Magic NeRF lens: interactive fusion of neural radiance fields for virtual facility inspection

Li, Ke and Schmidt, Susanne and Rolff, Tim and Bacher, Reinhard and Leemans, Wim and Steinicke, Frank (2024) Magic NeRF lens: interactive fusion of neural radiance fields for virtual facility inspection. Frontiers in Virtual Reality, 5. ISSN 2673-4192

[thumbnail of frvir-05-1377245.pdf] Text
frvir-05-1377245.pdf - Published Version

Download (3MB)

Abstract

Virtual reality (VR) has become an important interactive visualization tool for various industrial processes including facility inspection and maintenance. The capability of a VR application to present users with realistic simulations of complex systems and immerse them in, for example, inaccessible remote environments is often essential for using VR in real-world industrial domains. While many VR solutions have already been developed to support virtual facility inspection, previous systems provide immersive visualizations only with limited realism, because the real-world conditions of facilities are often difficult to reconstruct with accurate meshes and point clouds or typically too time-consuming to be consistently updated in computer-aided design (CAD) software toolkits. In this work, we present Magic NeRF Lens, a VR framework that supports immersive photorealistic visualizations of complex industrial facilities leveraging the recent advancement of neural radiance fields (NeRF). We introduce a data fusion technique to merge a NeRF model with the polygonal representation of it’s corresponding CAD model, which optimizes VR NeRF rendering through magic-lens-style interactions while introducing a novel industrial visualization design that can support practical tasks such as facility maintenance planning and redesign. We systematically benchmarked the performance of our framework, investigated users’ perceptions of the magic-lens-style visualization design through a visual search experiment to derive design insights, and performed an empirical evaluation of our system through expert reviews. To support further research and development of customized VR NeRF applications, the source code of the toolkit was made openly available.

Item Type: Article
Subjects: Research Scholar Guardian > Multidisciplinary
Depositing User: Unnamed user with email support@scholarguardian.com
Date Deposited: 10 Apr 2024 07:36
Last Modified: 10 Apr 2024 07:36
URI: http://science.sdpublishers.org/id/eprint/2678

Actions (login required)

View Item
View Item