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 approach 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 represents channels - laser beams. Multiple topologies of relatively shallow CNNs (i.e. 3-5 convolutional layers) are trained and evaluated, 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. |
Year | DOI | Venue |
---|---|---|
2018 | 10.1109/ICARSC.2018.8374167 | 2018 IEEE International Conference on Autonomous Robot Systems and Competitions (ICARSC) |
Keywords | DocType | Volume |
CNN,fast ground segmentation,velodyne LiDAR data,Velodyne point clouds,sparse 3D data,convolutional neural network,sparse point cloud,multichannel 2D signal,horizontal axis corresponds,rotation angle,vertical axis,convolutional layers,shallow CNNs,Velodyne sensor | Conference | abs/1709.02128 |
ISSN | ISBN | Citations |
2573-9360 | 978-1-5386-5222-0 | 1 |
PageRank | References | Authors |
0.35 | 1 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Martin Velas | 1 | 7 | 1.20 |
Michal Spanel | 2 | 41 | 7.85 |
Michal Hradis | 3 | 132 | 14.19 |
Adam Herout | 4 | 248 | 35.39 |