Camera Based Lumber Strength Classification System

In this paper, a new camera based lumber strength classification solution using solely knot features is presented. Two alternative classifiers, k-NN and SVM, are applied to classifing pinewood boards based on their breaking strength. Features for classification are formed using knot properties, which are extracted from four sides of the board using machine vision algorithms. Extracted properties include size, x,y,z-coordinates and the type of knot. These properties are used as such as features. They are also used to form different combination features like length or volume of the knot. In experiments, ground truth breaking strength of the boards was determined using a three point bending test. Our evaluation shows that when knots are present it is possible to classify pinewood boards with over 70% accuracy using a combination of knot based features.