TBPos: Dataset for Large-Scale Precision Visual Localization

Masud Fahim; Ilona Söchting; Luca Ferranti; Juho Kannala; Jani Boutellier

doi:10.1007/978-3-031-31435-3_6

TBPos: Dataset for Large-Scale Precision Visual Localization

Masud Fahim, Ilona Söchting, Luca Ferranti, Juho Kannala, Jani Boutellier^*

^*Corresponding author for this work

University of Vaasa

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

Abstract

Image based localization is a classical computer vision challenge, with several well-known datasets. Generally, datasets consist of a visual 3D database that captures the modeled scenery, as well as query images whose 3D pose is to be discovered. Usually the query images have been acquired with a camera that differs from the imaging hardware used to collect the 3D database; consequently, it is hard to acquire accurate ground truth poses between query images and the 3D database. As the accuracy of visual localization algorithms constantly improves, precise ground truth becomes increasingly important. This paper proposes TBPos, a novel large-scale visual dataset for image based positioning, which provides query images with fully accurate ground truth poses: both the database images and the query images have been derived from the same laser scanner data. In the experimental part of the paper, the proposed dataset is evaluated by means of an image-based localization pipeline.

Original language	English
Title of host publication	Image Analysis - Proceedings of SCIA 2023
Editors	Rikke Gade, Michael Felsberg, Joni-Kristian Kämäräinen
Publisher	Springer
Pages	84-94
Number of pages	11
ISBN (Print)	9783031314346
DOIs	https://doi.org/10.1007/978-3-031-31435-3_6
Publication status	Published - 2023
MoE publication type	A4 Conference publication
Event	Scandinavian Conference on Image Analysis - Levi, Kittilä, Finland Duration: 18 Apr 2023 → 21 Apr 2023 Conference number: 23

Publication series

Name	Lecture Notes in Computer Science
Volume	13885 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	Scandinavian Conference on Image Analysis
Abbreviated title	SCIA
Country/Territory	Finland
City	Kittilä
Period	18/04/2023 → 21/04/2023

Keywords

6DoF pose
Computer vision
Dataset
Visual localization

Access to Document

10.1007/978-3-031-31435-3_6

https://urn.fi/URN:NBN:fi-fe20231012139903

OpenUrl availability

Full text

Cite this

@inproceedings{e2116089862b489fb79c1bd2015f78f9,

title = "TBPos: Dataset for Large-Scale Precision Visual Localization",

abstract = "Image based localization is a classical computer vision challenge, with several well-known datasets. Generally, datasets consist of a visual 3D database that captures the modeled scenery, as well as query images whose 3D pose is to be discovered. Usually the query images have been acquired with a camera that differs from the imaging hardware used to collect the 3D database; consequently, it is hard to acquire accurate ground truth poses between query images and the 3D database. As the accuracy of visual localization algorithms constantly improves, precise ground truth becomes increasingly important. This paper proposes TBPos, a novel large-scale visual dataset for image based positioning, which provides query images with fully accurate ground truth poses: both the database images and the query images have been derived from the same laser scanner data. In the experimental part of the paper, the proposed dataset is evaluated by means of an image-based localization pipeline.",

keywords = "6DoF pose, Computer vision, Dataset, Visual localization",

author = "Masud Fahim and Ilona S{\"o}chting and Luca Ferranti and Juho Kannala and Jani Boutellier",

note = "Publisher Copyright: {\textcopyright} 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.; Scandinavian Conference on Image Analysis, SCIA ; Conference date: 18-04-2023 Through 21-04-2023",

year = "2023",

doi = "10.1007/978-3-031-31435-3_6",

language = "English",

isbn = "9783031314346",

series = "Lecture Notes in Computer Science",

publisher = "Springer",

pages = "84--94",

editor = "Rikke Gade and Michael Felsberg and Joni-Kristian K{\"a}m{\"a}r{\"a}inen",

booktitle = "Image Analysis - Proceedings of SCIA 2023",

address = "Germany",

}

Fahim, M, Söchting, I, Ferranti, L , Kannala, J & Boutellier, J 2023, TBPos: Dataset for Large-Scale Precision Visual Localization. in R Gade, M Felsberg & J-K Kämäräinen (eds), Image Analysis - Proceedings of SCIA 2023. Lecture Notes in Computer Science, vol. 13885 LNCS, Springer, pp. 84-94, Scandinavian Conference on Image Analysis, Kittilä, Finland, 18/04/2023. https://doi.org/10.1007/978-3-031-31435-3_6

TBPos: Dataset for Large-Scale Precision Visual Localization. / Fahim, Masud; Söchting, Ilona; Ferranti, Luca et al.
Image Analysis - Proceedings of SCIA 2023. ed. / Rikke Gade; Michael Felsberg; Joni-Kristian Kämäräinen. Springer, 2023. p. 84-94 (Lecture Notes in Computer Science; Vol. 13885 LNCS).

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

TY - GEN

T1 - TBPos

T2 - Scandinavian Conference on Image Analysis

AU - Fahim, Masud

AU - Söchting, Ilona

AU - Ferranti, Luca

AU - Kannala, Juho

AU - Boutellier, Jani

N1 - Conference code: 23

PY - 2023

Y1 - 2023

N2 - Image based localization is a classical computer vision challenge, with several well-known datasets. Generally, datasets consist of a visual 3D database that captures the modeled scenery, as well as query images whose 3D pose is to be discovered. Usually the query images have been acquired with a camera that differs from the imaging hardware used to collect the 3D database; consequently, it is hard to acquire accurate ground truth poses between query images and the 3D database. As the accuracy of visual localization algorithms constantly improves, precise ground truth becomes increasingly important. This paper proposes TBPos, a novel large-scale visual dataset for image based positioning, which provides query images with fully accurate ground truth poses: both the database images and the query images have been derived from the same laser scanner data. In the experimental part of the paper, the proposed dataset is evaluated by means of an image-based localization pipeline.

AB - Image based localization is a classical computer vision challenge, with several well-known datasets. Generally, datasets consist of a visual 3D database that captures the modeled scenery, as well as query images whose 3D pose is to be discovered. Usually the query images have been acquired with a camera that differs from the imaging hardware used to collect the 3D database; consequently, it is hard to acquire accurate ground truth poses between query images and the 3D database. As the accuracy of visual localization algorithms constantly improves, precise ground truth becomes increasingly important. This paper proposes TBPos, a novel large-scale visual dataset for image based positioning, which provides query images with fully accurate ground truth poses: both the database images and the query images have been derived from the same laser scanner data. In the experimental part of the paper, the proposed dataset is evaluated by means of an image-based localization pipeline.

KW - 6DoF pose

KW - Computer vision

KW - Dataset

KW - Visual localization

UR - http://www.scopus.com/inward/record.url?scp=85161406166&partnerID=8YFLogxK

U2 - 10.1007/978-3-031-31435-3_6

DO - 10.1007/978-3-031-31435-3_6

M3 - Conference contribution

AN - SCOPUS:85161406166

SN - 9783031314346

T3 - Lecture Notes in Computer Science

SP - 84

EP - 94

BT - Image Analysis - Proceedings of SCIA 2023

A2 - Gade, Rikke

A2 - Felsberg, Michael

A2 - Kämäräinen, Joni-Kristian

PB - Springer

Y2 - 18 April 2023 through 21 April 2023

ER -