Automatic classification of surface defects on red oak boards

Most of the conventional manufacturing processes of a furniture plant are labor oriented. In order to improve the productivity and quality of products, automation of the different stages in the production is needed. Automatic classification of surface defects on wood boards can serve as a step toward achieving a completely automated wood processing plant. It can be used for grading boards, which is usually performed by a human. The information obtained by automatic classification can also provide a basis to determine an optimal strategy for cutting boards. In this paper, image modeling and pattern recognition techniques were applied for automatic classification of surface defects on wood boards. Planed boards of red oak were used in the study. Nine classes, eight types of surface defects and clear wood, were considered in the classification. The estimated parameters based on an image modeling method with the mean of the gray levels of the images were used as features. A hierarchical tree classifier was constructed based on selected features. A classification accuracy of 97.2 percent was obtained in classifying nine classes.

You must be logged in to download any documents. Please login (login accounts are free) or learn how to Become a Member