Research on the algorithm for calculating tower span and line sag based on visual localization

Abstract Distribution networks are crucial for power systems, and accurate measurement of tower span and conductor sag is vital for safe operation. This paper proposes an automated method using drone imaging to measure these parameters. A low-altitude photogrammetry platform with YOLOv8 detects towers and conductors, while the ORB-SLAM2 framework enables real-time drone positioning and sparse reconstruction. A dense matching model generates high-fidelity 3D point clouds. Customized span and sag calculation models consider line asymmetry and wind deviation. The effectiveness of the method is validated through experiments, enhancing measurement efficiency and accuracy.