Steel plates are now found in nearly every aspect of daily life. They play an important role in the production of industrial, automotive, and technological products. Because degradation in these materials can affect every stage of industrial production, it is critical to detect these deteriorations as soon as possible. Steel surface deterioration is a symptom of these materials' internal and superficial failures. Image fault detection has grown in popularity in recent years. In this field, image-based non-contact fault detection methods are preferred because they are quick, dependable, and do not cause material damage. Using a Convolutional Neural Network, this study proposes a method for detecting deterioration on the surfaces of steel materials. The architect proposed in this study was found to be 95.21 percent successful in recognizing the defect classes.