SCPose-MLite: A Lightweight Neural Network for On-Orbit Attitude Estimation of Space Targets

  • Yuran Chen
  • , Xuesong Wu
  • , Kangjia Fu
  • , Qi Zhang
  • , Sunquan Yu
  • , Rui Zhong
  • , Xiucong Sun
  • , Xiang Zhang
  • , Teng Yi*
  • *Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Accurate pose estimation of space targets is of great significance for conducting on-orbit rendezvous, space debris removal and other tasks.In this paper,we propose a lightweight deep learning model, SCPose-MLite, which is based on the deep learning framework of URSONet and adopts MobileNet-V2 as the backbone network.By integrating the mixed pooling module, SCPose-MLite further enhances the generalization ability of the model while maintaining high accuracy.The experimental results show that SCPose-MLite is close to URSONet in terms of pose estimation accuracy, and at the same time, it has improved the generalization factor by 2.32 times, which can better meet the actual needs of space target pose estimation.Furthermore, this paper attempts to run the lightweight model on different devices, demonstrating that the training optimization brought by the mixed-precision quantization strategy and high-computing-power devices may fail in the training of low-parameter networks.

Original languageEnglish
Title of host publicationRCAR 2025 - IEEE International Conference on Real-Time Computing and Robotics
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages606-611
Number of pages6
ISBN (Electronic)9798331502058
DOIs
StatePublished - 2025
Event2025 IEEE International Conference on Real-Time Computing and Robotics, RCAR 2025 - Toyama, Japan
Duration: 1 Jun 20256 Jun 2025

Publication series

NameRCAR 2025 - IEEE International Conference on Real-Time Computing and Robotics

Conference

Conference2025 IEEE International Conference on Real-Time Computing and Robotics, RCAR 2025
Country/TerritoryJapan
CityToyama
Period1/06/256/06/25

Fingerprint

Dive into the research topics of 'SCPose-MLite: A Lightweight Neural Network for On-Orbit Attitude Estimation of Space Targets'. Together they form a unique fingerprint.

Cite this