AUTOMATED INSPECTION OF SPECULAR SURFACES BY MEANS OF PATTERN REFLECTIONS

A technical realization to inspect specular surfaces based on an imita tion of the human visual inspection process is presented. The human ob server does not look at the surface itself but at the environment refl ected by the surface and analyzes the distortions. To automate this vi sual approach and to extend it to a quantitative measurement method, a CCD-camera looks at the reflected image of an illuminated, programmab le LC-Display. With a sequence of patterns on the LCD and by means of inverse ray tracing the surface relief can be reconstructed. In contra st to the method of fringe projection, which is locally sensitive main ly to the absolute topography, the reflexion method is particularly se nsitive to the gradient of the surface.