Edge and Corner Defects Detection on Crude Ceramic Tiles

Ceramic tile production consists of several stages. Production starts with a specific material mixture and finishes with the visual inspection and classification by human operators. In order to increase the production quality and to lower the cost of the production in the ceramic tile industry visual inspection and classification needs to be automated. Automation of the visual inspection and classification process can be achieved in two stages: on a crude ceramic tile (before the baking process) and on a baked ceramic tile. In this paper we propose a method for visual inspection of a prebaked ceramic tile. The proposed method detects and localizes edge and corner defects. Method is implemented with Matlab/C++ environment and tested on a prototype conveyor line with an area scan camera. The presented results show that the edge and corner defect detection rate are very high and the method can be used for real-time application.