Readers are likely to be familiar with photogrammetry, a method of creating 3D geometry from a series of 2D photographs taken of an object or scene. To pull this off you need a lot of pictures, hundreds or thousands, all taken from slightly different perspectives. Unfortunately technology suffers where there are significant obstructions due to overlapping elements, and glossy or reflective surfaces that appear different colors in each photo can also cause problems.
But new research from NVIDIA Marries photogrammetry with artificial intelligence To create what the developers are calling the Instant Neural Radiation Field (NERF). Not only does their method require far fewer images, according to NVIDIA, less than a few dozen, but the AI is better able to deal with the pain points of traditional photogrammetry; Filling in the gaps in blocked areas and taking advantage of reflections to create more realistic 3D scenes that reconstruct what shiny materials looked like in their original environments.
If your machine has a CUDA-compatible NVIDIA graphics card, you can give the technology a shot now. Tutorial video after the break Will walk you through the setup and some basics, showing how the 3D reconstruction is progressively refined over just a few minutes and then can be searched in a game engine like a scene. The Instant-NERF tool includes camera-path keyframing to export animations with higher quality results than real-time previews. This technique seems to be better suited for outputting scenes and animations than models for 3D printing, although both are possible.
Don't have the latest and greatest Nvidia Silicon? Don't worry, you can still create some impressive 3D scans using "old school" photogrammetry— All you really need is a camera and a motorized turntable,
0 Comments