Advanced search
1 file | 338.99 KB

Semantically aware multilateral filter for depth upsampling in automotive LiDAR point clouds

Martin Dimitrievski (UGent) , Peter Veelaert (UGent) and Wilfried Philips (UGent)
Author
Organization
Abstract
We present a novel technique for fast and accurate reconstruction of depth images from 3D point clouds acquired in urban and rural driving environments. Our approach focuses entirely on the sparse distance and reflectance measurements generated by a LiDAR sensor. The main contribution of this paper is a combined segmentation and upsampling technique that preserves the important semantical structure of the scene. Data from the point cloud is segmented and projected onto a virtual camera image where a series of image processing steps are applied in order to reconstruct a fully sampled depth image. We achieve this by means of a multilateral filter that is guided into regions of distinct objects in the segmented point cloud. Thus, the gains of the proposed approach are two-fold: measurement noise in the original data is suppressed and missing depth values are reconstructed to arbitrary resolution. Objective evaluation in an automotive application shows state-of-the-art accuracy of our reconstructed depth images. Finally, we show the qualitative value of our images by training and evaluating a RGBD pedestrian detection system. By reinforcing the RGB pixels with our reconstructed depth values in the learning stage, a significant increase in detection rates can be realized while the model complexity remains comparable to the baseline.
Keywords
lidar, depth, upsampling, multilateral, filter, segmentation, autonomous, driverless, pedestrian

Downloads

  • IV2017 0095 FI.pdf
    • full text
    • |
    • open access
    • |
    • PDF
    • |
    • 338.99 KB

Citation

Please use this url to cite or link to this publication:

Chicago
Dimitrievski, Martin, Peter Veelaert, and Wilfried Philips. 2017. “Semantically Aware Multilateral Filter for Depth Upsampling in Automotive LiDAR Point Clouds.” In , 1058–1063.
APA
Dimitrievski, M., Veelaert, P., & Philips, W. (2017). Semantically aware multilateral filter for depth upsampling in automotive LiDAR point clouds (pp. 1058–1063). Presented at the 2017 IEEE Intelligent Vehicles Symposium.
Vancouver
1.
Dimitrievski M, Veelaert P, Philips W. Semantically aware multilateral filter for depth upsampling in automotive LiDAR point clouds. 2017. p. 1058–63.
MLA
Dimitrievski, Martin, Peter Veelaert, and Wilfried Philips. “Semantically Aware Multilateral Filter for Depth Upsampling in Automotive LiDAR Point Clouds.” 2017. 1058–1063. Print.
@inproceedings{8516730,
  abstract     = {We present a novel technique for fast and
accurate reconstruction of depth images from 3D point clouds
acquired in urban and rural driving environments. Our
approach focuses entirely on the sparse distance and
reflectance measurements generated by a LiDAR sensor. The
main contribution of this paper is a combined segmentation
and upsampling technique that preserves the important
semantical structure of the scene. Data from the point cloud is
segmented and projected onto a virtual camera image where a
series of image processing steps are applied in order to
reconstruct a fully sampled depth image. We achieve this by
means of a multilateral filter that is guided into regions of
distinct objects in the segmented point cloud. Thus, the gains of
the proposed approach are two-fold: measurement noise in the
original data is suppressed and missing depth values are
reconstructed to arbitrary resolution. Objective evaluation in
an automotive application shows state-of-the-art accuracy of
our reconstructed depth images. Finally, we show the
qualitative value of our images by training and evaluating a
RGBD pedestrian detection system. By reinforcing the RGB
pixels with our reconstructed depth values in the learning
stage, a significant increase in detection rates can be realized
while the model complexity remains comparable to the
baseline.},
  author       = {Dimitrievski, Martin and Veelaert, Peter and Philips, Wilfried},
  keyword      = {lidar,depth,upsampling,multilateral,filter,segmentation,autonomous,driverless,pedestrian},
  language     = {eng},
  location     = {Redondo Beach, CA, USA},
  pages        = {1058--1063},
  title        = {Semantically aware multilateral filter for depth upsampling in automotive LiDAR point clouds},
  url          = {http://dx.doi.org/10.1109/IVS.2017.7995854},
  year         = {2017},
}

Altmetric
View in Altmetric