Industry standards for recognition of marginal wood defects

This study of 46 U.S. wood products companies that manufacture a variety of hardwood products found no consistently applied formal or standard definition of what constitutes a marginal defect. Analysis indicated homogeneous yet nonspecific rules are used. While there were wide variations in criteria, similar quantitative terms were used. Two distinct classes of defects emerged. One group, encompassing holes, knots, and checks, appears to be well defined, while the other, encompassing stain, mineral, color harmony, and incipient decay, is left to subjective evaluation. For example, hole identification appears to be based solely on diameter, knot identification on diameter and occurrence, and check identification on length and width. For the second group, defect elimination and cutting decisions appear to be based more on how obvious a defect may be in the end product rather than on explicitly defined sizes and colors. In some cases, the cost of the end product does not directly reflect the number or size of wood defects allowed, but rather the manufacturing steps involved and the relative mechanization in manufacturing the product. A computer vision system now under development can identify size, color, and pattern differences and locate the board edge, type of defect, and its location. However, the cost and functional speed of the system is directly related to the specific requirements of the user. It is essential that measures used by industry be translatable into parameters usable by the vision system. With such standardization, one vision system can be tailored to identify specific defects for various manufacturers based on the quality needs of their individual product lines. The objective of this effort was to quantify information on the maximum allowable defects by product and species. This is the first effort and points to the need for additional work before a complete vision system is built.