After using my phone for a year as a scanner, I was getting acceptable but not spectacular scans of complex scenes, such as a server racks, plants and very small objects. Often, the wires/stems would simply get erased, or the switches would merge together from algorithms that Polycam uses, and the inaccuracies of photogrammetry.

I am still learning how to use this, since it’s quite a different process. I have been practicing capturing details, and I’m soon going to work on a refreshed model to use in my Creating a 3D, Interactive, Real-Time, Network Rack Model project.

Models made with this can easily balloon to immense size. For example, I have captured a small house plant and the resulting mesh was 6 million polygons. This clearly won’t run well on most computers, so the models often need post processing to simplify/smooth/solidify/etc the model so its easier to handle.