Sawmill industry systems need to stay up to date, adapt to changes, and continue working even when the technology around them changes. Software is never finished – but it still needs to last for decades.
Panu Toukola is the CTO of Lisker Oy and responsible for software suites, product development, and the direction of technology together with his team. In this article, he provides perspectives into the software running behind machine vision, the importance of architecture, and customer-oriented development.
Machine vision systems always have at their heart software suites that combine images, sensor data, and process data into optimal production solutions. Lisker’s software development covers the entire chain: the implementation of imaging, data processing, and algorithms into functional applications.
The in-house software team at Vääksy works closely together with systems development and other specialist teams. The work includes not only the building of new features, but also the continuous development of systems on the basis of information gained from sawmill conditions and customer feedback.
“Understanding algorithms, designing new algorithms, and converting them to software is what we do on a daily basis. This requires a combination of a deep understanding of measurement technology, mathematics, and software development skills,” says Panu.
Optimal use of timber increases yield – and reduces the environmental footprint.
When raw wood is utilised optimally by means of smart technology, natural resources can be conserved. Efficiency does not therefore mean only improved yields, but also a smaller environmental footprint throughout the supply chain.
Software architecture is close to Panu’s heart.
“In addition to machine vision, designing software structures is my passion – and I seek to ensure that Lisker systems are always up to date and never stuck on the same old solutions,” he explains.
Sawmill industry systems are investments with life cycles of up to several decades. This is why Lisker develops software architecture that can still be developed many years down the road.
A poorly designed architecture can easily hinder modification and slow down the development of the software. The end result is worse productivity – and in the worst case scenario a system that cannot be maintained anymore.
“The older the software, the more important the architecture. When the foundations are solid, we can scale, update, and integrate new technologies – such as AI tools – without having to rebuild the entire system.”
An example of this is wanting to connect a new machine vision module or AI-based analysis component to the system, but the software structure cannot support it. If the architecture is inflexible, the development can become very costly – or even completely impossible.
This outlook has been utilised for example in the development of Profi-TC, where the system was built from the ground up with the idea that new recognition means and control logics could be developed in stages.
The rapid development of artificial intelligence or other technologies requires adaptable software and architecture. No-one can predict with confidence what future technology will look like. That is why it is important to make sure that the system structure makes it possible to connect new components, allowing the best innovations to be adopted at just the right time.
Panu says that Lisker product development efforts are always a response to a real customer need and not outcomes of new and interesting technologies. Many systems, such as Profi-TC, were built for a purpose, not just to fill up a product portfolio.
The development of Profi-TC began with the need of sawmill customers to have a more precise rotation solution without moving parts and sensitive components. This resulted in an idea: would it be possible to monitor the rotation of a log just by observing its surface texture, without using markings or lasers? The solution was a combination of machine vision and texture analysis, which made it possible to put together an entirely novel system.
Development does not begin with a technology trend, but with a question: what does the sawmill actually need?
Products are developed in on-going interaction with customers. Just recently, a customer corrected Lisker’s promise – in a positive spirit: A rotation standard deviation of five degrees was promised for Profi-TC, but, according to the customer, the actual deviation was more like three degrees.
“This type of feedback is not exceptional, but an outcome of a partnership in which user experiences, data collection, and continuous iteration are the guidelines for development,” Panu concludes.
Lisker Oy
Vähäkuja 3 / 17200 VÄÄKSY / FINLAND
+358 3 872 3400 / info@lisker.fi / www.lisker.fi
