Journal of Digital Landscape Architecture

Seite drucken

Neural Radiance Fields for Landscape Architecture

In this paper, we examine potential applications of Neural Radiance Fields (NeRF) in the field of landscape architecture. NeRF is a state-of-the-art method for novel view synthesis and volumetric scene reconstruction based on real-world training data. Our paper addresses NeRF and its derived models with a focus on the use and application of Instant-NGP, a method developed by researchers from the technology company NVIDIA. We discuss experimental applications of NeRF based on the case study of the post-disaster landscape of Ahr Valley, Germany, affected by a 100-year flood in 2021. In particular, we are interested in the benefits of NeRF in comparison to other landscape modeling methods, such as Structure-from-Motion (SfM) or Multi-View-Stereo (MVS), which use similar data as input. This study shows that the application of NeRF technology can be a promising alternative for capturing and visualizing landscape scenes. The study focuses especially on tasks and situations where the larger spatial context – the landscape – is of interest and importance. The technological aspects of how NeRF models work are relevant, but our main focus is on their potential implications for the field of landscape architecture. Technical development and research in the scientific field of computer vision are accelerating rapidly. As users, rather than developers, of digital tools, we believe that NeRF technology requires professional validation through real-world landscape projects.

Autor / Author: Schob, Maximilian; Rekittke, Jörg
Institution / Institution: Norwegian University of Life Sciences (NMBU)/Norway; Norwegian University of Life Sciences (NMBU)/Norway
Seitenzahl / Pages: 15
Sprache / Language: Englisch
Veröffentlichung / Publication: JoDLA – Journal of Digital Landscape Architecture, 8-2023
Tagung / Conference: Digital Landscape Architecture 2023 – Future Resilient Landscapes
Veranstaltungsort, -datum / Venue, Date: Dessau Campus of Anhalt University, Germany 24-05-23 - 27-05-23
Schlüsselwörter (de):
Keywords (en): Neural Radiance Field, NeRF, Novel View Synthesis, Instant-NGP (Instant Neural Graphics Primitives), Ahr Valley Flood 2021
Paper review type: Full Paper Review
DOI: doi:10.14627/537740046
7733 -