Research on Intelligent Recognition and Analysis of Power Grid Infrastructure Construction Based on YOLO

Power grid construction sites pose challenges like clutter, occlusion, and varying scales, making manual monitoring insufficient. Employing image recognition technology is crucial to track personnel, safety compliance, and vehicle entry in real time. Despite limited data availability, our study achieves 90% accuracy using transfer learning with a small dataset. This adaptable model not only identifies elements in power grid sites but also accommodates diverse task demands.