Piazza Navona: 3D Gaussian Splatting and Photogrammetry
Vložit
- čas přidán 3. 02. 2024
- Comparison and discussion of 3D Gaussian Splat visualization vs. Photogrammetry for a complex sculptural object - The Fountain of the Four Rivers in Piazza Navona, Rome.
Very nice, I think your screen cap settings were optimized for streaming just so you know, got a bit blocky even at 4k when u orbit around things. I usually use Indistinguishable and MKV for output settings in OBS.
Ooooh good point - I don't have much experience with OBS, I'll look into this. thanks!
Nice stuff. Did you manage to get the 2000 images in one splat? If i try to use more than about 300 Images my system will kind of explode? Which tool are you using to show the pointcloud in the video? Thanks in advance.
Yes all 2k processed in one go, though I did downscale them to prevent running out of VRAM (40GB A100). The point cloud is just a screen-capture of the model in "Capturing Reality" (a photogrammetry software).
Awesome stuff as usual.
I'd love to see some of my favorite places point cloud embedded in acrylic by laser etching perhaps?
Embedded in glass?
I've seen some cool examples of the glass/crystal interior laser-etching.... it seems to work well for objects - would be curious to see how a space looks.
I wonder how far one could get with enhanced texturing of the ground based on satellite and photosphere images from Google maps (like in MSFS 2020).
Cuz otherwise you'd need to get approval to conduct a drone flight over the historical site to get that extra imagery and detail reference.
I’ve thought about screenshot-ing Google earth to create the rooftops/fill architecture, but the quality just isn’t high enough- I think what you’d want to do is get some reference materials for the paving/ground and model that by hand, then PBR texture it