0

I have a 3D point cloud that I attained from tracing out the outline of a shape with sensors attached to my fingertips. The resulting data has non-uniform density with large gaps between some of the points.

What are some good surface reconstruction algorithms to use on this kind of data that is recorded by hand and therefore has issues of varying density?

I have been attempting to use the Cocone, Robust Cocone, and Tight Cocone Surface Reconstruction algorithms from Tamal Dey to reconstruct the shape, but I am having difficulty because I believe my data is much less uniform than the example point sets provided with the algorithms. I have read Tamal's literature on each reconstruction algorithm because there are variables that can be altered in the algorithms, but I have been unable to find the right settings to get my data to work with any of the Cocone algorithms.

Does anyone understand the user settings in these algorithms?

What would be the best settings for very non-uniform data points? I can provide the 3D point data of the shape upon request.

Nikolay Kostov
  • 16,433
  • 23
  • 85
  • 123
Mike Petersen
  • 1
  • 2
  • I have implemented several variants of the Co-cone algorithms (with various ways of estimating the normals, it is mostly what makes the difference). If you send the data, I can try them on it and tell you which strategy works (if one of them works ....) – BrunoLevy Jul 24 '15 at 14:55
  • Thank you for your response BrunoLevy, much appreciated. I have experimented with the settings on my own and discovered the default settings worked well enough for my data. I also played around with several filtering and smoothing algorithms to help with the surface reconstruction which really helped to achieve an accurate shape. – Mike Petersen Jul 27 '15 at 14:23

0 Answers0