I am writing a function that extrudes a 2D shape along a 3D spline, as found in 3D modeling software.

I need a way of translating a set of points P so that they all lie in a new plane L (preserving distances between points).

  • P is all points where Z = 0;
  • L is defined by (0,0,0), with the normal vector N.

I am using Processing, it has common matrix and vector functionality (but no quaternions) A conceptual or pseudo-code answer is OK.

  • \$\begingroup\$ I just discovered a library for processing that includes quaternion rotation and euler angles. I've been trying to do something similar but looks like I should use this library. github.com/remixlab/proscene \$\endgroup\$
    – Andy Mac
    Feb 13 '15 at 7:10

I need a way of translating a set of points P

I suppose you mean rotating here?

Let Z = (0,0,1). If cross(N,Z) has length 0, it means that all your points already lie in the desired plane. Otherwise we can build a basis of the target plane:

  • U = normalize(cross(N,Z))
  • V = cross(N,U)

Now to transform a point P = (x,y,0) so that it lies in your target plane, simply do this:

P2 = x * U + y * V
  • \$\begingroup\$ does the line V = cross(N,V) contain a typo? V is not defined? Did you mean U? \$\endgroup\$
    – Andy Mac
    Feb 17 '15 at 3:41
  • 1
    \$\begingroup\$ oops, yes, it should be U! Fixed. \$\endgroup\$ Feb 17 '15 at 7:00

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.