Determining 3D "contour-esque" shapes out of 3 dimension matrix of samples

I am trying to calculate 3D shapes out of a 3 dimensional matrix of samples. My idea is that I would have a 3 dimensional matrix of data points, with each corresponding location in (X, Y, Z) space, and from this I would determine a set of 3 dimensional shapes in the (X, Y, Z) s开发者_开发技巧pace that each outline data points of similar values. Something similar to finding isolines, except in three dimensions.

Can anyone help me? Algorithm references or reusable code (in any language) welcome.

Thanks in advance,

Daniel

---

You may be looking for the Marching Cubes algorithm.

---

Sounds like a job for Marching cubes. Also, see this link