Week 4: 3D Scanning a Rubber Duck


This is a bonus page for the second part of week 4's assignment: scanning an object using a 3D scanner. As I had already spent most of my time this week modeling and 3D printing with a 2-material printer, I decided to go with something small for my 3D scanning. I used a rubber duck as an example object to scan, since it had some nice curvature and I was wondering how smooth the surface of the 3D object would turn out and if I could use it effectively in Blender. I also wanted to see if the rubbery material had any effect on the scan quality.

Scanning the Rubber Duck

I wanted to use the Revopoint 3D scanner we have in our fab lab to scan the rubber duck. Obviously, I needed to scan it from all angles to get a 3D representation of it. However, I can not hold it in my hand while doing so, as otherwise, I would scan my hand as well. A number of possibilities exist to avoid this issue, but they all follow the same basic idea of either rotating the laser scanner by itself or the object by itself in order to cover all 360°. Because the rubber duck was relatively small and lightweight, I decided to use a powered lazy susan we had that came with the 3D scanner and put the rubber duck in the middle.

I then started the Revopoint 3D scanning software on our iMac and previewed the current exposure. The preview screen is extremely useful and easy to work with, as it gives a live representation of the particles currently seen by the camera, a live video feed, and a distance map. The auto settings for the exposure time were already fairly well adjusted – although I did play with them, I could not get any better results. So I started my first 3D scan. Unfortunately, the rubber duck was placed a bit off-center on the lazy susan, which created an erronous 3D scan. This is because the camera tries to map these particles to locations on the back side of the 3D object that do not match their actual position in the real world, so the calculated construct collapses into a weird pile of vericies. This problem was easy to fix though.

After I placed the rubber duck perfectly at the center of the lazy susan, I took another 3D scan of it, and this time, the 3D object came out perfectly. The surface was surprisingly smooth and I could not find any hard lines / seams or buggy verticies, so the scan succeeded and I could have used the object in Blender or another modelling software to continue modeling it. Overall, I am happy with the result and genuinly surprised how well it worked!