A Quality Control System Prototype for Detecting Knot Defects in the Wooden Panel Manufacturing

Abstract

Product quality has become a necessary goal for all manufacturers in today's competitive market. Product defects, not detected, cause financial damages and reputation loss for the manufacturer. These defects can be due to quality of the inputs or misuse of the good quality inputs during the manufacturing process. This is also the case for wooden panel manufacturing where elements are the basic input. It is possible to reduce the loss of the manufacturer by using a method that minimizes the human error in the inspection of the elements. In this study, we, first, identified the quality control problems of the wooden panel manufacturers and basic steps in the automated element quality control. We then developed a prototype for the detection of knots, the most common defects in wooden panels. This prototype, with 80.0% true positive (with knot defect) and 82.0% true negative (without knot defect) rates, performs close to accuracy rates of a quality control inspector. The element image library created during the development of the system made publicly available for use in similar studies. This prototype is expected to be developed to detect other wood defects and to be applied in the wooden panel manufacturing.