Lublin students build an industrial vision system for Wikpol

You’ve created a project called “Vision system for verifying the positioning of loads on a packaging completion line in a robotic palletising system”. Where did the idea for this project come from, and what inspired it?
Our journey began when we heard about the “Rising Innovators” competition organised by Lublin University of Technology for secondary school students in the region. The competition selects the best solutions addressing real technical problems for companies. When Wikpol submitted their challenge – a vision system to verify the placement of cartons on a palletising line – we immediately felt it was a unique opportunity to test ourselves in practice. The project brought together multiple technological areas, and the chance to have a real impact on the production process while collaborating with professionals was a huge motivation for us and an experience that no school lesson could provide.

How did you react when you found out that your school would be collaborating with Wikpol on a real production line problem?
We were very enthusiastic. It was exciting to take part in such an advanced and practical project. We felt motivated because we knew we could contribute directly to solving a real technological problem. Visiting Wikpol’s facilities also helped us understand the real production environment we would be working with.

The vision system for carton inspection sounds highly technical. How did you turn the competition brief into a concrete plan?
During a meeting at Wikpol, we went over the problem and the company’s expectations in detail. Together we defined a preliminary plan, which became the foundation for our work. We decided the system would generate a reference 3D point cloud based on the dimensions and arrangement of cartons, then compare it with the real image from the sensor – regardless of lighting conditions in the hall.

How did you divide responsibilities within the team? Did everyone know from the start what they were best at, or did that become clear as you worked?
We had already worked together in this team before, so we knew each other’s strengths and weaknesses. That helped us divide tasks efficiently. Each of us focused on different parts of the project: one person handled camera selection and testing, another analysed software and libraries, and the third managed reporting and data visualisation. To stay organised and keep track of progress, we used a Kanban board in Trello with “to do,” “doing,” and “done” columns, which made coordination much easier.

What was the biggest challenge at the very start of the project?
The main challenge was technical limitations – specifically, the school’s graphics cards weren’t compatible with the software needed for the vision system. This made it difficult to start and required consulting a specialist from Lublin University of Technology.

Can you explain exactly how that specialist helped you? Why was their support crucial?
Dr Rafał Stęgierski came to us with his laptop, equipped with an NVIDIA graphics card, which finally allowed us to run the library for generating point clouds. His help meant we could continue the project and also learn how to diagnose hardware issues and match tools to technological requirements. He also shared his experience with industrial and humanoid robots, which was inspiring and highlighted the importance of collaborating with experts in engineering projects.

Since you were creating a vision system, you surely had to select appropriate hardware, like a camera. How did that process work, and which model did you choose in the end?
Choosing the camera was critical because the system relies on it heavily. We evaluated several models based on technical specifications and budget. The final choice, Luxonis Oak-D-Pro PoE, was made with input from Wikpol, who felt it best suited the project. The camera could communicate via Ethernet, had the right resolution, and an IP65 rating, making it suitable for the variable conditions of a production hall.

You mentioned earlier the reference 3D point cloud. How did the installation of the point cloud library go? Did you run into any difficulties?
After connecting the camera to the computer, we tried installing the library but ran into several errors. It was frustrating – we checked forums, tried different approaches, even tested the software in a Linux virtual machine, but problems persisted. Thanks to Lublin University of Technology, we discovered it wasn’t the software at fault, but the hardware – the school’s graphics card was incompatible.

Despite your dedication, the system didn’t fully implement carton layout recognition. Were you disappointed, or proud of how far you got despite the limitations?
We weren’t disappointed at all – quite the opposite. We were proud of what we achieved. The project gave us practical knowledge in vision systems, point cloud generation, and integrating multiple technologies. Despite hardware and time limitations, we completed key stages, and the experience will be invaluable in future projects.

What did you learn about vision systems and their industrial applications? Is this a field you want to pursue further?
We learned that vision systems have broad applications in industry and require consideration of many factors, like varying lighting conditions. In our project, we needed a point cloud instead of a simple image, which demanded powerful hardware. This field fascinates us, and we see great potential in combining vision systems with robotics and production line automation.

What surprised you most when working on a real industrial problem? Did anything turn out differently than you expected?
Even with previous experience, we were surprised by how complex it is to integrate cameras, software, and robots in real production conditions. We knew hardware and time constraints might appear, but we didn’t expect the gap between theory and practice to be so significant.

How would you evaluate your collaboration with Wikpol? Has working with industry changed how you see technology and the future job market?
Working with Wikpol helped us understand how a company implementing industrial robotics operates. We saw which skills are critical and what challenges come with deploying technology in a real production environment. It showed us the importance of practical skills and collaborating with engineers and experts.

How has participating in this project influenced your plans for further studies or careers? Are you considering continuing in automation, robotics, or IT?
The project reinforced our desire to develop in automation, robotics, and IT. We plan to continue with higher education – each of us is considering different universities, but we might stay at Lublin University of Technology to deepen our knowledge in industrial robotics and vision systems.

Would you like to continue this project or take on similar challenges with other Lublin companies? What technological topics interest you most today?
Absolutely – we are open to more technological challenges. Working on projects with industrial companies provides real opportunities for growth. We are particularly interested in robotics, automation, vision systems, and point cloud analysis in industry. We want to implement solutions that genuinely support production processes and encourage companies to collaborate with young innovators.