Abstarct: Nowadays one of the requirements to improve the quality and quantity of the industrial products are the use of machines that are automatically and lower its reliance on human factors. Despite advances in the ceramic tile’s automation industry, troubleshooting the product is still manually and done by human factors. Most available works in this area has focused on defects whose existence cannot definitely determine the fate of the product at the point of manufacture. Fracture Defects are important because defective products of this defect are likely to be recovered. And timely diagnosis of this defect, the product is sent to the next stage can be prevented. That it reduces production costs and adverse environmental effects. The mentioned cases provided the possibility of this diagnosis separately. Hence, in this thesis a simple and efficient method baxsed on image processing and machine vision techniques is presented for automatic detection of ceramic tiles fracture defects in borders and bodies. The proposed method uses Healthy boundaries retrieval approach. Using the approach proposed in this thesis, Discovery, enumeration, screening and grading fracture defects will be done in real time and with high accuracy. In addition, it is possible that the proposed method can be used in two different phases. First one is before entering the products to the furnace. In this stage, the products are weak. In this case, the proposed method can avoid of sending the defective product to the next stage. This action reduces wasteful costs. Another use of proposed method is in the calibration phase. Hence grading of products has been added to the other capabilities of the proposed method. Real-time implementation of the method is the most prominent feature of the proposed method. It is evident this issue in the industry is very important.