A Brief Review on Loop Closure Detection with 3D Point Cloud

A robust loop closure detection (LCD) in simultaneous localization and mapping (SLAM) can lead to more consistent and accurate map, which is a key part of SLAM. Currently, there are two main kind of data sources used for loop closure detection, the 2D image or 3D points. The 2D image-based LCD method has been extensively studied and many literatures have been published. However, this kind of method is sensitive to the illumination, the change of weather conditions and season et al. With the rapid development of self-driving technique, LiDAR-based SLAM has attracted many researchers because of its advantages. In this paper, we make a summarization on the recent developments of LCD method with 3D data source, especially with 3D point cloud input, and classify these methods into three groups, point feature based, segmentation/object based, and learning-based methods. For each group, the typical references are introduced. For the feature based method, four kind of features for LCD are introduced. Based on these introduction, the readers can get an overview on these LCD methods and even their working principles. Some discussions and the future trend on LCD with 3D point are also presented.

[1]  Michael Bosse,et al.  Keypoint design and evaluation for place recognition in 2D lidar maps , 2009, Robotics Auton. Syst..

[2]  Eduardo Mario Nebot,et al.  Scan-SLAM: Combining EKF-SLAM and Scan Correlation , 2005, FSR.

[3]  Jie Li,et al.  An Evaluation of Deep Learning in Loop Closure Detection for Visual SLAM , 2017, 2017 IEEE International Conference on Internet of Things (iThings) and IEEE Green Computing and Communications (GreenCom) and IEEE Cyber, Physical and Social Computing (CPSCom) and IEEE Smart Data (SmartData).

[4]  Andrew E. Johnson,et al.  Surface matching for object recognition in complex three-dimensional scenes , 1998, Image Vis. Comput..

[5]  Wolfram Burgard,et al.  Place recognition in 3D scans using a combination of bag of words and point feature based relative pose estimation , 2011, 2011 IEEE/RSJ International Conference on Intelligent Robots and Systems.

[6]  Markus Vincze,et al.  Ensemble of shape functions for 3D object classification , 2011, 2011 IEEE International Conference on Robotics and Biomimetics.

[7]  Renaud Dubé,et al.  Structure-based vision-laser matching , 2016, 2016 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).

[8]  Leonidas J. Guibas,et al.  Frustum PointNets for 3D Object Detection from RGB-D Data , 2017, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[9]  Hannes Sommer,et al.  Incremental-Segment-Based Localization in 3-D Point Clouds , 2018, IEEE Robotics and Automation Letters.

[10]  Abel Gawel,et al.  Point cloud descriptors for place recognition using sparse visual information , 2016, 2016 IEEE International Conference on Robotics and Automation (ICRA).

[11]  Margarita Chli,et al.  Viewpoint-Tolerant Place Recognition Combining 2D and 3D Information for UAV Navigation , 2018, 2018 IEEE International Conference on Robotics and Automation (ICRA).

[12]  Peter Biber,et al.  The normal distributions transform: a new approach to laser scan matching , 2003, Proceedings 2003 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2003) (Cat. No.03CH37453).

[13]  Xiaowu Chen,et al.  3D Mesh Labeling via Deep Convolutional Neural Networks , 2015, ACM Trans. Graph..

[14]  Leonidas J. Guibas,et al.  Volumetric and Multi-view CNNs for Object Classification on 3D Data , 2016, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[15]  Xiaqing Ding,et al.  LocNet: Global Localization in 3D Point Clouds for Mobile Vehicles , 2017, 2018 IEEE Intelligent Vehicles Symposium (IV).

[16]  Dirk Schulz,et al.  A fast histogram-based similarity measure for detecting loop closures in 3-D LIDAR data , 2015, 2015 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).

[17]  Michael Bosse,et al.  Place recognition using keypoint voting in large 3D lidar datasets , 2013, 2013 IEEE International Conference on Robotics and Automation.

