Shape measurement by inverse raytracing

摘要

The actual 3D shape of free formed transparent objects is of interest in different fields of research and production. Measuring shapes just by looking through an object is a challenging task for any hard- and software concepts. We report a stereo imaging system that computes the topography of the surface and the refractive index of freeform lenses. The system takes a set of images from different viewpoints of various patterns in a fix environment. Based on this 2D images, an optic design model is built up with the sample as unknown optical element. Then, the shape of the sample will be modified until a cost function is minimized. In a first approach the sample is described as a spherical lens. In this case two radii of curvature and a refractive index of the material can be determined. In a more general case, we discuss an approach to calculate the full topography of one surface of the sample from the recorded data. Hereby, an iterative method to calculate the topography will be demonstrate. To qualify the performance of this principle, different free formed optics were measured and the deviation to reference measurements will be shown. This work shows the advantages of combining methods from different disciplines as optical engineering, optics design and computer vision.