Comparison of ROS-Based Monocular Visual SLAM Methods: DSO, LDSO, ORB-SLAM2 and DynaSLAM

被引：19

作者：

Mingachev, Eldar ^{[1
]}

Lavrenov, Roman ^{[1
]}

Tsoy, Tatyana ^{[1
]}

Matsuno, Fumitoshi ^{[2
]}

Svinin, Mikhail ^{[3
]}

Suthakorn, Jackrit ^{[4
]}

Magid, Evgeni ^{[1
]}

机构：

[1] Kazan Fed Univ, Higher Inst Informat Technol & Intelligent Syst, Intelligent Robot Dept, Lab Intelligent Robot Syst LIRS, Kazan, Russia

[2] Kyoto Univ, Dept Mech Engn & Sci, Kyoto 6158540, Japan

[3] Ritsumeikan Univ, Informat Sci & Engn Dept, 1-1-1 Noji Higashi, Kusatsu, Shiga 5258577, Japan

[4] Mahidol Univ, Biomed Engn Dept, 4 999 Phuttamonthon, Salaya 73170, Nakhon Pathom, Thailand

来源：

INTERACTIVE COLLABORATIVE ROBOTICS, ICR 2020 | 2020年 / 12336卷

基金：

日本科学技术振兴机构; 俄罗斯基础研究基金会;

关键词：

Simultaneous localization and mapping; Visual SLAM; Monocular SLAM; Visual odometry; State estimation; Path planning; Benchmark testing; Robot sensing systems;

D O I：

10.1007/978-3-030-60337-3_22

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

Stable and robust path planning of a ground mobile robot requires a combination of accuracy and low latency in its state estimation. Yet, state estimation algorithms should provide these under computational and power constraints of a robot embedded hardware. The presented study offers a comparative analysis of four cutting edge publicly available within robot operating system (ROS) monocular simultaneous localization and mapping methods: DSO, LDSO, ORB-SLAM2, and DynaSLAM. The analysis considers pose estimation accuracy (alignment, absolute trajectory, and relative pose root mean square error) and trajectory precision of the four methods at TUM-Mono and EuRoC datasets.

引用

页码：222 / 233

页数：12

共 26 条

[1] Divergent trinocular vision observers design for extended Kalman filter robot state estimation [J].

Alonso Martinez-Garcia, Edgar ;

Rivero-Juarez, Joaquin ;

Abril Torres-Mendez, Luz ;

Enrique Rodas-Osollo, Jorge .

PROCEEDINGS OF THE INSTITUTION OF MECHANICAL ENGINEERS PART I-JOURNAL OF SYSTEMS AND CONTROL ENGINEERING, 2019, 233 (05) :524-547

[2]

Bereznikov D, 2020, PROCEEDINGS OF THE 2020 INTERNATIONAL CONFERENCE ON ARTIFICIAL LIFE AND ROBOTICS (ICAROB2020), P444

[3] DynaSLAM: Tracking, Mapping, and Inpainting in Dynamic Scenes [J].

Bescos, Berta ;

Facil, Jose M. ;

Civera, Javier ;

Neira, Jose .

IEEE ROBOTICS AND AUTOMATION LETTERS, 2018, 3 (04) :4076-4083

[4] Sparse 3D Point-Cloud Map Upsampling and Noise Removal as a vSLAM Post-Processing Step: Experimental Evaluation [J].

Bokovoy, Andrey ;

Yakovlev, Konstantin .

INTERACTIVE COLLABORATIVE ROBOTICS, ICR 2018, 2018, 11097 :23-33

[5] The EuRoC micro aerial vehicle datasets [J].

Burri, Michael ;

Nikolic, Janosch ;

Gohl, Pascal ;

Schneider, Thomas ;

Rehder, Joern ;

Omari, Sammy ;

Achtelik, Markus W. ;

Siegwart, Roland .

INTERNATIONAL JOURNAL OF ROBOTICS RESEARCH, 2016, 35 (10) :1157-1163

[6] Comparative Analysis of ROS-Based Monocular SLAM Methods for Indoor Navigation [J].

Buyval, Alexander ;

Afanasyev, Ilya ;

Magid, Evgeni .

NINTH INTERNATIONAL CONFERENCE ON MACHINE VISION (ICMV 2016), 2017, 10341

[7]

Carballo A, 2018, IEEE INT C INTELL TR, P3811, DOI 10.1109/ITSC.2018.8569764

[8]

Delmerico J, 2018, IEEE INT CONF ROBOT, P2502

[9] A solution to the simultaneous localization and map building (SLAM) problem [J].

Dissanayake, MWMG ;

Newman, P ;

Clark, S ;

Durrant-Whyte, HF ;

Csorba, M .

IEEE TRANSACTIONS ON ROBOTICS AND AUTOMATION, 2001, 17 (03) :229-241

[10]

Engel J., 2016, ARXIV

← 1 2 3 →