Automatic detection of sound knots and loose knots on sugi using gray level co-occurrence matrix parameters

Abstract

Knowledge of the occurrence of sound and loose knots on the surface of sawn sugi (Cryptomeria japonica L.f.) is important for its grading and application. This study examined an optical system for detecting sound and loose knots in sugi instead of human being using the combining information of the color and texture features. The proposed system could be conceptually divided into two components: a CCD camera scanning system and a defect detecting algorithm developed by the authors. In the algorithm, the contrast parameter calculated from a gray-level co-occurrence matrix was used to locate the potential defects represented by sound knots and loose knots. The rule-based approach, which was built according to the color feature histograms, was used to identify sound knots and loose knots. A series of samples containing single or multiple sound and/or loose knots were selected at random to verify the efficiency and accuracy of the proposed system. There were 94 sound knots and 86 loose knots on the surfaces of these samples, and the accuracy of locating the positions of sound knots and loose knots was 94.7% and 97.6%, respectively. The accuracies of identifying knots as sound or loose were 96.6% and 98.8%, respectively. The overall detection accuracy of the system was 93.9%. The results indicate that the proposed vision system is an efficient means of detecting sound knots and loose knots.