3D Localization in Urban Environments from Single Images

Anil Armagan, Martin Hirzer, Peter M. Roth, Vincent Lepetit

Research output: Chapter in Book/Report/Conference proceedingConference contributionResearchpeer-review

Abstract

In this paper, we tackle the problem of geolocalization in urban environments overcoming the limitations in terms of accuracy of sensors like GPS, compass and accelerometer. For that purpose, we adopt recent findings in image segmentation and machine learning and combine them with the valuable information given by 2.5D maps of buildings. In particular, we first extract the façades of buildings and their edges and use this information to estimate the orientation and location that best align an input image to a 3D rendering of the given 2.5D map. As this step builds on a learned semantic segmentation procedure, rich training data is required. Thus, we also discuss how the required training data can be efficiently generated via a 3D tracking system.
Original languageEnglish
Title of host publicationProceedings of the OAGM/AAPR & ARW Joint Workshop (OAGM/AAPR & ARW)
Publication statusPublished - 2017
EventOAGM/AAPR & ARW Joint Workshop 2017 - Wien, Austria
Duration: 10 May 201712 May 2017

Conference

ConferenceOAGM/AAPR & ARW Joint Workshop 2017
Abbreviated titleOAGM/AAPR ARW 2017
CountryAustria
CityWien
Period10/05/1712/05/17

Fingerprint

Information use
Image segmentation
Accelerometers
Learning systems
Global positioning system
Semantics
Sensors
Rendering (computer graphics)

Cite this

Armagan, A., Hirzer, M., Roth, P. M., & Lepetit, V. (2017). 3D Localization in Urban Environments from Single Images. In Proceedings of the OAGM/AAPR & ARW Joint Workshop (OAGM/AAPR & ARW)

3D Localization in Urban Environments from Single Images. / Armagan, Anil; Hirzer, Martin; Roth, Peter M.; Lepetit, Vincent.

Proceedings of the OAGM/AAPR & ARW Joint Workshop (OAGM/AAPR & ARW). 2017.

Research output: Chapter in Book/Report/Conference proceedingConference contributionResearchpeer-review

Armagan, A, Hirzer, M, Roth, PM & Lepetit, V 2017, 3D Localization in Urban Environments from Single Images. in Proceedings of the OAGM/AAPR & ARW Joint Workshop (OAGM/AAPR & ARW). OAGM/AAPR & ARW Joint Workshop 2017, Wien, Austria, 10/05/17.
Armagan A, Hirzer M, Roth PM, Lepetit V. 3D Localization in Urban Environments from Single Images. In Proceedings of the OAGM/AAPR & ARW Joint Workshop (OAGM/AAPR & ARW). 2017
Armagan, Anil ; Hirzer, Martin ; Roth, Peter M. ; Lepetit, Vincent. / 3D Localization in Urban Environments from Single Images. Proceedings of the OAGM/AAPR & ARW Joint Workshop (OAGM/AAPR & ARW). 2017.
@inproceedings{45d150df9bd8427ab935c0e8bf2b398e,
title = "3D Localization in Urban Environments from Single Images",
abstract = "In this paper, we tackle the problem of geolocalization in urban environments overcoming the limitations in terms of accuracy of sensors like GPS, compass and accelerometer. For that purpose, we adopt recent findings in image segmentation and machine learning and combine them with the valuable information given by 2.5D maps of buildings. In particular, we first extract the fa{\cc}ades of buildings and their edges and use this information to estimate the orientation and location that best align an input image to a 3D rendering of the given 2.5D map. As this step builds on a learned semantic segmentation procedure, rich training data is required. Thus, we also discuss how the required training data can be efficiently generated via a 3D tracking system.",
author = "Anil Armagan and Martin Hirzer and Roth, {Peter M.} and Vincent Lepetit",
year = "2017",
language = "English",
booktitle = "Proceedings of the OAGM/AAPR & ARW Joint Workshop (OAGM/AAPR & ARW)",

}

TY - GEN

T1 - 3D Localization in Urban Environments from Single Images

AU - Armagan, Anil

AU - Hirzer, Martin

AU - Roth, Peter M.

AU - Lepetit, Vincent

PY - 2017

Y1 - 2017

N2 - In this paper, we tackle the problem of geolocalization in urban environments overcoming the limitations in terms of accuracy of sensors like GPS, compass and accelerometer. For that purpose, we adopt recent findings in image segmentation and machine learning and combine them with the valuable information given by 2.5D maps of buildings. In particular, we first extract the façades of buildings and their edges and use this information to estimate the orientation and location that best align an input image to a 3D rendering of the given 2.5D map. As this step builds on a learned semantic segmentation procedure, rich training data is required. Thus, we also discuss how the required training data can be efficiently generated via a 3D tracking system.

AB - In this paper, we tackle the problem of geolocalization in urban environments overcoming the limitations in terms of accuracy of sensors like GPS, compass and accelerometer. For that purpose, we adopt recent findings in image segmentation and machine learning and combine them with the valuable information given by 2.5D maps of buildings. In particular, we first extract the façades of buildings and their edges and use this information to estimate the orientation and location that best align an input image to a 3D rendering of the given 2.5D map. As this step builds on a learned semantic segmentation procedure, rich training data is required. Thus, we also discuss how the required training data can be efficiently generated via a 3D tracking system.

M3 - Conference contribution

BT - Proceedings of the OAGM/AAPR & ARW Joint Workshop (OAGM/AAPR & ARW)

ER -