08-13-2008 01:41 AM
08-13-2008 01:58 AM
08-13-2008 03:13 AM
It is a paperboard. I try to make an automatic surface inspection with the photometric stereo.
(I am searching for 3D defects on the surface)
For this i have to illuminate the board from 4 directions (north, south, west and east) and make a picture from the object while holding the viewing geometry constant.
With the image fusion of the north and the south images and the fusion of west and east i have the possibility to calculate the 3D surface.
But for this it is important that the pictures have nearly the same and a constant brightness.
So after the image correction there should be only slight surface structures visible and not the illumination gradient.
08-13-2008 04:07 AM
I am searching for 3D defects on the surface
Did you try low angle lighting techniques using bar leds
08-13-2008 06:52 AM
yes,
sure i can still optimize the illumination, but this will not be enough.
i must calculate a calibration factor for each pixel
08-13-2008 06:59 AM