The application of improved Yolov5s in anomaly detection of circuit lines for small targets

The purpose of this study is to address the challenges and limitations of detecting small anomalies in circuit lines by modifying Yolov5s model. We optimized the model to improve its performance in detecting small abnormal targets. Taking the detection of exposed copper wires in meter wiring as an example, we conducted a series of experiments to validate the performance of the improved Yolov5s model. The experimental results showed that our improved model achieved a significant performance improvement on the self-made dataset for exposed copper wires in meter wiring. Compared to the baseline method, our improved model achieved a 4.5 percentage point increase in MAP [IoU=0.5:0.95], while maintaining accurate identification and precise localization of small abnormal targets. This provides a more reliable and efficient solution for circuit anomaly detection, with potential applications in the field.