Autonomous on-orbit satellite servicing benefits from an inspector satellite that can gain as much information as possible about the primary satellite. This includes performance of articulated objects such as solar arrays, antennas, and sensors. A method for building an articulated model from monocular imagery using tracked feature points and the known relative inspection route is developed. Two methods are also developed for tracking the articulation of a satellite in real-time given an articulated model using both tracked feature points and image silhouettes. Performance is evaluated for multiple inspection routes and the effect of inspection route noise is assessed. Additionally, a satellite model is built and used to collect stop-motion images simulating articulated motion over an inspection route under simulated space illumination. The images are used in the silhouette articulation tracking method and successful tracking is demonstrated qualitatively. Finally, a human pose tracking algorithm is modifxC;ied for tracking the satellite articulation demonstrating the applicability of human tracking methods to satellite articulation tracking methods when an articulated model is available.
Satellite Articulation Sensing using Computer Vision
2018
222 pages
Report
No indication
English
Computers, Control & Information Theory , Unmanned Spacecraft , Artificial intelligence , Artificial satellites , Computer vision , Computers , Data processing , Geosynchronous orbits , Image processing , Information processing , Information systems , Kalman filters , Neural networks , Particle swarm optimization , Pattern recognition , Simultaneous localization and mapping , Solar panels , Spacecraft , Two dimensional
Satellite Articulation Sensing using Computer Vision
AIAA | 2017
|Satellite Articulation Sensing Using Computer Vision (AIAA 2017-1329)
British Library Conference Proceedings | 2017
|