Aiming at the problems of low efficiency, strong subjectivity in traditional bearing surface defect detection and insufficient dimensional measurement accuracy, this paper proposes an integrated detection scheme SimAM-YOLO that combines the improved YOLOv5 algorithm with size measurement technology. Based on YOLOv5, the scheme replaces the original C3 module with the C2F network structure and embeds the SimAM attention mechanism to enhance the model's ability to extract defect features. Combined with OpenCV, it realizes the real-time measurement of the key dimension of bearing radius and constructs a visual system for bearing size measurement. Experimental results show that the improved mode... More >