mlsanityData to 14 Oct 2026

Stress-Testing Point Cloud Registration on Automotive LiDAR

arXiv 2204.07719 ↗First posted 16 Apr 2022cs.CV cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

Rigid Point Cloud Registration (PCR) algorithms aim to estimate the 6-DOF relative motion between two point clouds, which is important in various fields, including autonomous driving. Recent years have seen a significant improvement in global PCR algorithms, i.e. algorithms that can handle a large relative motion. This has been demonstrated in various scenarios, including indoor scenes, but has only been minimally tested in the Automotive setting, where point clouds are produced by vehicle-mounted LiDAR sensors. In this work, we aim to answer questions that are important for automotive applications, including: which of the new algorithms is the most accurate, and which is fastest? How transferable are deep-learning approaches, e.g. what happens when you train a network with data from Boston, and run it in a vehicle in Singapore? How small can the overlap between point clouds be before the algorithms start to deteriorate? To what extent are the algorithms rotation invariant? Our results are at times surprising. When comparing robust parameter estimation methods for registration, we find that the fastest and most accurate is not one of the newest approaches. Instead, it is a modern variant of the well known RANSAC technique. We also suggest a new outlier filtering method, Grid-Prioritized Filtering (GPF), to further improve it. An additional contribution of this work is an algorithm for selecting challenging sets of frame-pairs from automotive LiDAR datasets. This enables meaningful benchmarking in the Automotive LiDAR setting, and can also improve training for learning algorithms.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 22 of its 37 references that are in the map. First posted on arXiv 16 Apr 2022.

Builds on

22 of 37 references.

References · 22

  1. 2019nuScenes: A multimodal dataset for autonomous drivingcited by 4,523
  2. 20194D Spatio-Temporal ConvNets: Minkowski Convolutional Neural Networkscited by 1,208
  3. 2018Open3D: A Modern Library for 3D Data Processingcited by 721
  4. 20163DMatch: Learning Local Geometric Descriptors from RGB-D Reconstructionscited by 336
  5. 2019Deep Closest Point: Learning Representations for Point Cloud Registrationcited by 321
  6. 2020PREDATOR: Registration of 3D Point Clouds with Low Overlapcited by 290
  7. 2020TEASER: Fast and Certifiable Point Cloud Registrationcited by 274
  8. 2019PointNetLK: Robust & Efficient Point Cloud Registration using PointNetcited by 233
  9. 2020Deep Global Registrationcited by 229
  10. 2020RPM-Net: Robust Point Matching using Learned Featurescited by 182
  11. 2020D3Feat: Joint Learning of Dense Detection and Description of 3D Local Featurescited by 177
  12. 2019PRNet: Self-Supervised Learning for Partial-to-Partial Registrationcited by 146
10 more
  1. 2021PointDSC: Robust Point Cloud Registration using Deep Spatial Consistencycited by 142
  2. 2017Graph-Cut RANSACcited by 133
  3. 2020DeepGMR: Learning Latent Gaussian Mixture Models for Registrationcited by 91
  4. 2019DeepICP: An End-to-End Deep Neural Network for 3D Point Cloud Registrationcited by 77
  5. 2019PCRNet: Point Cloud Registration Network using PointNet Encodingcited by 63
  6. 2021HRegNet: A Hierarchical Network for Large-scale Outdoor LiDAR Point Cloud Registrationcited by 48
  7. 2020PointGMM: a Neural GMM Network for Point Cloudscited by 33
  8. 2020A Benchmark for Point Clouds Registration Algorithmscited by 5
  9. 2018Performance Evaluation of 3D Correspondence Grouping Algorithmscited by 4
  10. 2020Best Buddies Registration for Point Cloudscited by 3

Built on by

Not cited yet.