Robust 3D reconstruction with omni-directional camera based on structure from motion

Mengcheng Song; Hiroshi Watanabe; Junich Hara

doi:10.1109/IWAIT.2018.8369715

Robust 3D reconstruction with omni-directional camera based on structure from motion

Mengcheng Song, Hiroshi Watanabe, Junich Hara

School of Fundamental Science and Engineering

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

4 Citations (Scopus)

Abstract

3D scenes reconstruction with omni-directional camera is of utmost significance as the camera can capture the 360 degrees scene. It dramatically improves the performance of Structure from Motion (SfM) since the tracked interest points will not miss with a long base-line and sharp rotation. This paper proposes a system with preprocessing and modification based on conventional SfM pipeline for equirectangular images produced by omni-directional camera. The results show that proposed system can well estimate the accurate ego-motion and sparse 3D structure of a synthetic scene as well as a real-world scene by solving the problem of distortion and lower quality at certain area of the equirectangular images.

Original language	English
Title of host publication	2018 International Workshop on Advanced Image Technology, IWAIT 2018
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	1-4
Number of pages	4
ISBN (Electronic)	9781538626153
DOIs	https://doi.org/10.1109/IWAIT.2018.8369715
Publication status	Published - 2018 May 30
Event	2018 International Workshop on Advanced Image Technology, IWAIT 2018 - Chiang Mai, Thailand Duration: 2018 Jan 7 → 2018 Jan 9

Publication series

Name	2018 International Workshop on Advanced Image Technology, IWAIT 2018

Other

Other	2018 International Workshop on Advanced Image Technology, IWAIT 2018
Country	Thailand
City	Chiang Mai
Period	18/1/7 → 18/1/9

Keywords

3D structure
Omni-directional camera
ego-motion
structure from motion

ASJC Scopus subject areas

Computer Networks and Communications
Computer Vision and Pattern Recognition
Media Technology

Access to Document

10.1109/IWAIT.2018.8369715

Fingerprint Dive into the research topics of 'Robust 3D reconstruction with omni-directional camera based on structure from motion'. Together they form a unique fingerprint.

Cite this

Song, M., Watanabe, H., & Hara, J. (2018). Robust 3D reconstruction with omni-directional camera based on structure from motion. In 2018 International Workshop on Advanced Image Technology, IWAIT 2018 (pp. 1-4). (2018 International Workshop on Advanced Image Technology, IWAIT 2018). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/IWAIT.2018.8369715

Robust 3D reconstruction with omni-directional camera based on structure from motion. / Song, Mengcheng; Watanabe, Hiroshi; Hara, Junich.

2018 International Workshop on Advanced Image Technology, IWAIT 2018. Institute of Electrical and Electronics Engineers Inc., 2018. p. 1-4 (2018 International Workshop on Advanced Image Technology, IWAIT 2018).

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

Song, M, Watanabe, H & Hara, J 2018, Robust 3D reconstruction with omni-directional camera based on structure from motion. in 2018 International Workshop on Advanced Image Technology, IWAIT 2018. 2018 International Workshop on Advanced Image Technology, IWAIT 2018, Institute of Electrical and Electronics Engineers Inc., pp. 1-4, 2018 International Workshop on Advanced Image Technology, IWAIT 2018, Chiang Mai, Thailand, 18/1/7. https://doi.org/10.1109/IWAIT.2018.8369715

Song M, Watanabe H, Hara J. Robust 3D reconstruction with omni-directional camera based on structure from motion. In 2018 International Workshop on Advanced Image Technology, IWAIT 2018. Institute of Electrical and Electronics Engineers Inc. 2018. p. 1-4. (2018 International Workshop on Advanced Image Technology, IWAIT 2018). https://doi.org/10.1109/IWAIT.2018.8369715

@inproceedings{5635245b815747b5b2618254384fa7d5,

title = "Robust 3D reconstruction with omni-directional camera based on structure from motion",

abstract = "3D scenes reconstruction with omni-directional camera is of utmost significance as the camera can capture the 360 degrees scene. It dramatically improves the performance of Structure from Motion (SfM) since the tracked interest points will not miss with a long base-line and sharp rotation. This paper proposes a system with preprocessing and modification based on conventional SfM pipeline for equirectangular images produced by omni-directional camera. The results show that proposed system can well estimate the accurate ego-motion and sparse 3D structure of a synthetic scene as well as a real-world scene by solving the problem of distortion and lower quality at certain area of the equirectangular images.",

keywords = "3D structure, Omni-directional camera, ego-motion, structure from motion",

author = "Mengcheng Song and Hiroshi Watanabe and Junich Hara",

note = "Publisher Copyright: {\textcopyright} 2018 IEEE. Copyright: Copyright 2018 Elsevier B.V., All rights reserved.; 2018 International Workshop on Advanced Image Technology, IWAIT 2018 ; Conference date: 07-01-2018 Through 09-01-2018",

year = "2018",

month = may,

day = "30",

doi = "10.1109/IWAIT.2018.8369715",

language = "English",

series = "2018 International Workshop on Advanced Image Technology, IWAIT 2018",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "1--4",

booktitle = "2018 International Workshop on Advanced Image Technology, IWAIT 2018",

}

TY - GEN

T1 - Robust 3D reconstruction with omni-directional camera based on structure from motion

AU - Song, Mengcheng

AU - Watanabe, Hiroshi

AU - Hara, Junich

PY - 2018/5/30

Y1 - 2018/5/30

N2 - 3D scenes reconstruction with omni-directional camera is of utmost significance as the camera can capture the 360 degrees scene. It dramatically improves the performance of Structure from Motion (SfM) since the tracked interest points will not miss with a long base-line and sharp rotation. This paper proposes a system with preprocessing and modification based on conventional SfM pipeline for equirectangular images produced by omni-directional camera. The results show that proposed system can well estimate the accurate ego-motion and sparse 3D structure of a synthetic scene as well as a real-world scene by solving the problem of distortion and lower quality at certain area of the equirectangular images.

AB - 3D scenes reconstruction with omni-directional camera is of utmost significance as the camera can capture the 360 degrees scene. It dramatically improves the performance of Structure from Motion (SfM) since the tracked interest points will not miss with a long base-line and sharp rotation. This paper proposes a system with preprocessing and modification based on conventional SfM pipeline for equirectangular images produced by omni-directional camera. The results show that proposed system can well estimate the accurate ego-motion and sparse 3D structure of a synthetic scene as well as a real-world scene by solving the problem of distortion and lower quality at certain area of the equirectangular images.

KW - 3D structure

KW - Omni-directional camera

KW - ego-motion

KW - structure from motion

UR - http://www.scopus.com/inward/record.url?scp=85048827140&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85048827140&partnerID=8YFLogxK

U2 - 10.1109/IWAIT.2018.8369715

DO - 10.1109/IWAIT.2018.8369715

M3 - Conference contribution

AN - SCOPUS:85048827140

T3 - 2018 International Workshop on Advanced Image Technology, IWAIT 2018

SP - 1

EP - 4

BT - 2018 International Workshop on Advanced Image Technology, IWAIT 2018

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2018 International Workshop on Advanced Image Technology, IWAIT 2018

Y2 - 7 January 2018 through 9 January 2018

ER -

Robust 3D reconstruction with omni-directional camera based on structure from motion

Abstract

Publication series

Other

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Cite this