Photometric Stereo  (Estimating 3D Shape from Shading)

  • Aside : How to compute surface normals for parametric surfaces.
  • The surface reconstruction setup and assumptions.
  • The photometric stereo reconstruction equations and how to solve them.
  • How to reduce the influence of shadows in photometric surface reconstruction.
  • Recovering normals and surface albedo from a set of N images.
  • The differential geometry of Monge Patches.
  • Estimating the difference images (x-partial-differential and y-partial-differential) from the components of the estimated normals.
  • Checking the solution for consistency, i.e., checking the mixed second partial derivatives.
  • Reconstructing surface shape by numerical contour integration.
Project : Implement a theoretically ideal photometric stereo system in MATLAB.

 

Using range sensors to obtain 3D depth of object surfaces :

  • Active range sensing and examples
  • Passive range sensing and examples

 Active Triangulation

  • Problem setup : Reconstructing depth by projecting a laser stripe onto a 3D object.
  • Derivation of the sensor model.
  • Equations to be solved and their solution.
Project : Implement a theoretically ideal active triangulation system in MATLAB.