Surface Defect Isolation in Ceramic Tile Based on Texture Feature Analysis Using Radon Transform and FCM

Applying image processing technology and machine vision in industry have had significant development in recent decade. Tile and ceramic industry was not excluded form this matter. By using image processing techniques in production line of this industry, it is possible to detect surface defections such as edge defections, cracks and coloring defections. Surface defection is the most common type of defection in tile and ceramic industry. In this article a new and simple approach for detecting surface defections is introduced. The proposed algorithm has four parts. First, angle correction using Radon transform, then, background elimination using Fuzzy C-Means, after that ,analyzing feature texture , and finally, classification with neural network. The available algorithm in the article has good ability in distinguishing the healthy tile from defective one. Compared with results of other methods; the used algorithm in this article has better results on actual data base, in presence of camera noise and environmental lighting effects.