TU Darmstadt / ULB / TUbiblio

Automated extraction of image coordinates for Optical Indoor Positioning

Willert, Verena and Willert, Volker and Gering, Stefan and Raß, Sebastian and Etzel, Johannes (2011):
Automated extraction of image coordinates for Optical Indoor Positioning.
In: International Conference on Indoor Positioning and Indoor Navigation, Guimaraes, 21.-23.09.2011, [Conference or Workshop Item]

Abstract

In this paper, we focus on the accuracy of optical indoor positioning based on pictures taken by a cell phone camera. We restrict ourselves to automated relative pose estimation given only one image including the projection of a known object with at least three known reference points with known absolute position. To infer the relative pose from the image coordinates of the reference points we first have to detect and classify an object, afterwards localize the image coordinates, and finally apply spatial resection. We show, that if an object is correctly classified then the quality of the positioning heavily depends on the accuracy of the localization of the image coordinates and on the choice of the spatial resection algorithm. To this end, we compare three different spatial resection algorithms and present two combinations of object classification and image coordinate localization techniques using doors as known objects. The accuracy and fault tolerance of the different methods will conclusively be discussed in terms of quality and robustness.

Item Type: Conference or Workshop Item
Erschienen: 2011
Creators: Willert, Verena and Willert, Volker and Gering, Stefan and Raß, Sebastian and Etzel, Johannes
Title: Automated extraction of image coordinates for Optical Indoor Positioning
Language: English
Abstract:

In this paper, we focus on the accuracy of optical indoor positioning based on pictures taken by a cell phone camera. We restrict ourselves to automated relative pose estimation given only one image including the projection of a known object with at least three known reference points with known absolute position. To infer the relative pose from the image coordinates of the reference points we first have to detect and classify an object, afterwards localize the image coordinates, and finally apply spatial resection. We show, that if an object is correctly classified then the quality of the positioning heavily depends on the accuracy of the localization of the image coordinates and on the choice of the spatial resection algorithm. To this end, we compare three different spatial resection algorithms and present two combinations of object classification and image coordinate localization techniques using doors as known objects. The accuracy and fault tolerance of the different methods will conclusively be discussed in terms of quality and robustness.

Divisions: 18 Department of Electrical Engineering and Information Technology > Institut für Automatisierungstechnik und Mechatronik > Control Methods and Robotics
18 Department of Electrical Engineering and Information Technology > Institut für Automatisierungstechnik und Mechatronik
18 Department of Electrical Engineering and Information Technology
Event Title: International Conference on Indoor Positioning and Indoor Navigation
Event Location: Guimaraes
Event Dates: 21.-23.09.2011
Date Deposited: 06 Oct 2011 13:16
Related URLs:
Alternative keywords:
Alternative keywordsLanguage
visual indoor positioning, spatial resection, image detection, image classificationEnglish
Export:

Optionen (nur für Redakteure)

View Item View Item