With the intelligent upgrading of power grids,accurate equipment identification has become a key factor in enhancing management efficiency. To address the challenge of transmission line equipment identification,this paper proposes a digital twin-oriented identification method for transmission line equipment. Under the framework of a digital twin-based transmission line system, the YOLOv7 algorithm is improved by integrating a cross-connection structured feature pyramid network
an attention mechanism
and a minimum point-distance intersection-over-union loss function to enhance the model’s recognition accuracy. Simulation analysis demonstrates that,compared to other object detection methods,the proposed approach exhibits significant advantages in recognition precision and adaptability to complex scenarios.