Surface extraction a fluid .

Question

At a given time-step I have a sampled a lot points from a fluid , and I want to extract the points which lie at the surface of the fluid. Does any one know a good algorithm and any available codes to do this?

I am aware of surface reconstruction, but the assumption there is that the sampled points are on/near the surface. So I guess that would not be too useful here.

Answer 1

Is finding the convex hull of the points adequate? How many points do you have?

Maybe start with a convex hull and then modify it to allow a certian amount of concavity where there are large parts of the surface without any points nearby.

Otherwise try fitting a spline or similar polynomial function to the points. You need some sort of cost metric to measure how good your fit is so that you don't bend your surface too much to reach the inner points. (Unless sharp high curvature sections are allowed for breaking waves etc?)

Is this a tank that get's moved about causing sloshing or something similar - if so there may be some physics model you can used to suggest the shape of surfaces likely to be found. The pattern of motion and speed of waves in the fluid may tell you how many waves etc you'd see.