Reconfigurable Voxels: A New Representation for LiDAR-Based Point Clouds

Tai Wang, Xinge Zhu, Dahua Lin
Proceedings of the 2020 Conference on Robot Learning, PMLR 155:286-295, 2021.

Abstract

LiDAR is an important method for autonomous driving systems to sense the environment. The point clouds obtained by LiDAR typically exhibit sparse and irregular distribution, thus posing great challenges to the detection of 3D objects, especially those that are small and distant. To tackle this difficulty, we propose Reconfigurable Voxels, a new approach to constructing representations from 3D point clouds. Specifically, we devise a biased random walk scheme, which adaptively covers each neighborhood with a fixed number of voxels based on the local spatial distribution and produces a representation by integrating the points in the chosen neighbors. We found empirically that this approach effectively improves the stability of voxel features, especially for sparse regions. Experimental results on multiple benchmarks, including nuScenes, Lyft, and KITTI, show that this new representation can remarkably improve the detection performance for small and distant objects, without incurring noticeable overhead cost

Cite this Paper

BibTeX
@InProceedings{pmlr-v155-wang21b, title = {Reconfigurable Voxels: A New Representation for LiDAR-Based Point Clouds}, author = {Wang, Tai and Zhu, Xinge and Lin, Dahua}, booktitle = {Proceedings of the 2020 Conference on Robot Learning}, pages = {286--295}, year = {2021}, editor = {Kober, Jens and Ramos, Fabio and Tomlin, Claire}, volume = {155}, series = {Proceedings of Machine Learning Research}, month = {16--18 Nov}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v155/wang21b/wang21b.pdf}, url = {https://proceedings.mlr.press/v155/wang21b.html}, abstract = {LiDAR is an important method for autonomous driving systems to sense the environment. The point clouds obtained by LiDAR typically exhibit sparse and irregular distribution, thus posing great challenges to the detection of 3D objects, especially those that are small and distant. To tackle this difficulty, we propose Reconfigurable Voxels, a new approach to constructing representations from 3D point clouds. Specifically, we devise a biased random walk scheme, which adaptively covers each neighborhood with a fixed number of voxels based on the local spatial distribution and produces a representation by integrating the points in the chosen neighbors. We found empirically that this approach effectively improves the stability of voxel features, especially for sparse regions. Experimental results on multiple benchmarks, including nuScenes, Lyft, and KITTI, show that this new representation can remarkably improve the detection performance for small and distant objects, without incurring noticeable overhead cost} }
Endnote
%0 Conference Paper %T Reconfigurable Voxels: A New Representation for LiDAR-Based Point Clouds %A Tai Wang %A Xinge Zhu %A Dahua Lin %B Proceedings of the 2020 Conference on Robot Learning %C Proceedings of Machine Learning Research %D 2021 %E Jens Kober %E Fabio Ramos %E Claire Tomlin %F pmlr-v155-wang21b %I PMLR %P 286--295 %U https://proceedings.mlr.press/v155/wang21b.html %V 155 %X LiDAR is an important method for autonomous driving systems to sense the environment. The point clouds obtained by LiDAR typically exhibit sparse and irregular distribution, thus posing great challenges to the detection of 3D objects, especially those that are small and distant. To tackle this difficulty, we propose Reconfigurable Voxels, a new approach to constructing representations from 3D point clouds. Specifically, we devise a biased random walk scheme, which adaptively covers each neighborhood with a fixed number of voxels based on the local spatial distribution and produces a representation by integrating the points in the chosen neighbors. We found empirically that this approach effectively improves the stability of voxel features, especially for sparse regions. Experimental results on multiple benchmarks, including nuScenes, Lyft, and KITTI, show that this new representation can remarkably improve the detection performance for small and distant objects, without incurring noticeable overhead cost
APA
Wang, T., Zhu, X. & Lin, D.. (2021). Reconfigurable Voxels: A New Representation for LiDAR-Based Point Clouds. Proceedings of the 2020 Conference on Robot Learning, in Proceedings of Machine Learning Research 155:286-295 Available from https://proceedings.mlr.press/v155/wang21b.html.

Related Material