grx0542@uoft02.utoledo.edu (03/20/90)
Does anyone have any leads or papers on fitting cylinders to arbitrarily spaced 3d data - there is no order to the data, just a cluster of data points. These are not circular scans about the axis of the cylinder, but scattered points on the surface. These points may also be clustered on one side of the cylinder. I am right now working on finding the direction cosines of the axis using least squares approximations but am having trouble formulating the objective function. I assume I would have to use an optimization technique like pattern search or graident search methods. Any suggestions would be appreciated. Siva Balakumar (grx0542@uoft02.utoledo.edu) CAEDM Center, SE 3090 U of Toledo, Toledo Oh 43606