Various knots I created based on 3D scans with my Artec Leo. The scans were used as a guide for remeshing. The meshes were simulated with Kangaroo. The shoelace knot was the culmination so far because there are both self intersections to avoid aswell as collisions in between the two loops.
Beautiful. Thanks for posting these. Does the scanner give you most of the info to build the meshes. Just watched a video on Kangaroo. Would love to see how you used it to manipulate these drawings.
The scanner only scans what I can see. I started with scanning a piece of a climbing rope. The first challenge was the positioning of the knot so it did not change when I had to move it to see it from all sides. I used needles and they are thin enough the scanner doesn’t see them.
On the screenshot below you can see the points I was able to capture. Blackcolor shows the backfaces, so basically whatever is black are holes which I couldn’t capture.
Here’s a fusion of a total of 7 separate scans:
And a view from another side, showing the bridging where the rope is too close to itself. This could be improved but I decided to rebuild in Rhino so I didn’t worry too much about this bridge problem.
The process of finding the “axis” or “centerline” of the rope is an iteration in GH, where I start with a manually created curve as close to the centerline as possible. Then I can create sections and use the right sections to improve the centerline. It gets complicated when the rope loops and is close to itself.
It was an interesting investigation. First of all to test the new high definition mode on the scanner. The results are mainly to showcase my work but at least I could use the technique for a few renders…
Excellent work Martin. The 3D scanning as a 3D underlay to drive your intent is a great idea.
Why not use Knot Theory to achieve this if it is not a stupid question?
I’m aware there are programs or plugins to create 3D shoelaces. Not sure what Knot Theory is and what it can do.
Anyway, there are a few reasons I’m not using off the shelf solutions:
- I’m stubborn and I like to explore my world.
- I wanted to see whether I can solve the problem with Kangaroo.
- I own an Artec Leo and the new HD mode allows to capture much more detail.
Actually it started with that shoe scan where I tested the HD mode. I was very excited with the shoe laces since I was able to scan almost everything without holes. But then my friend who’s obsessed with rebuilding all sorts of things with quad meshes said the shoe laces are not good enough for his needs, mainly because it doesn’t show exactly what happens inside the knot. I couldn’t just give up and I wanted to see how I can generate a better solution. The first try was with the round climbing rope. Then I also wanted to simulate flat webbing…
Knot theory in a nutshell should be not too hard in GH (I didn’t search*)…
Although i haven’t figured out it could work…
It’s been done… http://katlas.org/wiki/Gauss_Codes
Im avoiding coding as much as possible so im not going to attempt it.
That scanner rocks! A bit out of my budget however…
Thanks, I actually already visited that mathartblog website
One of the first things I tried to parametrize in GH back around 2011 was a trefoil…
The mathematical approach is interesting. It leads to clean results, sometimes too clean. The goal with the scanner is to be able to play with irregular things, shaped by nature, or by hands. Otherwise the scanning wouldn’t be necessary.
This is BADA$$. @martinsiegrist are you working on developing a LIBRARY OF KNOTS & TIES!?
Na, the scans were mainly just prove of concept, no particular use…
I wante to test the capability of the new HD mode on my Artec Leo scanner and try various ways to reconstruct the knots.