ORB-SLAM-CNN: Lessons in Adding Semantic Map Construction to Feature-Based SLAM

Andrew M. Webb; Gavin Brown; Mikel Luján

doi:10.1007/978-3-030-23807-0_19

ORB-SLAM-CNN: Lessons in Adding Semantic Map Construction to Feature-Based SLAM

Andrew M. Webb^*, Gavin Brown, Mikel Luján

^*Corresponding author for this work

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

Abstract

Recent work has integrated semantics into the 3D scene models produced by visual SLAM systems. Though these systems operate close to real time, there is lacking a study of the ways to achieve real-time performance by trading off between semantic model accuracy and computational requirements. ORB-SLAM2 provides good scene accuracy and real-time processing while not requiring GPUs [1]. Following a ‘single view’ approach of overlaying a dense semantic map over the sparse SLAM scene model, we explore a method for automatically tuning the parameters of the system such that it operates in real time while maximizing prediction accuracy and map density.

Original language	English
Title of host publication	Towards Autonomous Robotic Systems - 20th Annual Conference, TAROS 2019, Proceedings
Editors	Kaspar Althoefer, Jelizaveta Konstantinova, Ketao Zhang
Publisher	Springer Nature
Pages	221-235
Number of pages	15
ISBN (Print)	9783030238063
DOIs	https://doi.org/10.1007/978-3-030-23807-0_19
Publication status	Published - 2019
Event	20th Annual Conference on Towards Autonomous Robotic Systems - London, United Kingdom Duration: 3 Jul 2019 → 5 Jul 2019

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	11649 LNAI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	20th Annual Conference on Towards Autonomous Robotic Systems
Abbreviated title	TAROS 2019
Country/Territory	United Kingdom
City	London
Period	3/07/19 → 5/07/19

Keywords

Online parameter tuning
Semantic segmentation
SLAM

Access to Document

10.1007/978-3-030-23807-0_19

Cite this

Webb, A. M., Brown, G., & Luján, M. (2019). ORB-SLAM-CNN: Lessons in Adding Semantic Map Construction to Feature-Based SLAM. In K. Althoefer, J. Konstantinova, & K. Zhang (Eds.), Towards Autonomous Robotic Systems - 20th Annual Conference, TAROS 2019, Proceedings (pp. 221-235). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11649 LNAI). Springer Nature. https://doi.org/10.1007/978-3-030-23807-0_19

Webb, Andrew M. ; Brown, Gavin ; Luján, Mikel. / ORB-SLAM-CNN : Lessons in Adding Semantic Map Construction to Feature-Based SLAM. Towards Autonomous Robotic Systems - 20th Annual Conference, TAROS 2019, Proceedings. editor / Kaspar Althoefer ; Jelizaveta Konstantinova ; Ketao Zhang. Springer Nature, 2019. pp. 221-235 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{cdeb09a341314ad5a3c215ac0e66ec30,

title = "ORB-SLAM-CNN: Lessons in Adding Semantic Map Construction to Feature-Based SLAM",

abstract = "Recent work has integrated semantics into the 3D scene models produced by visual SLAM systems. Though these systems operate close to real time, there is lacking a study of the ways to achieve real-time performance by trading off between semantic model accuracy and computational requirements. ORB-SLAM2 provides good scene accuracy and real-time processing while not requiring GPUs [1]. Following a {\textquoteleft}single view{\textquoteright} approach of overlaying a dense semantic map over the sparse SLAM scene model, we explore a method for automatically tuning the parameters of the system such that it operates in real time while maximizing prediction accuracy and map density.",

keywords = "Online parameter tuning, Semantic segmentation, SLAM",

author = "Webb, {Andrew M.} and Gavin Brown and Mikel Luj{\'a}n",

year = "2019",

doi = "10.1007/978-3-030-23807-0_19",

language = "English",

isbn = "9783030238063",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Nature",

pages = "221--235",

editor = "Kaspar Althoefer and Jelizaveta Konstantinova and Ketao Zhang",

booktitle = "Towards Autonomous Robotic Systems - 20th Annual Conference, TAROS 2019, Proceedings",

address = "United States",

note = "20th Annual Conference on Towards Autonomous Robotic Systems, TAROS 2019 ; Conference date: 03-07-2019 Through 05-07-2019",

