Computer Vision and Pattern Recognition Laboratory (CVPRL)

Academy of Finland project DeepTimber started

The DeepTimber project uses the next generation digitalisation to study the optimisation of sawing

The DeepTimber project "Sawing Optimization via Deep Learning and Multi-instrument Imaging" funded by Academy of Finland started as a coperation between LUT CVPRL and the LUT mathematicians.

Sustainability and profitability of sawmills depend on their ability to streamline sawing processes, optimize the quality of end product, and minimize waste products due to suboptimal sawing processes. Increased efficiency can be for instance achieved by digitalization; for which many technical solutions already exist, such as automatic sorting and grading. As part of this project we aim at the next stage of digitalization, by creating an internet of things for full log tracking from initial measurements to the final product. Which will be supported by the main focus of this project on optimizing the sawing processes. Novel imaging techniques from multi instrument measurements will enable the creating of a digital twin for each log, and a virtual sawing can be performed to determine optimal sawing patterns and angles controlling the locations of knots and other defects. This will make the sawmill process more resource-efficient and will promote sustainability of the forestry sector.

The two-year project was started in April 1, 2020. The principal investigator at LUT is Associate Professor Lassi Roininen from LUT Uncertainty Quantification and Inverse Problems Laboratory and the project manager is Adjunct Professor, Dr. Tuomas Eerola from LUT CVPRL. The project partners are University of Oulu from Finland and Royal Institute of Technology (KTH) from Sweden.