Monocular vision based mobile robot 3D map building

Maohai Li; Lining Sun

doi:10.4028/www.scientific.net/AMM.43.49

Monocular vision based mobile robot 3D map building

Maohai Li^*
, Lining Sun

^*Corresponding author for this work

School of Mechatronics Engineering

Soochow University
Harbin Institute of Technology
School of Computer Science and Technology, Harbin Institute of Technology

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

A robust dense 3D feature map is built only with monocular vision and odometry. Monocular vision mounted on the robot front-end tracks the 3D natural landmarks, which are structured with matching Scale Invariant Feature Transform (SIFT) feature matching pairs. SIFT features are highly distinctive and invariant to image scaling, rotation, and change in 3D viewpoints. A fast SIFT feature matching algorithm is implemented with the KD-Tree based nearest search approach in the time cost of Ο(log ₂ ^N), and matches with large error are eliminated by epipolar line restriction. A map building algorithm based on 3D spatial SIFT landmarks is designed and implemented. Experiment results on Pioneer mobile robot in a real indoor environment show the superior performance of our proposed method.

Original language	English
Title of host publication	Advance in Mechatronics Technology
Pages	49-52
Number of pages	4
DOIs	https://doi.org/10.4028/www.scientific.net/AMM.43.49
State	Published - 2011
Event	6th China-Japan International Conference on Mechatronics, CJCM'2010 - Zhenjiang, China Duration: 10 Sep 2010 → 12 Sep 2010

Publication series

Name	Applied Mechanics and Materials
Volume	43
ISSN (Print)	1660-9336
ISSN (Electronic)	1662-7482

Conference

Conference	6th China-Japan International Conference on Mechatronics, CJCM'2010
Country/Territory	China
City	Zhenjiang
Period	10/09/10 → 12/09/10

Keywords

Feature map
KD-Tree
Mobile robot
Monocular vision
Scale invariant feature transform

Access to Document

10.4028/www.scientific.net/AMM.43.49

Cite this

@inproceedings{6e149e9248284e578ea890cabcef90b3,

title = "Monocular vision based mobile robot 3D map building",

abstract = "A robust dense 3D feature map is built only with monocular vision and odometry. Monocular vision mounted on the robot front-end tracks the 3D natural landmarks, which are structured with matching Scale Invariant Feature Transform (SIFT) feature matching pairs. SIFT features are highly distinctive and invariant to image scaling, rotation, and change in 3D viewpoints. A fast SIFT feature matching algorithm is implemented with the KD-Tree based nearest search approach in the time cost of Ο(log 2 N), and matches with large error are eliminated by epipolar line restriction. A map building algorithm based on 3D spatial SIFT landmarks is designed and implemented. Experiment results on Pioneer mobile robot in a real indoor environment show the superior performance of our proposed method.",

keywords = "Feature map, KD-Tree, Mobile robot, Monocular vision, Scale invariant feature transform",

author = "Maohai Li and Lining Sun",

year = "2011",

doi = "10.4028/www.scientific.net/AMM.43.49",

language = "英语",

isbn = "9783037850008",

series = "Applied Mechanics and Materials",

pages = "49--52",

booktitle = "Advance in Mechatronics Technology",

note = "6th China-Japan International Conference on Mechatronics, CJCM'2010 ; Conference date: 10-09-2010 Through 12-09-2010",

}

TY - GEN

T1 - Monocular vision based mobile robot 3D map building

AU - Li, Maohai

AU - Sun, Lining

PY - 2011

Y1 - 2011

N2 - A robust dense 3D feature map is built only with monocular vision and odometry. Monocular vision mounted on the robot front-end tracks the 3D natural landmarks, which are structured with matching Scale Invariant Feature Transform (SIFT) feature matching pairs. SIFT features are highly distinctive and invariant to image scaling, rotation, and change in 3D viewpoints. A fast SIFT feature matching algorithm is implemented with the KD-Tree based nearest search approach in the time cost of Ο(log 2 N), and matches with large error are eliminated by epipolar line restriction. A map building algorithm based on 3D spatial SIFT landmarks is designed and implemented. Experiment results on Pioneer mobile robot in a real indoor environment show the superior performance of our proposed method.

AB - A robust dense 3D feature map is built only with monocular vision and odometry. Monocular vision mounted on the robot front-end tracks the 3D natural landmarks, which are structured with matching Scale Invariant Feature Transform (SIFT) feature matching pairs. SIFT features are highly distinctive and invariant to image scaling, rotation, and change in 3D viewpoints. A fast SIFT feature matching algorithm is implemented with the KD-Tree based nearest search approach in the time cost of Ο(log 2 N), and matches with large error are eliminated by epipolar line restriction. A map building algorithm based on 3D spatial SIFT landmarks is designed and implemented. Experiment results on Pioneer mobile robot in a real indoor environment show the superior performance of our proposed method.

KW - Feature map

KW - KD-Tree

KW - Mobile robot

KW - Monocular vision

KW - Scale invariant feature transform

UR - https://www.scopus.com/pages/publications/78651299405

U2 - 10.4028/www.scientific.net/AMM.43.49

DO - 10.4028/www.scientific.net/AMM.43.49

M3 - 会议稿件

AN - SCOPUS:78651299405

SN - 9783037850008

T3 - Applied Mechanics and Materials

SP - 49

EP - 52

BT - Advance in Mechatronics Technology

T2 - 6th China-Japan International Conference on Mechatronics, CJCM'2010

Y2 - 10 September 2010 through 12 September 2010

ER -

Monocular vision based mobile robot 3D map building

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this