Real-Time Steel Surface Defect Detection Using YOLOv8 and OpenCV: Enhancing Accuracy and Efficiency in Industrial Applications

With its excellent surface quality and good mechanical properties, steel has become an indispensable and important material in the machinery manufacturing industry. However, in the process of steel production, due to the complex industrial background, there are always many surface defects. Defects on the surface of steel will not only affect its appearance quality, but also reduce its durability, and sometimes even cause machine damage and some accident casualties. The traditional steel surface defect detection method is mainly based on manual inspection, which is time-consuming, labor-intensive, and inefficient. Therefore, the research on steel surface defect detection is a very meaningful topic. In this study, a real-time steel surface defect detection method combining YOLOv8 and OpenCV is proposed to improve defect detection accuracy in industrial environments. Steel quality is critical in industrial applications where surface defects such as cracks, scratches, and oxidation can seriously affect performance and safety. Traditional inspection methods, such as manual inspection and magnetic leakage inspection, often have shortcomings in accuracy, adaptability, and real-time performance, especially in high-speed production lines. In this study, the deep learning ability of YOLOv8 is used to strengthen the detection of small objects, and the image preprocessing function of OpenCV (including edge detection and noise suppression) is combined to enhance the presentation of defect features, so as to improve the stability of the model in complex environments. The experimental results show that the combined system significantly improves the accuracy of defect detection, environmental adaptability, and processing speed. This method not only improves the efficiency of steel surface defect detection but also lays the foundation for wider application in the field of intelligent manufacturing. Future research will further explore adaptive filtering and lightweight feature enhancement to optimize the performance of the system in diverse industrial scenarios.