Image Based Quality Inspection in Smart Manufacturing Systems: A Literature Review

Abstract

Quality inspection is an important component of today’s smart manufacturing systems (SMS). Their prominence stems from the objective of manufacturing companies to i) deliver high-quality products, ii) inspire brand loyalty, iii) keep within regulations, and iv) minimize waste of resources (incl. such as employee and machine time, scrap materials) to maximize profit. There is a wide variety of quality inspection systems deployed on the shop floor utilizing different technologies, ranging from human operators to high-fidelity sensor systems to image based systems. In this work we focus on the latter, modern image-based quality inspection systems and processes. Vision based quality inspection has seen an increase in applications and academic attention over the past decade, aligned with the dawn of Industry 4.0. On the one hand, digital camera systems (including the optics, sensors, and connectivity) have become more powerful and at the same time more affordable. On the other hand, the analytics – namely artificial intelligence and machine learning algorithms – have made tremendous advancements in terms of results as well as accessibility. Most notably, deep neural networks and deep learning have elevated the potential of computer vision in quality inspection applications to the next level. In this paper, we will conduct a comprehensive literature review analysing image based quality inspection systems in SMS over the last decade. We will focus particularly on the question of how image based in-situ quality inspection of three-dimensional parts is currently conducted. The results will provide an overview of the different available image based quality inspection approaches, their benefits and challenges, as well as specific application areas and/or industries.