE-based channel recalibration to improve detection of small objects while maintaining feasible edge-level computation. Experimental results show GELAN-ViT-SE achieving higher AP than GELAN-t and GELAN-ViT on SODv2 and operating with acceptable latency and memory on Jetson Orin Nano, illustrating practical onboard applicability. Overall, the work advances edge AI capabilities for space sustainability by enabling accurate, low-latency onboard SOD for collision avoidance.

Abstract

Effective Edge AI for space object detection (SOD) tasks that can facilitate real-time collision assessment and avoidance is essential with the increasing space assets in near-Earth orbits. In SOD, low Earth orbit (LEO) satellites must detect other objects with high precision and minimal delay. We explore an Edge AI solution based on deep-learning-based vision sensing for SOD tasks and propose a deep learning model based on Squeeze-and-Excitation (SE) layers, Vision Transformers (ViT), and YOLOv9 framework. We evaluate the performance of these models across various realistic SOD scenarios, demonstrating their ability to detect multiple satellites with high accuracy and very low latency.
Paper Structure (5 sections, 3 figures, 2 tables)

This paper contains 5 sections, 3 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Model
  3. Dataset
  4. Performance Evaluation
  5. Conclusion

Figures (3)

  • Figure 1: The architecture of the RepNCSPELAN4_SE module. The SE layer is highlighted with a distinct background color.
  • Figure 2: The architecture of the GELAN-ViT-SE model.
  • Figure 3: Sample SODv2 images with detection bounding boxes and confidence scores for satellite objects.