HOTA Reconstructed.
Reality Scan image processing & alignment
Radiance field trained in PostShot
The project is a proof of concept for what Gaussian splatting makes possible: capturing a real place in high fidelity while keeping it lightweight, interactive and accessible online. Drone imagery was captured with a DJI Mini 5 Pro, aligned in RealityScan, then trained across multiple iterations in PostShot to balance visual quality against file size. The resulting radiance field was framed, scaled and cleaned in SuperSplat, then built into a custom flythrough with interactable points of interest.
Cleaning and animation in SuperSplat editor
Optimised Gaussian Splat asset
outcomes.
I chose HOTA for its visual richness and its connection to the fields of art and design, but the technique reaches well beyond a single tour. The same pipeline can preserve heritage sites, drive high-fidelity real estate and cultural experiences, and feed the immersive AR and VR spaces that are quickly becoming a real demand. This project is where I am taking my practice: capturing real places with emerging technology and making them explorable by anyone, anywhere.
