Towards SLAM-based Outdoor Localization using Poor GPS and 2.5 D Building Models

Clemens Arth, Ruyu Liu, Jianhua Zhang, Shengyong Chen

Research output: Contribution to conferencePaperResearchpeer-review

Abstract

In this paper, we address the topic of outdoor localization and tracking using monocular camera setups with poor GPS priors. We leverage 2.5 D building maps, which are freely available from open-source databases such as OpenStreetMap.
The main contributions of our work are a fast initialization method and a non-linear optimization scheme. The initialization upgrades a visual SLAM reconstruction with an absolute scale. The non-linear optimization uses the 2.5 D building model footprint, which further improves the tracking accuracy and the scale estimation. A pose optimization step relates the vision-based camera pose estimation from SLAM to the position information received through GPS, in order to fix the common problem of drift. We evaluate our approach on a set of challenging scenarios. The experimental results show that our approach achieves improved accuracy and robustness with an advantage in run-time over previous setups.
Original languageEnglish
Publication statusAccepted/In press - 2019
Event2019 IEEE International Symposium on Mixed and Augmented Reality - Bejing, China
Duration: 14 Oct 201918 Oct 2019

Conference

Conference2019 IEEE International Symposium on Mixed and Augmented Reality
Abbreviated titleISMAR 2019
CountryChina
CityBejing
Period14/10/1918/10/19

Fingerprint

Global positioning system
Cameras

Cite this

Arth, C., Liu, R., Zhang, J., & Chen, S. (Accepted/In press). Towards SLAM-based Outdoor Localization using Poor GPS and 2.5 D Building Models. Paper presented at 2019 IEEE International Symposium on Mixed and Augmented Reality, Bejing, China.

Towards SLAM-based Outdoor Localization using Poor GPS and 2.5 D Building Models. / Arth, Clemens; Liu, Ruyu; Zhang, Jianhua; Chen, Shengyong.

2019. Paper presented at 2019 IEEE International Symposium on Mixed and Augmented Reality, Bejing, China.

Research output: Contribution to conferencePaperResearchpeer-review

Arth, C, Liu, R, Zhang, J & Chen, S 2019, 'Towards SLAM-based Outdoor Localization using Poor GPS and 2.5 D Building Models' Paper presented at 2019 IEEE International Symposium on Mixed and Augmented Reality, Bejing, China, 14/10/19 - 18/10/19, .
Arth C, Liu R, Zhang J, Chen S. Towards SLAM-based Outdoor Localization using Poor GPS and 2.5 D Building Models. 2019. Paper presented at 2019 IEEE International Symposium on Mixed and Augmented Reality, Bejing, China.
Arth, Clemens ; Liu, Ruyu ; Zhang, Jianhua ; Chen, Shengyong. / Towards SLAM-based Outdoor Localization using Poor GPS and 2.5 D Building Models. Paper presented at 2019 IEEE International Symposium on Mixed and Augmented Reality, Bejing, China.
@conference{bf55e9ece2624c83b55e1107094941d3,
title = "Towards SLAM-based Outdoor Localization using Poor GPS and 2.5 D Building Models",
abstract = "In this paper, we address the topic of outdoor localization and tracking using monocular camera setups with poor GPS priors. We leverage 2.5 D building maps, which are freely available from open-source databases such as OpenStreetMap.The main contributions of our work are a fast initialization method and a non-linear optimization scheme. The initialization upgrades a visual SLAM reconstruction with an absolute scale. The non-linear optimization uses the 2.5 D building model footprint, which further improves the tracking accuracy and the scale estimation. A pose optimization step relates the vision-based camera pose estimation from SLAM to the position information received through GPS, in order to fix the common problem of drift. We evaluate our approach on a set of challenging scenarios. The experimental results show that our approach achieves improved accuracy and robustness with an advantage in run-time over previous setups.",
author = "Clemens Arth and Ruyu Liu and Jianhua Zhang and Shengyong Chen",
year = "2019",
language = "English",
note = "2019 IEEE International Symposium on Mixed and Augmented Reality, ISMAR 2019 ; Conference date: 14-10-2019 Through 18-10-2019",

}

TY - CONF

T1 - Towards SLAM-based Outdoor Localization using Poor GPS and 2.5 D Building Models

AU - Arth, Clemens

AU - Liu, Ruyu

AU - Zhang, Jianhua

AU - Chen, Shengyong

PY - 2019

Y1 - 2019

N2 - In this paper, we address the topic of outdoor localization and tracking using monocular camera setups with poor GPS priors. We leverage 2.5 D building maps, which are freely available from open-source databases such as OpenStreetMap.The main contributions of our work are a fast initialization method and a non-linear optimization scheme. The initialization upgrades a visual SLAM reconstruction with an absolute scale. The non-linear optimization uses the 2.5 D building model footprint, which further improves the tracking accuracy and the scale estimation. A pose optimization step relates the vision-based camera pose estimation from SLAM to the position information received through GPS, in order to fix the common problem of drift. We evaluate our approach on a set of challenging scenarios. The experimental results show that our approach achieves improved accuracy and robustness with an advantage in run-time over previous setups.

AB - In this paper, we address the topic of outdoor localization and tracking using monocular camera setups with poor GPS priors. We leverage 2.5 D building maps, which are freely available from open-source databases such as OpenStreetMap.The main contributions of our work are a fast initialization method and a non-linear optimization scheme. The initialization upgrades a visual SLAM reconstruction with an absolute scale. The non-linear optimization uses the 2.5 D building model footprint, which further improves the tracking accuracy and the scale estimation. A pose optimization step relates the vision-based camera pose estimation from SLAM to the position information received through GPS, in order to fix the common problem of drift. We evaluate our approach on a set of challenging scenarios. The experimental results show that our approach achieves improved accuracy and robustness with an advantage in run-time over previous setups.

M3 - Paper

ER -