The 3D reconstruction of power lines is a critical step in power inspection. To improve the accuracy of power line extraction and reconstruction, this paper proposes an automated method for the extraction and reconstruction of power lines. First, the LiDAR point cloud data is processed using an elevation threshold to identify tower points and determine the midpoints of the towers. A connection line is then drawn along the center points of two towers. The segmentation line that divides the power line should be perpendicular to the connection line of the tower center points, so that each segment of the transmission line point cloud file is divided into multiple point clouds, each containing a single-span power line. Then, elevation thresholds and the DBSCAN clustering algorithm are used to accurately remove towers and other points, isolating the individual power line. Next, PCA is applied to determine the primary direction of the power line, reducing the data dimensions while preserving key features. Finally, the least squares method is used to fit the conductor line equation, obtaining the optimal parameters for precise modeling. Experimental results show that the average accuracy of power line extraction using this method is 99.06%, with an average reconstruction residual of 0.15 m. In summary, this method can accurately reconstruct the 3D shape of power lines.