}

Webb, AM, Brown, G & Luján, M 2019, ORB-SLAM-CNN: Lessons in Adding Semantic Map Construction to Feature-Based SLAM. in K Althoefer, J Konstantinova & K Zhang (eds), Towards Autonomous Robotic Systems - 20th Annual Conference, TAROS 2019, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11649 LNAI, Springer Nature, pp. 221-235, 20th Annual Conference on Towards Autonomous Robotic Systems, London, United Kingdom, 3/07/19. https://doi.org/10.1007/978-3-030-23807-0_19

ORB-SLAM-CNN: Lessons in Adding Semantic Map Construction to Feature-Based SLAM. / Webb, Andrew M.; Brown, Gavin; Luján, Mikel.
Towards Autonomous Robotic Systems - 20th Annual Conference, TAROS 2019, Proceedings. ed. / Kaspar Althoefer; Jelizaveta Konstantinova; Ketao Zhang. Springer Nature, 2019. p. 221-235 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11649 LNAI).

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

TY - GEN

T1 - ORB-SLAM-CNN

T2 - 20th Annual Conference on Towards Autonomous Robotic Systems

AU - Webb, Andrew M.

AU - Brown, Gavin

AU - Luján, Mikel

PY - 2019

Y1 - 2019

N2 - Recent work has integrated semantics into the 3D scene models produced by visual SLAM systems. Though these systems operate close to real time, there is lacking a study of the ways to achieve real-time performance by trading off between semantic model accuracy and computational requirements. ORB-SLAM2 provides good scene accuracy and real-time processing while not requiring GPUs [1]. Following a ‘single view’ approach of overlaying a dense semantic map over the sparse SLAM scene model, we explore a method for automatically tuning the parameters of the system such that it operates in real time while maximizing prediction accuracy and map density.

AB - Recent work has integrated semantics into the 3D scene models produced by visual SLAM systems. Though these systems operate close to real time, there is lacking a study of the ways to achieve real-time performance by trading off between semantic model accuracy and computational requirements. ORB-SLAM2 provides good scene accuracy and real-time processing while not requiring GPUs [1]. Following a ‘single view’ approach of overlaying a dense semantic map over the sparse SLAM scene model, we explore a method for automatically tuning the parameters of the system such that it operates in real time while maximizing prediction accuracy and map density.

KW - Online parameter tuning

KW - Semantic segmentation

KW - SLAM

UR - http://www.scopus.com/inward/record.url?scp=85068986638&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-23807-0_19

DO - 10.1007/978-3-030-23807-0_19

M3 - Conference contribution

AN - SCOPUS:85068986638

SN - 9783030238063

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 221

EP - 235

BT - Towards Autonomous Robotic Systems - 20th Annual Conference, TAROS 2019, Proceedings

A2 - Althoefer, Kaspar

A2 - Konstantinova, Jelizaveta

A2 - Zhang, Ketao

PB - Springer Nature

Y2 - 3 July 2019 through 5 July 2019

ER -

Webb AM, Brown G, Luján M. ORB-SLAM-CNN: Lessons in Adding Semantic Map Construction to Feature-Based SLAM. In Althoefer K, Konstantinova J, Zhang K, editors, Towards Autonomous Robotic Systems - 20th Annual Conference, TAROS 2019, Proceedings. Springer Nature. 2019. p. 221-235. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). Epub 2019 Jun 28. doi: 10.1007/978-3-030-23807-0_19

ORB-SLAM-CNN: Lessons in Adding Semantic Map Construction to Feature-Based SLAM

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Robotics and Artificial Intelligence for Nuclear (RAIN)

Cite this

ORB-SLAM-CNN: Lessons in Adding Semantic Map Construction to Feature-Based SLAM

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Projects

Robotics and Artificial Intelligence for Nuclear (RAIN)

Cite this