April 9, 2026
With CT, X-ray and AI, sawmills can see inside the wood – and control how it is sawn
As raw material prices rise and margins are squeezed, every decision in production becomes critical. In sawmills, this has led to a shift from experience-based assessments to data-driven optimization, where CT scanning, X-ray and AI are used to analyze and control how each log should be sawn.
Traditionally, much of sawmill production has been based on experience. Logs have been sorted and sawn based on external dimensions and assumptions about what they look like on the inside.
It’s like having a sausage, but you don’t know what’s inside it. You see the shape, but you don’t know whether it’s a kabanoss or a salsiccia, says Jens Flodin, Production Manager at Norra Timber.
Norra Timber operates three sawmills in northern Sweden and produces close to 800,000 cubic meters of sawn timber annually. In a production of that scale, every decision about the raw material has a direct impact on the outcome, and for a long time, those decisions have been based on experience.
With CT scanning, the conditions change. Instead of relying on what is visible on the outside, the internal properties of the log can be analyzed before sawing and used as the basis for how each log is processed.
A three-dimensional image of every log
During log sorting, the logs pass through a rotating CT scanner. Using X-ray technology, density differences in the material are measured, and the results are built up into a three-dimensional model of the log’s internal structure – slice by slice, with a resolution of around ten millimeters.
Every new log that pops up is fully measured. We get a complete 3D reconstruction where we can see knots, cracks and density variations, says Jens Flodin.
This information is used directly in optimization. For each log, up to 35 different sawing alternatives are tested, where different combinations of dimensions and products are simulated before a decision is made.
We test up to 35 different sawing patterns for each log and choose the one that delivers the best result, says Jens Flodin.
When data starts learning from data
The technology used at Norra Timber’s facilities has been developed by MiCROTEC, which works with measuring and interpreting the internal properties of wood to maximize value in wood processing.
The combination of CT and AI makes it possible to predict sawmill output with significantly higher precision than before. We can use the three-dimensional data to understand how the material actually behaves throughout the entire process, says Philipp Bock, CTO at MiCROTEC.
Wood differs from many other materials in that every log is unique. This places high demands on both measurement and analysis compared to more homogeneous materials.
It’s about combining sensor data with AI in a way that allows us to identify patterns in a material that is not standardized. Only then can we truly start to predict the outcome, says Philipp Bock.
The systems are built on collecting large volumes of data and using them to train the models. The combination of three-dimensional CT data and AI makes it possible to identify patterns and characteristics that would otherwise be difficult to detect.
AI can only be as good as the data it is trained on. That’s why the link to actual production output is crucial for the systems to become accurate, says Philipp Bock.
When the systems are interconnected, data can be shared between different stages of the process, meaning that improvements in one part of production can have effects elsewhere.
In practice, one scanner can learn from another scanner. That allows us to improve precision across the entire system, not just at a single point, says Philipp Bock.
More value from every log
For Norra Timber, this represents a shift in how the raw material is used. Instead of adding extra safety margins to ensure the outcome, sawing can be controlled more precisely toward the product that is actually in demand.
We can use a smaller log to produce a product that is actually slightly better than before, says Jens Flodin.
This makes it possible to increase the value of each individual log, while at the same time reducing the amount of material that does not meet specifications. The increased precision also means that the raw material can be used more efficiently.
You don’t need to use more timber to make money – you can use the timber you have in a smarter way, he says.
This doesn’t stop at how an individual log is sawn. The same data can be used to see what actually comes out at the other end of the process and compare it with what was originally planned.
Previously, people did most of the interpretation. Now their role is increasingly to monitor and fine-tune the systems, says Philipp Bock.
The full article from Wood Products & Technology can be found here (in Swedish).
About Wood Products & Technology
Wood Products & Technology is the Nordic region’s largest meeting place for the wood industry, held in Gothenburg, Sweden, 25–27 August. The entire value chain gathers — from sawmills and production to architecture and urban development — to share knowledge, discover new technology and do business. Register for a free submission ticket »
Visit MiCROTEC at booth B07:52.