AI for Spotting and Predicting Timber Defects
A lab study using NIR camera analysis trained AI to detect wood species and accurately identify damaged wood, forming the foundation for an industrial prototype.
Did you know that Sweden is the world’s second-biggest exporter of pulp, paper, and sawn wood products? Timber is a significant industry.
The Study
This project is a lab study where log cross-sections from spruce and pine with varying degrees and types of defects were scanned with a NIR camera in a lab setup.
Objectives
The aim was to train the AI to:
- Detect wood species
- Accurately identify and predict the proportion of damaged wood
Future Development
This forms the foundation for developing an industrial prototype that could be deployed in sawmills and timber processing facilities.
Impact
Automated defect detection can improve efficiency in the timber industry by enabling faster and more consistent quality assessment.
