Improving the Traceability of Wood-based Sheet Leftovers using Computer Vision

Abstract

Being able to provide traceability over raw material leftovers is fundamental to reducing waste, achieving leaner production processes, and promoting overall efficiency. Even if this makes sense to virtually all industries, in the low-volume, custom-production woodworking businesses, is of paramount importance if efficient integration of leftovers in the production process should take place. However, this is easier to say than done. This paper describes a methodology that is being devised to improve traceability for small and medium carpentry industries. This approach takes place within a broader R&D project and deals with the development of a storage rack that resorts to computer vision and machine learning to facilitate data gathering and digitization. Preliminary results regarding the computer vision methodology are provided.