Faculty of Information Technology, BUT

Publication Details

CNN for Very Fast Ground Segmentation in Velodyne LiDAR Data

VEĽAS Martin, ŠPANĚL Michal, HRADIŠ Michal and HEROUT Adam. CNN for Very Fast Ground Segmentation in Velodyne LiDAR Data. In: Proceedings of IEEE International Conference on Autonomous Robot Systems and Competitions. Torres Vedras: Institute of Electrical and Electronics Engineers, 2018, pp. 97-103. ISBN 978-1-5386-5221-3. Available from: http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=8374163&isnumber=8374143
Czech title
Využití konvolučních sítí pro velmi rychlou segmentaci země v datech LiDARu Velodyne
Type
conference paper
Language
english
Authors
URL
Keywords
convolutional neural networks, ground segmentation, Velodyne, LiDAR
Abstract
This paper presents a novel method for ground segmentation in Velodyne point clouds. We propose an encoding of sparse 3D data from the Velodyne sensor suitable for training a convolutional neural network (CNN). This general purpose framework is used for segmentation of the sparse point cloud into ground and non-ground points. The LiDAR data are represented as a multi-channel 2D signal where the horizontal axis corresponds to the rotation angle and the vertical axis indexes channels (i.e. laser beams). We proposed multiple topologies of relatively shallow CNNs (i.e. 3-5 convolutional layers) and evaluated them using a manually annotated dataset we prepared. The results show significant improvement of performance over the state-of-the-art method by Zhang et al. in terms of speed and also minor improvements in terms of accuracy.
Published
2018
Pages
97-103
Proceedings
Proceedings of IEEE International Conference on Autonomous Robot Systems and Competitions
Conference
IEEE International Conference on Autonomous Robot Systems and Competitions, Torres Vedras, PT
ISBN
978-1-5386-5221-3
Publisher
Institute of Electrical and Electronics Engineers
Place
Torres Vedras, PT
DOI
BibTeX
@INPROCEEDINGS{FITPUB11346,
   author = "Martin Ve\'{l}as and Michal \v{S}pan\v{e}l and Michal Hradi\v{s} and Adam Herout",
   title = "CNN for Very Fast Ground Segmentation in Velodyne LiDAR Data",
   pages = "97--103",
   booktitle = "Proceedings of IEEE International Conference on Autonomous Robot Systems and Competitions",
   year = 2018,
   location = "Torres Vedras, PT",
   publisher = "Institute of Electrical and Electronics Engineers",
   ISBN = "978-1-5386-5221-3",
   doi = "10.1109/ICARSC.2018.8374167",
   language = "english",
   url = "https://www.fit.vut.cz/research/publication/11346"
}
Back to top