End-to-End Deep Learning-Driven Automation for Enhanced Inspection in Industry
| dc.contributor.author | Salib, Philopatear | |
| dc.date.accessioned | 2024-05-27T15:30:44Z | |
| dc.date.available | 2024-05-27T15:30:44Z | |
| dc.date.issued | 2024-05-27 | |
| dc.date.submitted | 2024-05-24 | |
| dc.description.abstract | This study explores the integration of machine learning into industrial automation, specifically focusing on transformer inspection to address labor shortages and enhance operational efficiencies. It develops a specialized defect detection system using ResNet architectures with various depths, thereby significantly advancing fault detection capabilities within industrial environments. This research integrates an end-to-end system for automatic inspection, encompassing data collection, augmentation, and labeling of objects for inspection, alongside leveraging ResNet architectures for in-depth training and hyperparameter tuning. The research aims to reduce reliance on skilled labor while increasing the accuracy of inspections. A Raspberry Pi-based monitoring system is designed, implemented, and evaluated, revealing substantial improvements in both the precision and efficiency of transformer inspections. Achieving an accuracy rate exceeding 90\% stands out as a major accomplishment, emphasizing the robustness of the machine learning model and the effectiveness of its training and optimization processes. Comprehensive reliability and repeatability tests are conducted under real-time conditions, with multiple users adjusting the orientation and placement of objects in varied lighting and location settings. Despite these challenges, the system consistently and accurately determines the status of the objects, demonstrating its ability to operate effectively in diverse and unpredictable environments. This consistent performance confirms the system’s readiness for industrial use and its reliability, making it a critical solution for environments where high accuracy and consistency are paramount. The potential of deep learning to revolutionize industrial inspection processes is affirmed, paving the way for future enhancements and broader applications of the technology. | en |
| dc.identifier.uri | http://hdl.handle.net/10012/20604 | |
| dc.language.iso | en | en |
| dc.pending | false | |
| dc.publisher | University of Waterloo | en |
| dc.subject | inspection | en |
| dc.subject | transformers | en |
| dc.subject | deep learning | en |
| dc.subject | ResNet | en |
| dc.subject | zero-shot learning | en |
| dc.subject | multi-camera systems | en |
| dc.subject | computer vision | en |
| dc.subject | real-time | en |
| dc.title | End-to-End Deep Learning-Driven Automation for Enhanced Inspection in Industry | en |
| dc.type | Master Thesis | en |
| uws-etd.degree | Master of Applied Science | en |
| uws-etd.degree.department | Mechanical and Mechatronics Engineering | en |
| uws-etd.degree.discipline | Mechanical Engineering | en |
| uws-etd.degree.grantor | University of Waterloo | en |
| uws-etd.embargo.terms | 0 | en |
| uws.contributor.advisor | Melek, William | |
| uws.contributor.affiliation1 | Faculty of Engineering | en |
| uws.peerReviewStatus | Unreviewed | en |
| uws.published.city | Waterloo | en |
| uws.published.country | Canada | en |
| uws.published.province | Ontario | en |
| uws.scholarLevel | Graduate | en |
| uws.typeOfResource | Text | en |