Corner defect detection based on dot product in ceramic tile images

One of the important problems in ceramic tile industry is tiles' quality classification with automatic systems by applying machine instead of human. Tiles' quality can be divided into color analysis, dimension verification, and surface defect detection. It's very difficult for human to control all of them, because of harsh industrial environment with noise, extreme temperature and humidity. In this paper, we present a method for visual inspection of ceramic tile corners. We use a method based on image processing techniques and dot product vectors if angle was more than 92 degree or less than 89 degree. Our ceramic is a defective tile. Our algorithm is evaluated on a set of images which has been taken of a Flaw master system in a tile factory and has 12.5% error in both normal and defective tile. The obtained results show efficiency our approach in corner defect detection.