[18]  Sebastian Scherer,et al.  VoxNet: A 3D Convolutional Neural Network for real-time object recognition , 2015, 2015 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).

[19]  Jianxiong Xiao,et al.  3D ShapeNets: A deep representation for volumetric shapes , 2014, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[20]  Andrew E. Johnson,et al.  Spin-Images: A Representation for 3-D Surface Matching , 1997 .

[21]  Maks Ovsjanikov,et al.  PCPNet Learning Local Shape Properties from Raw Point Clouds , 2017, Comput. Graph. Forum.

[22]  Javier González,et al.  Scene structure registration for localization and mapping , 2016, Robotics Auton. Syst..

[23]  John J. Leonard,et al.  Toward Object-based Place Recognition in Dense RGB-D Maps , 2015 .

[24]  Gary R. Bradski,et al.  Fast 3D recognition and pose using the Viewpoint Feature Histogram , 2010, 2010 IEEE/RSJ International Conference on Intelligent Robots and Systems.

[25]  Huosheng Hu,et al.  3-D-Laser-Based Scene Measurement and Place Recognition for Mobile Robots in Dynamic Indoor Environments , 2013, IEEE Transactions on Instrumentation and Measurement.

[26]  Xiaqing Ding,et al.  LocNet: Global localization in 3D point clouds for mobile robots , 2017, ArXiv.

[27]  Michael Bosse,et al.  Place Recognition Using Regional Point Descriptors for 3D Mapping , 2009, FSR.

[28]  Javier González,et al.  Fast place recognition with plane-based maps , 2013, 2013 IEEE International Conference on Robotics and Automation.

[29]  Wolfram Burgard,et al.  Robust place recognition for 3D range data based on point features , 2010, 2010 IEEE International Conference on Robotics and Automation.

[30]  Surya P. N. Singh,et al.  A Pipeline for the Segmentation and Classification of 3D Point Clouds , 2010, ISER.

[31]  Nico Blodow,et al.  Aligning point cloud views using persistent feature histograms , 2008, 2008 IEEE/RSJ International Conference on Intelligent Robots and Systems.

[32]  Andreas Nüchter,et al.  Automatic Appearance-Based Loop Detection from 3 D Laser Data Using the Normal Distributions Transform , 2009 .

[33]  Zhong Liu,et al.  An Effective 3D Shape Descriptor for Object Recognition with RGB-D Sensors , 2017, Sensors.

[34]  Thomas B. Schön,et al.  Learning to close loops from range data , 2011, Int. J. Robotics Res..

[35]  Renaud Dubé,et al.  SegMatch: Segment based place recognition in 3D point clouds , 2016, 2017 IEEE International Conference on Robotics and Automation (ICRA).

[36]  Leonidas J. Guibas,et al.  FPNN: Field Probing Neural Networks for 3D Data , 2016, NIPS.

[37]  Tim Bailey,et al.  Scan segments matching for pairwise 3D alignment , 2012, 2012 IEEE International Conference on Robotics and Automation.

[38]  Ingmar Posner,et al.  Voting for Voting in Online Point Cloud Object Detection , 2015, Robotics: Science and Systems.

[39]  Renaud Dubé,et al.  SegMap: 3D Segment Mapping using Data-Driven Descriptors , 2018, Robotics: Science and Systems.

[40]  Meng Wang,et al.  3D deep shape descriptor , 2015, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[41]  Nico Blodow,et al.  Fast Point Feature Histograms (FPFH) for 3D registration , 2009, 2009 IEEE International Conference on Robotics and Automation.

[42]  Renaud Dubé,et al.  SegMatch: Segment based loop-closure for 3D point clouds , 2016, ArXiv.

[43]  Leonidas J. Guibas,et al.  PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[44]  Leonidas J. Guibas,et al.  PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space , 2017, NIPS.

[45]  Subhransu Maji,et al.  Multi-view Convolutional Neural Networks for 3D Shape Recognition , 2015, 2015 IEEE International Conference on Computer Vision (ICCV).