Fehlstellenerkennung an Holzoberflächen mit digitaler Bildverarbeitung

Different problems of the automatic visual inspection of wood surfaces using digital image processing were examined. After a detailed review of the state of the art the experimental set-up of a trial system is described. This system allows the acquisition of gray value images of wood surfaces with a CCD line scan camera which are stored in a computer and processed with several image processing algorithms. Solutions have been developed for the automatic detection of different defects regarding growth, production and processing of sawn timber, parquet wood, veneers and wood-based panels. Most of the cases showed that simple segmentations with fixed gray value thresholds are not sufficient for this purpose, and that correspondingly dynamic gray value thresholds, adjusted digital filter operations or methods based on gray value statistics have to be used. A continuation of the work is recommended regarding the realization of the processes examined by simulation in a real time system.