Identification and selection strategy of relay core components based on YOLOv8n

To address the issues of low welding efficiency in the core components of new energy relays and the difficulty in component identification and screening, this paper proposes a target detection method based on YOLOv8 for real-time identification and screening of core components. Firstly, the model adopts the improved YOLOv8n convolutional Neural Network (CNN) architecture, mainly improving the C2f module of the YOLOv8n model. Then, the data set of relay core components is constructed and divided. Finally, the target detection model of new energy relay core components is trained and applied to material identification in the welding production process to achieve accurate identification and classification of related parts. The experimental results show that the accuracy of the method for the core components of copper rod and copper sheet is 95% and 96%, respectively, which provides technical support for the intelligent production of new energy relays.