A 3D vision system using unique color encoding

Abstract

For an active robot vision system, the 3D sensor is often moved to several viewpoints around the object for acquisition of the features of interest. In such a case, its configuration needs to be changed for better perception of the different features. This paper introduces the 3D vision system and investigates the self-recalibration problem in a structured light system with six degrees of freedom in which the sensing configuration is changeable. The relative pose between the projector and camera of the system is calibrated by taking a single view of the scene, so that the 3D measurements and reconstruction can be performed immediately when and if the configuration of the system is changed. A unique color encoding method is used so that only one image is required for the 3D reconstruction.