Research on Lightweight Steel Rebar End Surface Detection Algorithm Based on Improved YOLOv8

This paper proposes a steel bar detection algorithm model based on the YOLOv8 object detection algorithm. It improves the network structure by using EMA (Exponential Moving Average) attention to replace the original C2f module, and introduces the Dyhead detection head to achieve the unification of object detection head and attention, thereby enhancing the feature extraction capability of the base model. Experiments show that the improved method improves the mAP by 1.01 % compared with the original algorithm model, which can effectively improve the accuracy and speed of steel bar quantity detection, and is more in line with the needs of industrial automation.