2021 IEEE International Conference on Robotics and Automation (ICRA) 2021
DOI: 10.1109/icra48506.2021.9560835
|View full text |Cite
|
Sign up to set email alerts
|

Voxelized GICP for Fast and Accurate 3D Point Cloud Registration

Abstract: This paper presents a point cloud downsampling algorithm for fast and accurate trajectory optimization based on global registration error minimization. The proposed algorithm selects a weighted subset of residuals of the input point cloud such that the subset yields exactly the same quadratic point cloud registration error function as that of the original point cloud at the evaluation point. This method accurately approximates the original registration error function with only a small subset of input points (2… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

0
82
0

Year Published

2021
2021
2024
2024

Publication Types

Select...
4
3
3

Relationship

1
9

Authors

Journals

citations
Cited by 199 publications
(82 citation statements)
references
References 34 publications
0
82
0
Order By: Relevance
“…This ultimately assumes that the environment is smooth and hence can be viewed as a plane locally. However, the computation load of generalized-ICP is usually large [33]. Other works based on Normal Distribution Transformation (NDT) [34]- [36] also register raw points, but NDT has lower stability compared to ICP and may diverge in some scenes [36].…”
Section: A Lidar(-inertial) Odometrymentioning
confidence: 99%
“…This ultimately assumes that the environment is smooth and hence can be viewed as a plane locally. However, the computation load of generalized-ICP is usually large [33]. Other works based on Normal Distribution Transformation (NDT) [34]- [36] also register raw points, but NDT has lower stability compared to ICP and may diverge in some scenes [36].…”
Section: A Lidar(-inertial) Odometrymentioning
confidence: 99%
“…We use the Voxelized Generalized Iterative Closest Point (VGICP) in [28]. This voxelized version of GICP aggregates the voxel distribution on each point, parallelizing the optimization and achieving similar accuracy to GICP but substantially faster (it runs at 30 Hz on a low-end CPU).…”
Section: Robot Odometrymentioning
confidence: 99%
“…Multiple approaches and implementations for this exist. For example, (Segal, Haehnel and Thrun 2009) for generalized ICP (GICP), (Koide et al 2021) for voxelized GICP approach with GPU based implementation and (Pan et al 2018) for global optimal matching and hybrid metric based approach. Other approaches also exist, each with its own robustness, performance and time trade-offs with respect to processing computer hardware and the nature of the scanned environment.…”
Section: Automatic Point Cloud Alignmentmentioning
confidence: 99%