Image-based localization for an indoor VR/AR construction training system

Ludovico Carozza, Frédéric Bosché, Mohamed Abdel-Wahab

    Research output: Contribution to conferencePaperpeer-review

    188 Downloads (Pure)

    Abstract

    Virtual /Augmented Reality (VR/AR) technologies have been increasingly used in recent years to support different areas of the construction industry. Its simulation capabilities can enable different construction stakeholders to evaluate the impact of their choices not only on the built environment, but also with regard to the correct execution of operational procedures. Training providers, such as Further Education (FE) colleges, can also enhance their trainee’s experience through the simulation of realistic construction contexts whilst eliminating health and safety risks. Current approaches for the simulation of learning environments in Construction, such as Virtual Learning Environment (VLEs), provide limited degree of interactivity during the execution of real working tasks. Whilst immersive approaches (e.g. CAVE-based) can provide enhanced visualization of simulated environments, they require complex and expensive set-up with limited practical interaction in real construction projects context.
    This paper outlines a localization approach employed in the development of an Immersive Environment (IE) for Construction training, cheaper than CAVE-based approaches and with the potential to be rolled-out widely across the FE sector for maximizing the benefit to the construction industry. Pose estimation of the trainee is achieved by processing images acquired by a monocular camera integral with his head while performing tasks in a virtual construction environment. Realistic perception of the working environment and its potentially hazardous conditions can thus be consistently delivered to the trainee through immersive display devices (e.g. goggles). Preliminary performance of the localization approach is reported in the context of working at heights (which has a wide applicability to a range of construction trades, such as scaffolders and roofers), whilst highlighting the potential benefits for trainees. Current limitations of the localization approach are also discussed suggesting directions for future development.
    Original languageEnglish
    Pages363-372
    Publication statusPublished - 2013
    Event13th International Conference on Construction Applications of Virtual Reality - London, United Kingdom
    Duration: 30 Oct 201331 Oct 2013

    Conference

    Conference13th International Conference on Construction Applications of Virtual Reality
    Abbreviated titleCONVR 2013
    Country/TerritoryUnited Kingdom
    CityLondon
    Period30/10/1331/10/13

    Fingerprint

    Dive into the research topics of 'Image-based localization for an indoor VR/AR construction training system'. Together they form a unique fingerprint.

    Cite this