The AI program conducts research and development in industrial AI, with a focus on our core business. The aim is to investigate how the technology can contribute to improved production. Recent developments in hardware, software, and AI models have made it significantly easier to use AI in actual industrial operations.
A concrete example can be found at the Aitik open-pit mine. Here, we extract 38–42 million metric tons of ore annually, which is then transported by large mining trucks to be crushed and concentrated. The enormous volumes take a heavy toll on the truck beds, where damage can lead to increased maintenance costs, equipment damage, and safety risks.
To detect damage earlier and improve maintenance planning, we are now investigating how cameras, computer vision, and AI can be used to monitor the condition of the truck beds.
How can AI detect truck bed damage earlier?
For many years, truck beds at Aitik have been inspected manually or with drones. To better track how damage appears and develops over time, we installed a camera above the KID2 crusher in 2023, where the trucks unload their material.
When a truck has unloaded, photos are taken automatically. The project is then being developed in stages: from collecting the right images and getting the technology to work in the demanding mining environment, to training the AI model to identify different types of damage.
Why AI was the right tool
Photographing the truck beds is one thing. Being able to quickly interpret the images and determine what they show is another. This is where AI comes into play.
By training models on image data, AI can learn to recognize and distinguish between different conditions and anomalies. The goal is for the technology to eventually be able to classify different types of damage on the truck beds.
"Computer vision is already a mature AI application. The next step is to classify different types of damage: what are we actually seeing? Over time, this could become a service that alerts the truck repair shop when something needs attention", says Rasmus Tammia.
AI is also used to filter out low-quality images. This allows maintenance staff to access more relevant information without having to manually review the entire image stream.
Keeping people at the center
Technology is only part of the solution. One of the biggest challenges in the project has been understanding what end users actually need in their daily work. That’s why the project is being developed and tested in close collaboration between Aitik’s IT, maintenance, and operations departments.
"It’s about understanding the everyday problems end users face. We’re building these AI applications to support the operator. People are still at the center in one way or another", says Rasmus Tammia.
