Computer Vision-based Learning for Event Modelling in Digital Twins

Abstract

This study explores the integration of computer vision (CV) techniques for machine learning development of digital twins for manufacturing applications. Digital twins represent a transformative approach to process monitoring and optimization in a field prone to uncertainties and operational inefficiencies. However, existing implementations using sensors for event modelling often encounter limitations in scalability and compatibility, especially with legacy machines, and in monitoring non-machine events such as maintenance events. The research aims to use computer vision as a tool to overcome these challenges. The proposed approach employs computer vision-based learning models to capture and analyse real-time data from physical processes

both machine and non-machine events, addressing the constraints of traditional digital twin systems and providing a more comprehensive and non-invasive monitoring. This approach enhances the adaptability and utility of digital twins, making them low-cost, more scalable, accurate, and responsive to user needs in complex manufacturing environments. © This is an open access article published by the IET under the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0/)