The role of machine vision in industry 4.0: A textile manufacturing perspective

As we move forward to the fourth industrial revolution, findings in the state-of-the-art enhance the technological advancements in the research community as well as the business one, in particular the textile manufacturing, being here reviewed. In industry 4.0, the silos-free integrated systems sense the entire manufacturing environment by using sensor- and vision-based systems while enable multi-correlation zero-faults results with the minimum required data. Furthermore, the role of computer vision is vital in the deployment of cyber-physical systems, as the rich textural information provided by images through the scene’s appearance, can be translated into various findings. The paper at hand examines the existing appearance-based techniques in textile manufacturing over the previous five years. Moreover, open challenges and research gaps are highlighted. Through our analysis, we show that high performances are achieved mainly when vision-based systems are employed. Opportunities regarding one-batch fabric production without the need to promote multi-level integration are also proposed. Lastly, our findings exhibit that the self-adaptable visual features in textile manufacturing can be integrated with existing systems, e.g., cyber-physical ones, to observe the fabric factory in real-time while defects are corrected without scraps creation.