Abstract. It has been a challenge for electric power management to automatically extract power lines from LiDAR point clouds. However, environmental and technical issues have made management more challenging in complicated areas where power lines are in close proximity to buildings and/or trees. In this study, the structure and types of the data captured by a LiDAR sensor in regions containing line corridors were analysed. The crucial stage is appropriately identifying from the data the essential parts of a power line corridor route. The point cloud dataset used in the study belongs to the Borssele region in Zeeland, the Netherlands. By manually labelling the dataset, three classes were identified: wire, pylon, and others. For the classification of point clouds, the Random Forest method was utilised. To assess the obstacles posed by the class wire, 5 m, 10 m, and 15 m 3D buffer zones are created. The visual presentation of obstacles within the buffer zone is achieved by assigning them a separate class code and indicating that they are inside and partially within. Based on the results, the correctness values of the classes of wire and others are considered to be satisfactory. However, the class pylon contains points with incorrect labels after the classification. As a result, the accuracy of the pylon class is much lower than the accuracy of the other two classes.

Abstract. The study deals with the automatic supervised classification of urban objects from point clouds collected by the vehicle-based Mobile Laser Scanning (MLS) system. A benchmark dataset representing the Technical University of Munich (TUM) City Campus was used. The main contribution of this article is evaluating the performance difference between kNN, cylindrical and spherical local neighborhood relations in point-based classification of an MLS system using local geometric and shape-based features. The Random Forest (RF) classifier was performed for 8 manually marked classes in the benchmark set: artificial terrain, natural terrain, high vegetation, low vegetation, building, hardscape, artifact and vehicle. We reveal that the cylindrical neighborhood with 13 attributes provides an improvement of 5.2% compared to the spherical neighborhood, while the kNN gave almost the same result as the cylindrical neighborhood (0.8% improvement) in the shortest time. Finally, a new feature set was created by combining the most important features obtained from different neighborhood types. As a result, we achieved 96.9% overall accuracy by using 19 significant features obtained from all neighborhood types for the TUM-MLS1 point cloud.