“In response to the problem of low recognition accuracy of lightweight object detection algorithms in steel surface defect detection tasks, researchers have proposed a multi-scale enhanced feature fusion algorithm for steel surface defect object detection. This algorithm innovatively adopts an adaptive weighted fusion module, which achieves weighted fusion of deep semantics and shallow details, effectively improving the feature representation ability. At the same time, the algorithm also introduces a spatial feature enhancement module, which enhances the fusion features from three independent directions, improves the stability of the network structure, and mines more key information. The experimental results show that the detection accuracy of this algorithm reaches 80.47%, which is 6.81% higher than the original algorithm. In addition, the parameter and computational complexity of the algorithm are relatively small, and it can quickly and accurately detect defect information on the surface of steel, which has high application value. This research achievement provides a new solution for detecting surface defects in steel and opens up new directions for research in related fields.”