2026-09-20T03:29:42Zhttps://dspace.mit.edu/server/oai/requestoai:dspace.mit.edu:1721.1/1568272024-09-17T03:50:19Zcom_1721.1_7582com_1721.1_7581col_1721.1_131023
Frey, Daniel
Darby, Brady J.
Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science
2024-09-16T13:51:33Z
2024-09-16T13:51:33Z
2024-05
2024-07-11T14:36:50.410Z
https://hdl.handle.net/1721.1/156827
Developments of computer vision techniques in the past decade have rapidly accumulated and enabled the application of vision systems to use cases that were once out of reach. In conjunction with standard image processing techniques, deep learning models for vision tasks have received increasing attention, and they both see considerable utility in space exploration. Specifically, real-time obstacle detection and motion planning require advanced vision logic. However, retroactive data analysis is an area with less emphasis but promising application for computer vision. This thesis project explores how both image processing and deep learning-based computer vision methods can be leveraged to analyze drill bits on board the Mars 2020 Perseverance Rover, a Jet Propulsion Laboratory (JPL) mission. The effectiveness of thresholding and segmentation on two critical tasks, drill bit identification and mechanical wear detection, is demonstrated. Then, transfer learning of convolutional neural networks (CNNs) is applied to the same tasks, allowing comparison of results. This thesis also explores a means of presenting processed image outputs to non-technical operators in order to assist manual analysis of drill bit wear state.
M.Eng.
Massachusetts Institute of Technology
Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0)
Copyright retained by author(s)
https://creativecommons.org/licenses/by-nc-nd/4.0/
Computer Vision Techniques for Drill Bit Identification and Mechanical Wear Detection
Thesis
application/pdf
Master
Master of Engineering in Computer Science and Molecular Biology
Publication
unknown
unknown
<Publication xmlns="https://www.openaire.eu/cerif-profile/1.1/" id="3fb437f7-7c05-4703-bc5a-105bd7cc49d4">
<Type xmlns="https://www.openaire.eu/cerif-profile/vocab/COAR_Publication_Types">http://purl.org/coar/resource_type/c_1843</Type>
<Title>Computer Vision Techniques for Drill Bit Identification and Mechanical Wear Detection</Title>
<PublishedIn>
<Publication>
</Publication>
</PublishedIn>
<PublicationDate>2024-05</PublicationDate>
<Authors>
<Author>
<DisplayName>Darby, Brady J.</DisplayName>
<Affiliation>
<OrgUnit>
</OrgUnit>
</Affiliation>
</Author>
</Authors>
<Editors>
</Editors>
<Publishers>
<Publisher>
<DisplayName>Massachusetts Institute of Technology</DisplayName>
<OrgUnit />
</Publisher>
</Publishers>
<License>https://creativecommons.org/licenses/by-nc-nd/4.0/</License>
<Abstract>Developments of computer vision techniques in the past decade have rapidly accumulated and enabled the application of vision systems to use cases that were once out of reach. In conjunction with standard image processing techniques, deep learning models for vision tasks have received increasing attention, and they both see considerable utility in space exploration. Specifically, real-time obstacle detection and motion planning require advanced vision logic. However, retroactive data analysis is an area with less emphasis but promising application for computer vision. This thesis project explores how both image processing and deep learning-based computer vision methods can be leveraged to analyze drill bits on board the Mars 2020 Perseverance Rover, a Jet Propulsion Laboratory (JPL) mission. The effectiveness of thresholding and segmentation on two critical tasks, drill bit identification and mechanical wear detection, is demonstrated. Then, transfer learning of convolutional neural networks (CNNs) is applied to the same tasks, allowing comparison of results. This thesis also explores a means of presenting processed image outputs to non-technical operators in order to assist manual analysis of drill bit wear state.</Abstract>
<Access xmlns="http://purl.org/coar/access_right"
>
</Access>
</Publication>