Object Detection Algorithm for Power Transmission Lines

With extensive research in the field of deep learning, significant progress has been made in object detection. As a result, they have been applied in the power industry, such as identifying violations by workers, re identifying worker identities, detecting equipment defects and environmental hazards, and tracking personnel movement trajectories. With the rapid development of technology, an increasing number of edge embedded devices are being deployed in the power sector to efficiently perform data collection and processing tasks. However, these devices generally face challenges of limited computing resources and insufficient storage capacity. Due to the large parameters, computationally intensive nature, and high memory usage of most neural network models, their direct deployment on edge devices is severely constrained, hindering the widespread adoption of the technology in practical applications. This article proposes a drone detection algorithm based on YOLOv8. We (1) combine attention mechanisms to help the network focus on key regions and features, thereby improving performance with minimal additional computation; (2) Redesign the neck of the network based on data features to improve accuracy while maintaining the lightweight nature of the model. Our model can be effectively deployed on resource constrained embedded devices, providing robust detection results.