Go to page
 

Bibliographic Metadata

Title
Robust 3D Position Estimation in Wide and Unconstrained Indoor Environments
AuthorMossel, Annette
Published in
Sensors, 2015, Vol. 15, Issue 12, page 31482-31524
Published2015
Edition
Published version
LanguageEnglish
Document typeJournal Article
Keywords (EN)indoor positioning / optical / wide area / virtual reality / surveying
URNurn:nbn:at:at-ubtuw:3-1190 Persistent Identifier (URN)
DOI10.3390/s151229862 
Restriction-Information
 The work is publicly available
Files
Robust 3D Position Estimation in Wide and Unconstrained Indoor Environments [17.72 mb]
Links
Reference
Classification
Abstract (English)

In this paper, a system for 3D position estimation in wide, unconstrained indoor environments is presented that employs infrared optical outside-in tracking of rigid-body targets with a stereo camera rig. To overcome limitations of state-of-the-art optical tracking systems, a pipeline for robust target identification and 3D point reconstruction has been investigated that enables camera calibration and tracking in environments with poor illumination, static and moving ambient light sources, occlusions and harsh conditions, such as fog. For evaluation, the system has been successfully applied in three different wide and unconstrained indoor environments, (1) user tracking for virtual and augmented reality applications, (2) handheld target tracking for tunneling and (3) machine guidance for mining. The results of each use case are discussed to embed the presented approach into a larger technological and application context. The experimental results demonstrate the systems capabilities to track targets up to 100 m. Comparing the proposed approach to prior art in optical tracking in terms of range coverage and accuracy, it significantly extends the available tracking range, while only requiring two cameras and providing a relative 3D point accuracy with sub-centimeter deviation up to 30 m and low-centimeter deviation up to 100 m.

Stats
The PDF-Document has been downloaded 47 times.
License
CC-BY-License (4.0)Creative Commons Attribution 4.0 International License