FinlandManufacturing2023 – ongoingKoskisen Oyj
Neural-network defect detection on a veneer composing line
Koskisen added neural-network AI to a 30-year-old veneer line to tell birch bark from sound knots
Problem
Birch plywood is hard to produce. The line's machine vision could not reliably tell birch bark from sound knots, so defects were sorted less accurately than the mill wanted.
Approach
Koskisen installed a new veneer composing analyser in 2023 and used Raute's neural-network AI on it. The pilot ran on a line that had been in use for about 30 years and was one of the last older workstations at the plywood mill. The AI detects defects in real time.
Results
No published figures
Takeaway
Koskisen and Raute reported more accurate detection than on sister lines with older technology, and higher operator satisfaction. The article gave no figures for accuracy or output.
Categories
- Industry
- Manufacturing
- Business function
- Production and quality
- Innovation type
- Process innovation
- AI technology
- Computer vision
- Adoption
- Vendor solution
- Evidence
- Company's own claim
Sources
Source summaries
A trade article on Raute's neural-network AI added to the veneer composing analyser at Koskisen's plywood mill. It says bark and sound knots are now identified more accurately and detection errors are lower than on sister lines. It gives no numbers and no exact pilot date.
Full details are in the original sources.
