High Dynamic Range Imagine Algorithm Based 3D Profilometry

As one of the usual methods of automated optic inspection, 3D profilometry has a large application filed, so there has many researches on this field. In the past, many researches focused on the systematic algorithm of 3D profilometry, including system calibration, phase value calculation and phase unwrapping, but the inspected objects are restricted only on non-specular reflection objects. Because the metallic objects which will cause specular reflection compose large amount of industrial inspected object, so this paper is focused on metallic objects'3D profilometry, and introduce a new algorithm to overcome the disadvantage of 3D profilometry in measuring metallic objects.Traditional 3D profilometry has several steps including 3D system calibration, phase deformed image capture, phase value calculation and phase unwrapping, in these steps phase calculation is the key step which has final relation to 3D information's accuracy, and the calculation of phase value uses a series of phase deformed images with different phase shift. Because of the dynamic range limitation of traditional camera, traditional phase deformed image will be over-saturated in the affect of specular reflection, so the phase information will be lost in this region, and render the 3D information recovery in this region be lost.In this paper, we use high dynamic range imaging algorithm to combine into 3D profilometry to overcome the affect of specular reflection. High dynamic range imaging algorithm use a series of images with different exposure time to recovery the scene's true irradiance, and use this irradiance to replace pixel value in phase calculation, which will effectively avoid the phase value errors that is introduced by specluar reflection and recover the ground-truth 3D information of object surface.