English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

The correct ego-motion estimation basically relies on the understanding ofcorrespondences between adjacent LiDAR scans. However, given the complexscenarios and the low-resolution LiDAR, finding reliable structures foridentifying correspondences can be challenging. In this paper, we delve intostructure reliability for accurate self-supervised ego-motion estimation andaim to alleviate the influence of unreliable structures in training, inferenceand mapping phases. We improve the self-supervised LiDAR odometry substantiallyfrom three aspects: 1) A two-stage odometry estimation network is developed,where we obtain the ego-motion by estimating a set of sub-regiontransformations and averaging them with a motion voting mechanism, to encouragethe network focusing on representative structures. 2) The inherent alignmenterrors, which cannot be eliminated via ego-motion optimization, aredown-weighted in losses based on the 3D point covariance estimations. 3) Thediscovered representative structures and learned point covariances areincorporated in the mapping module to improve the robustness of mapconstruction. Our two-frame odometry outperforms the previous state of the artsby 16%/12% in terms of translational/rotational errors on the KITTI dataset andperforms consistently well on the Apollo-Southbay datasets. We can even rivalthe fully supervised counterparts with our mapping module and more unlabeledtraining data.

Robust Self-Supervised LiDAR Odometry via Representative Structure Discovery and 3D Inherent Error Modeling