Sanchez, Daniela, Butt, Sajid et Qureshi, Ahmed.
2025.
« AI-enabled machine vision model for manual process monitoring and cycle time measurement ».
In Proceedings of the CSME-CFDSC-CSR 2025 International Congress (Montreal, QC, Canada, May 25-28, 2025)
Coll. « Progress in Canadian Mechanical Engineering », vol. 8.
Prévisualisation |
PDF
258 - AI-enabled machine vision model fo.pdf - Version publiée Licence d'utilisation : Tous les droits réservés aux détenteurs du droit d'auteur. Télécharger (361kB) | Prévisualisation |
Résumé
In high-variety, high-volume production, optimizing manual assembly is essential to sustain productivity, ensure flexibility, and minimize inefficiency. Unlike automated systems, manual assembly relies on human operators performing complex, repetitive tasks, which are prone to variations in execution time, skill level, and consistency. Traditional cycle time tracking methods, often manual and error-prone, lack precision and granularity. This research explores AI-powered machine vision to automate action recognition and cycle time measurement, offering a scalable, data-driven solution for real time process optimization. By leveraging advanced video analytics, the system identifies and classifies operator actions, recognizes key process steps, measures cycle times accurately, and detects inefficiencies. These insights enable precise time studies, helping manufacturers pinpoint bottlenecks, streamline workflows, and enhance efficiency. Unlike traditional sensor-based methods, AI-powered machine vision provides non-intrusive, adaptable tracking, making it ideal for dynamic production environments. Integrating machine vision aligns with digital manufacturing principles, using data-driven approaches for continuous improvement and quality control. The system delivers real-time insights and predictive analytics, enabling proactive adjustments to production planning and resource allocation. This approach strengthens flexible manufacturing, allowing rapid adaptation to product variability and shifting demands. Ultimately, the solution equips manufacturers to manage volume and diversity fluctuations without compromising quality or throughput, enhancing productivity and competitiveness.
| Type de document: | Compte rendu de conférence |
|---|---|
| Éditeurs: | Éditeurs ORCID Hof, Lucas A. NON SPÉCIFIÉ Di Labbio, Giuseppe NON SPÉCIFIÉ Tahan, Antoine NON SPÉCIFIÉ Sanjosé, Marlène NON SPÉCIFIÉ Lalonde, Sébastien NON SPÉCIFIÉ Demarquette, Nicole R. NON SPÉCIFIÉ |
| Date de dépôt: | 18 déc. 2025 15:32 |
| Dernière modification: | 18 déc. 2025 15:32 |
| URI: | https://espace2.etsmtl.ca/id/eprint/32494 |
Actions (Authentification requise)
 |
Dernière vérification avant le dépôt |