AI for defect recognition in wooden surfaces
| Reference number | |
| Coordinator | WOODEYE AB - INNOVATIV VISION AB |
| Funding from Vinnova | SEK 500 000 |
| Project duration | October 2019 - April 2020 |
| Status | Completed |
| Venture | AI - Competence, ability and application |
| Call | Start your AI journey! |
Purpose and goal
The project has become the starting point for WoodEye AB stepping into the AI world. The company has learned how to run an AI project in everything from the feasibility study to how to collect and label image data and then training different networks. The results have been good and will further strengthen WoodEye´s scanners to perform in absolute world class. Already, plans are being made for the first customer deliveries so that the timber industry can benefit from the improved detection as quickly as possible.
Expected results and effects
* Developed AI networks have helped improve detection * WoodEye has significantly increased AI expertise within the company * Developed AI networks are already integrated into WoodEye´s software * First delivery with AI -based detection is planned for summer 2020 * No trimming of the AI network is needed. It detects the learned defects regardless of wood type
Planned approach and implementation
** Denna text är maskinöversatt ** We have collected high quality images from 10,000 boards scanned with various WoodEye scanners worldwide. These images have been labeled with the help of a program called WoodAI LabelTool developed within the project. This data has been used to train multiple networks to classify dark and light twigs of various sizes in the images.