Researchers are exploring various methods for detecting metal surface defects, including the use of YOLO algorithms and other technologies. Different approaches are being proposed, such as anchor-free algorithms, dynamic fine-grained multi-branch encoders, and multi-scale feature fusion. The goal appears to be the development of efficient, lightweight, and high-precision defect detection systems for metals like steel.