Automated power line recognition and 3D reconstruction for intelligent grid monitoring

Monitoring and management of power line corridors are essential for ensuring the safe and reliable operation of power transmission systems. Traditional manual inspection methods are not only inefficient but also pose significant safety risks, while certain existing automated approaches suffer from limited effectiveness in complex terrains or in the presence of discontinuities in point cloud data, resulting in insufficient accuracy in power line extraction and frequent reconstruction failures. To address these challenges, this study proposes a novel power line reconstruction method termed Weighted Multi-feature Multi-plane Projection Geometric Fusion (WM-MPGF). The proposed method comprises two sequential stages: Weighted Multi-Feature SVM (WMF-SVM) and Multi-plane Projection and Geometric Joint Reconstruction (MPG-Recon). Specifically, WMF-SVM introduces a weighted multi-feature support vector machine framework that integrates elevation data derived from the Digital Elevation Model (DEM) with spatial features extracted via Principal Component Analysis (PCA), while optimizing feature weights through the entropy weight method to enhance the accuracy of power line identification. Subsequently, MPG-Recon performs geometric analysis to construct directional projection planes and applies the Hough transform to project power line points and determine their dominant orientations on the XOY and YOZ planes. The Davies-Bouldin index is employed to determine the optimal number of clusters, thereby enabling accurate estimation of the number of power lines. By integrating the K-means clustering algorithm, the method achieves effective separation of multiple power lines and ensures high-precision fitting of individual conductors. Experimental results indicate that the proposed approach achieves average fitting errors of 5.41 cm on the XOY plane and 5.68 cm in the vertical direction, successfully capturing the three-dimensional structural characteristics of power lines. The method constructs a robust 3D model and provides critical technical support for advanced applications in power line corridor monitoring and maintenance.