Research into automated quality control in the manufacturing industry

How do you prevent errors in production processes when there is a shortage of experienced skilled workers and the pressure to improve efficiency increases constantly? That question lies at the heart of DigitalZH, a research project within the Smart Sustainable Manufacturing Research Group at The Hague University of Applied Sciences.

Together with businesses and knowledge partners, the research group explored how smart cameras, sensors, and robots can help automate quality checks in the manufacturing industry. The aim is to detect errors earlier, reduce rework, and support businesses in making better, data-driven decisions.

The challenge of transparent materials 

A manufacturer of window systems submitted the first practical case: the automatic inspection of transparent acrylic sheets. That proved to be more difficult than it sounds. With transparent materials, reflections, glare, and refraction play a major role, which makes it difficult for standard vision systems to identify shapes or defects properly. “With transparent products, current camera systems mainly see what is behind the material”, researcher Henk van der Burgh explains. “That makes it difficult to determine whether the geometry of the product is correct”. Because the results were not sufficiently reliable, researchers at The Hague University of Applied Sciences and the company decided to develop a new approach together.

From manual checks to smart inspection

At the RE/Manufacturing Lab of The Hague University of Applied Sciences, the research team developed a demonstrator with a camera system that can automatically inspect acrylic sheets. The system reads QR codes milled into acrylic sheets, detects dust, scratches and other defects, and checks the geometry of components. Finally, it verifies whether the end product was assembled correctly. These checks take place before the bonding process, so that errors are identified early in the production process. “Until now, experienced quality control staff carried out these checks”, Henk says. “That knowledge is in their minds. We are now translating those quality criteria into a digital inspection system”. Automated inspection does not eliminate the role of people entirely, however. “Final inspection will always remain the responsibility of an operator”, he emphasises. “The system does help to identify deviations more quickly”.

Experimenting with light and technology

To inspect transparent materials properly, researchers experimented with different types of lighting and camera technology. Variations in light frequencies and illumination made defects and adhesive layers more visible to the camera. Based on those images, the software was trained to recognise deviations automatically. “What is interesting is that we are not working with the most expensive systems here”, says Jenny Coenen, Lector in Smart Sustainable Manufacturing at The Hague University of Applied Sciences. “We first demonstrate that something is technically possible, for example with simple cameras or open-source software. That also makes the research accessible and affordable for small and medium-sized enterprises. If a solution works, SMEB can then decide for themselves which system best suits their production environment”.

Robotic inspection

In addition to static inspection systems, DigitalZH also examines mobile inspection technology. In collaboration with SAM|XL, for example, the project is working with an inspection robot ‘Spot-the-dog’, which can inspect factories. This robot is equipped with cameras and sensors to detect, for example, air leaks in industrial installations. “We can also deploy this system during assembly.  If an inspection shows that there is a leak somewhere, staff will normally check all the connections with leak-detection spray”, Henk explains. “With the robot, we can identify much more quickly where the problem lies”. 

Accurate sorting

Other demonstrators being developed by the research team are systems that automatically check whether components were sorted into the correct bins. By combining vision technology with, for example, a weighing system, a machine can determine whether the correct number of components is present.

Making digital technology widely accessible

The DigitalZH project forms part of a broader ambition: to make digital technology accessible to SME in South Holland.  According to the researchers, a wider societal issue is also at play. Because of an ageing workforce and staff shortages, it is increasingly more difficult for manufacturing businesses to recruit enough experienced skilled workers. “Through inspection technology, we are capturing part of that knowledge, so that less experienced staff can also carry out these inspection tasks”, Henk says. 

Collaboration between education, research, and professional practice

An important feature of the project is the close collaboration between education, research, and business community. Students also play an active role, for example through internship and graduation projects.  One business is currently working with a graduating student on the automatic inspection of the adhesive layer between two acrylic sheets.  This example shows how practice-oriented research can contribute to innovation in the industry. Jenny: “Businesses often have a concrete problem, but not always the time or the resources to investigate it themselves. We do the exploratory work first: testing, experimenting, and showing what is possible. This is how we inspire businesses to take this further”.

More Information

Would you like to know more about the DigitalZH project or collaborate with the Smart Sustainable Manufacturing research gsroup? Please visit the projectpage or contact Jenny Coenen [email protected].