An adaptive structured light pattern for the 3D profiling of colored objects

Coded structured light is a fast and reliable technique to acquire the geometry of a measuring surface since the pattern codification might solve the correspondence problem in a single shot. Colored patterns reduce pattern segmentation complexity but their efficiency decreases in the presence of colored surfaces. We propose a color coded pattern which is automatically adapted depending on the color of the measured surface, which is a topic not completely solved in literature. In 1, Caspi showed that the projection of colored patterns on a saturated object is possible to get the 3D geometry of the surface but the surface color was not acquired and so the pattern lacks of an automatically adaptation. Hence, an accurate color calibration of the structured light system is required, which has been based on 2: Nayar's radiometric calibration computes the non-linear relationships between camera and projector. Then, the system identifies the color of the surface and the projected colors might change with the aim of making easy the pattern segmentation and the labeling.