An algorithm for automatic detection of pole-like street furniture objects from Mobile Laser Scanner point clouds |
| |
| Institution: | 1. Department of Mining Exploitation, University of Oviedo, 33004 Oviedo, Spain;2. Department of Natural Resources and Environmental Engineering, University of Vigo, 36310 Vigo, Spain;1. Institute of Photogrammetry and Remote Sensing, Karlsruhe Institute of Technology (KIT), Englerstr. 7, 76131 Karlsruhe, Germany;2. Université Paris-Est, IGN, SRIG, MATIS, 73 Avenue de Paris, 94160 Saint-Mandé, France;1. Fujian Key Laboratory of Sensing and Computing for Smart City, Xiamen University, Xiamen, FJ 361005, China;2. Faculty of Computer and Software Engineering, Huaiyin Institute of Technology, Huaian, JS 223003, China;3. Department of Geography & Environmental Management, University of Waterloo, Waterloo, ON N2L3G1, Canada;4. College of Geography & Remote Sensing, Nanjing University of Information Science and Technology, Nanjing, JS 210044, China;1. State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University, Wuhan 430079, China;2. Guangdong Electric Power Research Institute, Guangzhou 510080, China;1. State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University, Wuhan 430079, China;2. Engineering Research Center for Spatio-tempoal Data Smart Acquisition and Application, Ministry of Education of China, Wuhan University, Wuhan 430079, China;3. Wuhan Geomatics Institute, Wuhan 430022, China;1. School of Mining Engineering, Campus Lagoas s/n, University of Vigo, 36310 Vigo, Spain;2. Defense University Center at the Naval Academy, Plaza de España, 36920 Marín, Spain |
| |
| Abstract: | An algorithm for automatic extraction of pole-like street furniture objects using Mobile Laser Scanner data was developed and tested. The method consists in an initial simplification of the point cloud based on the regular voxelization of the space. The original point cloud is spatially discretized and a version of the point cloud whose amount of data represents 20–30% of the total is created. All the processes are carried out with the reduced version of the data, but the original point cloud is always accessible without any information loss, as each point is linked to its voxel. All the horizontal sections of the voxelized point cloud are analyzed and segmented separately. The two-dimensional fragments compatible with a section of a target pole are selected and grouped. Finally, the three-dimensional voxel representation of the detected pole-like objects is identified and the points from the original point cloud belonging to each pole-like object are extracted.The algorithm can be used with data from any Mobile Laser Scanning system, as it transforms the original point cloud and fits it into a regular grid, thus avoiding irregularities produced due to point density differences within the point cloud.The algorithm was tested in four test sites with different slopes and street shapes and features. All the target pole-like objects were detected, with the only exception of those severely occluded by large objects and some others which were either attached or too close to certain features. |
| |
| Keywords: | Urban Simplification Detection Laser scanning Algorithms Mobile |
| 本文献已被 ScienceDirect 等数据库收录！ |
|
