How to smooth a surface generated from interpolation on nearly random data points?

I am creating surfaces from large point clouds and one of the MATLAB functions I have been working with is the surf command. The only problem with it is that it creates a surface using every single point in my data set and the generated surface is very jagged. If there was a way to remove the inner points and keep only the outer layer of points, I could create a smooth surface that would better represent the point cloud shape volume.

Any suggestions, and in advance, thank you for your efforts to help me with my problem.

Answer 1

[Image Analyst]

Is this a 2D cloud (scatterplot) and surface, or a 3D cloud and 3D surface (a closed volume)? Maybe you might want to look into alpha shapes http://cgm.cs.mcgill.ca/~godfried/teaching/projects97/belair/alpha.html and related things like snakes, balloons, active contours, etc., or else just do a regression fit instead of an interpolation.

[Luke]

Thanks, I took at look at your recommendation and it sparked an idea to switch from interpolation to triangulations and it all went downhill from there.

Answer 2

[Walter Roberson]

No can do. Not unless you want to confine the surface to being a convex hull, or are otherwise able to state explicit criteria for how to determine whether a point should be discarded or not.