View Full Version : Fitting a sphere

01-07-2018, 11:21 AM

Any suggestion for sphere recognition on point clouds?


01-07-2018, 07:02 PM
I'm just taking a guess on this, so forgive me if I misunderstood the question. I think you could take a sample of points from your point cloud and test each point to see if they have the characteristics of being points from a sphere. So take a sample of points and find the center of the sphere using some math. Then you test several more points to see if their distance to the center of the sphere is close to being the radius of the sphere(there will be some margin for error). The more points the better chance of being sure its a sphere.