3D visualization of distribution network overhead lines using point cloud edge registration

A 3D visualization system is designed for distribution network overhead line operation scenarios based on a point cloud edge registration algorithm to address the common challenge of data latency, which hinders real-time updates in existing technologies and compromises the accurate reflection of the latest grid status. Overhead line data are acquired through oblique photogrammetry. The normal vector and curvature of the point cloud data are estimated using PCA analysis, while the coordinate transformation of the distribution lines is accomplished via a genetic algorithm. The fusion of laser point cloud data and oblique photogrammetric data is realized by solving for rotation and translation parameters. A feature point registration algorithm combining coarse and fine registration steps is adopted to achieve precise point cloud alignment. A 3D visualization model is constructed using a triangulated irregular network. Experimental results show that when the reduction ratio exceeds 60%, the positional deviation of the primitive points is negligible, with a runtime of 369 ms, a mean square registration error of 0.097 mm, and no holes or distortions present in the model.