3

I'm trying to think of the best way to conduct some sort of analysis between two 3D models of the same object.

The first scan is of the original item and the second scan is after it has been put under some load x.

An example would be trying to find the difference between two types of metal.

I would like to be able to scan the initial metal cylinder, apply a measured load, scan it again, and then finally apply some sort of algorithm to compare the difference.

Is it possible to do this efficiently (maybe using Mablab) over say 50 - 100 items for an object around 5inch^3?

I am assuming I will need to work out some sort of utility function as the total mass should be the same?

Would machine learning be beneficial in this case?

Any suggestions or direction would be amazing.

Thank you :)

EDIT: The scan files are coming through as '.stl'

Spektre
  • 49,595
  • 11
  • 110
  • 380
Procrastix
  • 43
  • 4
  • 1
    Congrats on a very interesting first question. I don't think I will be the right person to answer it. But I think the question could benefit if you explained a little bit more what you imagine the scan data will look like. 3D rotational? Stereoscopic? 2Dimage etc. – visibleman Nov 08 '17 at 06:52
  • Thank you @visibleman :) I believe it is coming through as a .stl file. – Procrastix Nov 08 '17 at 07:24
  • You should 1. align the meshes so they got the same orientation and position in space. 2. compute their substraction (either as voxel map or as polygonal mesh). That should reveal/emphasize any changes. If you got some sort of specific shape you could use its analytical representation (fit it from each mesh and compare the parameters directly) For example if you got cylinders you can fit them from mesh (convert them to slices perpendicular to its axis which can be curve) and can compare the axis distortion as well as surface shape distortion per slice. Without knowing more is hard to say more – Spektre Nov 08 '17 at 07:46
  • Another option is to map the surface vertexes between both meshes (resample one to match the other) and than you can do per vertex comparison and displacement analysis (or even stress estimation). btw why machine learning, image processing and statistic tags? – Spektre Nov 08 '17 at 07:48
  • Thank you Spektre! I didn't entirely know which one this problem would fall under? – Procrastix Nov 08 '17 at 11:15
  • When you analytical representation, would you say that would be the initial scan? If so, will these be difficult to line up in software or will it be fine assuming I scan it in the exact orientation both times? – Procrastix Nov 08 '17 at 11:19
  • 1
    @DanielGormly that depends on the shape and circumstances of the scan like noise, accuracy, target precision, the difference itself etc. Without any specifics we can only guess. Also if you want to reply to specific user (`nick`) present in current question thread Add `@nick` to your comment so the site will automatically notify him or she or it. Author of the thread is notified automatically (but us all others not). – Spektre Nov 10 '17 at 07:03

0 Answers0