Investigating Visual Localization Using Geospatial Meshes

This paper investigates the use of geospatial mesh data for visual localization, focusing on city-scale aerial meshes as map representations for locating ground-level query images captured by smartphones. Visual localization, essential for applications such as robotics and augmented reality, traditi...

Full description

Saved in:
Bibliographic Details
Main Authors: F. Vultaggio, P. Fanta-Jende, M. Schörghuber, A. Kern, M. Gerke
Format: Article
Language:English
Published: Copernicus Publications 2024-12-01
Series:The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Online Access:https://isprs-archives.copernicus.org/articles/XLVIII-2-W8-2024/447/2024/isprs-archives-XLVIII-2-W8-2024-447-2024.pdf
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1846122165441658880
author F. Vultaggio
F. Vultaggio
P. Fanta-Jende
M. Schörghuber
A. Kern
M. Gerke
author_facet F. Vultaggio
F. Vultaggio
P. Fanta-Jende
M. Schörghuber
A. Kern
M. Gerke
author_sort F. Vultaggio
collection DOAJ
description This paper investigates the use of geospatial mesh data for visual localization, focusing on city-scale aerial meshes as map representations for locating ground-level query images captured by smartphones. Visual localization, essential for applications such as robotics and augmented reality, traditionally relies on Structure-from-Motion (SfM) reconstructions or image collections as maps. However, mesh-based approaches offer dense spatial representation, memory efficiency, and real-time rendering capabilities. In this work, we evaluate initialization strategies, image matching techniques, and pose refinement methods for mesh-based localization pipelines, comparing the performance of both traditional and deep-learning-based techniques in image matching between real and synthetic views. We created a dataset from nadir and oblique aerial imagery and accurately georeferenced smartphone images to test cross-modal localization. Our findings demonstrate that combining global feature retrieval with GNSS-based spatial filtering yields significant improvements in accuracy and efficiency, achieving submeter positional and subdegree rotational errors. This study advances scalable visual localization using meshes and highlights the potential of integrating smartphone GNSS data for improved performance in urban environments.
format Article
id doaj-art-2c70677d24c140789eb8a94fbe318688
institution Kabale University
issn 1682-1750
2194-9034
language English
publishDate 2024-12-01
publisher Copernicus Publications
record_format Article
series The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
spelling doaj-art-2c70677d24c140789eb8a94fbe3186882024-12-15T02:43:15ZengCopernicus PublicationsThe International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences1682-17502194-90342024-12-01XLVIII-2-W8-202444745410.5194/isprs-archives-XLVIII-2-W8-2024-447-2024Investigating Visual Localization Using Geospatial MeshesF. Vultaggio0F. Vultaggio1P. Fanta-Jende2M. Schörghuber3A. Kern4M. Gerke5Austrian Institute of Technology, Center for Vision, Automation and Control, Unit Assistive and Autonomous Systems, AustriaTechnische Universität Braunschweig, Institute of Geodesy and Photogrammetry, GermanyAustrian Institute of Technology, Center for Vision, Automation and Control, Unit Assistive and Autonomous Systems, AustriaAustrian Institute of Technology, Center for Vision, Automation and Control, Unit Assistive and Autonomous Systems, AustriaTechnische Universität Braunschweig, Institute of Flight Guidance, GermanyTechnische Universität Braunschweig, Institute of Geodesy and Photogrammetry, GermanyThis paper investigates the use of geospatial mesh data for visual localization, focusing on city-scale aerial meshes as map representations for locating ground-level query images captured by smartphones. Visual localization, essential for applications such as robotics and augmented reality, traditionally relies on Structure-from-Motion (SfM) reconstructions or image collections as maps. However, mesh-based approaches offer dense spatial representation, memory efficiency, and real-time rendering capabilities. In this work, we evaluate initialization strategies, image matching techniques, and pose refinement methods for mesh-based localization pipelines, comparing the performance of both traditional and deep-learning-based techniques in image matching between real and synthetic views. We created a dataset from nadir and oblique aerial imagery and accurately georeferenced smartphone images to test cross-modal localization. Our findings demonstrate that combining global feature retrieval with GNSS-based spatial filtering yields significant improvements in accuracy and efficiency, achieving submeter positional and subdegree rotational errors. This study advances scalable visual localization using meshes and highlights the potential of integrating smartphone GNSS data for improved performance in urban environments.https://isprs-archives.copernicus.org/articles/XLVIII-2-W8-2024/447/2024/isprs-archives-XLVIII-2-W8-2024-447-2024.pdf
spellingShingle F. Vultaggio
F. Vultaggio
P. Fanta-Jende
M. Schörghuber
A. Kern
M. Gerke
Investigating Visual Localization Using Geospatial Meshes
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
title Investigating Visual Localization Using Geospatial Meshes
title_full Investigating Visual Localization Using Geospatial Meshes
title_fullStr Investigating Visual Localization Using Geospatial Meshes
title_full_unstemmed Investigating Visual Localization Using Geospatial Meshes
title_short Investigating Visual Localization Using Geospatial Meshes
title_sort investigating visual localization using geospatial meshes
url https://isprs-archives.copernicus.org/articles/XLVIII-2-W8-2024/447/2024/isprs-archives-XLVIII-2-W8-2024-447-2024.pdf
work_keys_str_mv AT fvultaggio investigatingvisuallocalizationusinggeospatialmeshes
AT fvultaggio investigatingvisuallocalizationusinggeospatialmeshes
AT pfantajende investigatingvisuallocalizationusinggeospatialmeshes
AT mschorghuber investigatingvisuallocalizationusinggeospatialmeshes
AT akern investigatingvisuallocalizationusinggeospatialmeshes
AT mgerke investigatingvisuallocalizationusinggeospatialmeshes