Game engines, photogrammetry and deep learning for Antarctic heritage visualization: 2020 Work-in-progress
Morse, P and Staal, T and Reading, A, Game engines, photogrammetry and deep learning for Antarctic heritage visualization: 2020 Work-in-progress, Scientific Committee on Antarctic Research (SCAR) Open Science Conference 2020: Antarctic - Global Connections, 3-7 August 2020, Online, pp. 1-10. (2020) [Non Refereed Conference Paper]
Antarctic Heritage presents unrivalled opportunities for contemporary computational visualization
techniques. These range from compelling immersive heritage experiences for the general public,
through to the more exacting development of accurate digital archives for scholarly use.
Game engines have a wide variety of heritage applications as development environments for
computational humanities, digital museology and GLAM-sector applications. Reconstruction of
historic Antarctic sites using satellite and other geophysical data in concert with photogrammetric
scene reconstruction enable the construction of physically accurate heritage site models. These can
be displayed as immersive screen experiences (e.g. VR, Augmented Reality and Dome
environments) and afford novel visual analytics approaches to Antarctic heritage data. Associated
historical textual, map, photographic and film materials can be restored, animated, translated into
3D scenes and actors, and colourised using machine learning techniques (‘Deep Learning’)
employed in the film, special effects and games industries.
Immersive interactive simulations that embed historic materials demonstrate new ways of interacting
with museum collections and scientific archives, new digital methodologies of historical scholarship
and effective ways of exposing fragile archival materials for general and specialist audiences.
Interactive post-cinematic narratives suggest novel opportunities for dramatising the experience of
significant artefacts, bringing place, biography, history and science alive. Remote environments,
both in space and time, become far more accessible and available to contemporary enquiry.
A demonstration model of the Mawson’s Huts Historic Site will be presented, using a computer