SegMatch: Segment based place recognition in 3D point clouds

By Renaud Dubé, Daniel Dugas, Elena Stumm, Juan I. Nieto, Roland Yves Siegwart, César Cadena. IEEE International Conference on Robotics and Automation, 2016

365 citations recorded by the source corpus.

Place recognition in 3D data is a challenging task that has been commonly approached by adapting image-based solutions. Methods based on local features suffer from ambiguity and from robustness to environment changes while methods based on global features are viewpoint dependent. We propose SegMatch, a reliable place recognition algorithm based on the matching of 3D segments. Segments provide a good compromise between local and global descriptions, incorporating their strengths while reducing their individual drawbacks. SegMatch does not rely on assumptions of ‘perfect segmentation’, or on the existence of ‘objects’ in the environment, which allows for reliable execution on large scale, unstructured environments. We quantitatively demonstrate that SegMatch can achieve accurate localization at a frequency of 1Hz on the largest sequence of the KITTI odometry dataset. We furthermore show how this algorithm can reliably detect and close loops in real-time, during online operation. In addition, the source code for the SegMatch algorithm is made publicly available1.

Subfields: Aerospace Engineering, Environmental Engineering, Geology

View Paper: https://www.semanticscholar.org/paper/1cd9ba4bd3669c9a95847b57f2dc060628bfc107

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Aerospace Engineering

1. Chen Wang
2. Lihua Xie
3. Han Wang
4. François Michaud
5. Mathieu Labbé
6. Sunwook Choi
7. Giseop Kim
8. Ayoung Kim

In Environmental Engineering

1. Cyrill Stachniss
2. Jens Behley
3. Giseop Kim
4. Lihua Xie
5. Brendan Englot
6. Han Wang
7. Sunwook Choi
8. Chen Wang

In Geology

1. Sunwook Choi
2. Chen Wang
3. Lihua Xie
4. Han Wang
5. Cyrill Stachniss
6. Giseop Kim
7. Ayoung Kim
8. Jens Behley

Citations in the corpus

• LPD-Net: 3D Point Cloud Learning for Large-Scale Place Recognition and Environment Analysis (2018, methodology): “Existing researches have validated that F3D, FV and FZ are effective in solving the large-scale 3D scene analysis problem [24], and F2D and FZ are effective in solving the largescale localization problem in self-driving tasks[3, 4].”
• RTAB‐Map as an open‐source lidar and visual simultaneous localization and mapping library for large‐scale and long‐term online operation (2018, methodology): “In these lidar ‐ based SLAM approaches, only SegMatch can be used for multisession or multirobot mapping (Dubé et al., 2017).”
• Scan Context: Egocentric Spatial Descriptor for Place Recognition Within 3D Point Cloud Map (2018, methodology): “Recently, SegMatch [23] introduced a segment-based matching algorithm.”
• Scan Context++: Structural Place Recognition Robust to Rotation and Lateral Variations in Urban Environments (2021, methodology): “Not being a global descriptor as M2DP and SCD are, SegMatch only had a short range in its PR and DR curve.”
• Efficient Surfel-Based SLAM using 3D Laser Range Data in Urban Environments (2018, background): “[5] investigated an approach that matches segments extracted from a scan to find loop closures via segment-based features.”
• LeGO-LOAM: Lightweight and Ground-Optimized Lidar Odometry and Mapping on Variable Terrain (2018, methodology): “A segmentation-based registration algorithm is proposed in [18].”
• Intensity Scan Context: Coding Intensity and Geometry Relations for Loop Closure Detection (2020, methodology): “It is effective in estimating affine transform between two point clouds and is used for place recognition in the later work such as [15].”