A New Method for Defect Detection on Images Captured from Wood Veneer: Optimization of Energy Function on Natural Texture Eliminated Image

In this paper a new method is introduced for defect detection in veneer images to investigate wood quality. In this method firstly, separating of defects from natural background of veneer is modeled a hypothesis testing problem. In the next step, the natural texture of veneer is eliminated by using morphology concept. Finally the correct boundaries of defects are extracted by optimizing energy function on above homogen area. Performance of the proposed algorithm is evaluated on real captured images containing several kinds of surface defects. The results demonstrate that the proposed method detects the defects approximately 18% better than some present approaches. Furthermore, it may be shown that better detection of defects in the proposed algorithm not only does not lead to extracting more false defects, but also it decreases rate of false detections approximately 8.2% compared to the existing algorithms. Consequently, it may be concluded that the proposed method may be used as a suitable alternative for detecting defects in veneer surfaces.