point cloud data from machine vision

[abcds]

I would like to know if i can use the CCD camera to generate point cloud data from a surface to use it in reverse engineering applications
The process of reverse engineering for a surface consists of geenration of points by the CCD camera, segmentation, then using some algorithms in which can create the reconstructed surface then transfer it into CAD software (IGES)

[nestor-sg]

I believe this process will get complicated unless someone else might have some algorithms or suggestions. The Vision toolkit has none built-in functions for this purpose, however if you have your own idea and have questions on the Vision functions please let us know. The best example is a sphere, in which with lightning you could easily say that according to the shade it is translated to the surface depth, so with the intensity values of each pixel you could build kind of a surface mesh 2D array for your CAD software. The problem is that most likely shades are not going to be directly applied to your objects so you will need to look for other information as edges, intensity changes, frequency analysis, etc.

Some of the Digital Image Processing books will touch this sub
ject, 3D analysis, you may want to reference any of those books to see what type of pixel analysis they are doing. We have a small list of the DIP books that we are aware of in our Reference Books section at www.ni.com.

Hopefully someone will give you a better guidance. Good luck!

Nestor.