Pose estimation of Didymos' moon using CNN-based image processing algorithm for HERA mission

Aurelio Kaluthantrige, J. Feng, J. Gil-Fernández

Research output: Contribution to conferencePaper

9 Downloads (Pure)

Abstract

In this work we develop an appearance-based methodology to estimate the continuous pose (attitude) of Dimorphos during the COP using a Convolutional Neural Networks (CNN)-based Image Processing (IP) algorithm. The key of the developed algorithm is to leverage the lit limb of the secondary to estimate the quaternions representing the orientation of Dimorphos with respect to the spacecraft. By relying on the shape of the target, the developed algorithm is not constrained on the prior knowledge of the spinning state of the target and on the visibility of features on the surface of the target.


Original languageEnglish
PagesIAA-PDC-23-PO-111
Publication statusPublished - 6 Apr 2023
Event8th IAA Planetary Defense Conference (PDC) 2023 - Vienna, Austria
Duration: 3 Apr 20237 Apr 2023
Conference number: 8th

Conference

Conference8th IAA Planetary Defense Conference (PDC) 2023
Country/TerritoryAustria
CityVienna
Period3/04/237/04/23

Keywords

  • autonomous navigation
  • pose estimation
  • neural networks
  • image processing
  • appearance-based model

Fingerprint

Dive into the research topics of 'Pose estimation of Didymos' moon using CNN-based image processing algorithm for HERA mission'. Together they form a unique fingerprint.

Cite this