Quality inspection of veneer using soft-computing methods

This paper describes the use of a complex modular image processing system for veneer classification. An introduction into problems that arise when producing and processing veneer is given, namely flaws in the wood and faults in the final product, in this case spring boards for slatted frames. A system has been developed that detects and classifies the faults. A line-scan camera is used for capturing the veneer images of which the features are extracted with statistical methods. The features are classified with either fuzzy clustering methods or neural networks. Additionally genetic algorithms are used for optimization purposes