A framework for optical features selection and management for camera-only autonomous navigation in the proximity to small celestial objects
| dc.contributor.advisor | Chemak, L | |
| dc.contributor.advisor | Sanchez Cuartielles, Joan Pau | |
| dc.contributor.author | Di Fraia, Marco Z | |
| dc.date.accessioned | 2023-03-13T14:08:55Z | |
| dc.date.available | 2023-03-13T14:08:55Z | |
| dc.date.issued | 2021-08 | |
| dc.description | © Cranfield University 2021. All rights reserved. No part of this publication may be reproduced without the written permission of the copyright owner. | en_UK |
| dc.description.abstract | Small celestial bodies such as asteroids and comets are abundantly present in the Solar System, yet their surfaces remain largely unexplored. Achieving regular access to these surfaces would have a major impact on capabilities such as planetary defence and in situ resource utilisation and lead to significant scientific insights. However, missions close to small celestial objects remain challenging in at least two aspects: technically, due to weak gravity fields, complex operational environments and latency from long communication times, and commercially, with the applications still being few and cost-ineffective. A potential solution to reducing development and operational costs and obtaining robust, scalable operations, could be using small, camera-only spacecraft with an elevated degree of autonomy. Enabling a camera-based autonomy requires building appropriate computer vision pipelines. All computer vision pipelines start with the detection of features - salient patterns within the scene. This thesis presents multiple methods and tools enabling the appropriate selection and management of different features for autonomous navigation in proximity to asteroids. To that end, relevant contributions developed during this work consist of: The development of a software toolbox for prototyping and testing optical navigation technologies through a parametrisable synthetic 3D visual environment; An analysis of the response of feature detectors to internal factors (e.g., feature model) and external factors (e.g., illumination). This response, once known, can be used for designing the system or to obtain situational awareness An assessment of the response of template matching methods when the template (model) does not perfectly match the observed target (asteroid, with illumination). Through the above contributions, it was shown that considering environmental cues and the perception model helps in achieving robust camera-only navigation processes. This capability could lead to small satellites autonomously exploring hundreds or thousands of small celestial objects or be employed on more powerful spacecraft for redundancy. | en_UK |
| dc.description.coursename | PHD | en_UK |
| dc.identifier.uri | https://dspace.lib.cranfield.ac.uk/handle/1826/19289 | |
| dc.language.iso | en | en_UK |
| dc.relation.ispartofseries | PHD;PHD-21-DI FRAIA | |
| dc.rights | © Cranfield University, 2015. All rights reserved. No part of this publication may be reproduced without the written permission of the copyright holder. | |
| dc.subject | Space | en_UK |
| dc.subject | omputer Vision | en_UK |
| dc.subject | Optical Navigation | en_UK |
| dc.subject | Feature Detection | en_UK |
| dc.title | A framework for optical features selection and management for camera-only autonomous navigation in the proximity to small celestial objects | en_UK |
| dc.type | Thesis | en_UK |