Sensor error modeling and compensation for range images captured by a 3D ranging system

The improvements in structured lighting-based 3D optical camera measurement systems have made the non-contact data acquisition (3D ranging) systems more applicable. A major critical issue in range image acquisition concerns the data accuracy. In this paper, factors that affect accuracy are carefully considered for a specific range image acquisition system, the four-dimensional imager system. The surface point distance errors, due to the optical distortion, field curvature and depth of field of the ranging system, have been studied. Two model-based methods, the area model (AM) and line model (LM), are proposed to model the systematic errors at different distances and orientation angles. An error lookup table is built with these models and used to reduce the systematic error of acquired data. The experimental results show that the proposed algorithm effectively improves ranging accuracy. Up to half of the systematic errors can be reduced using the AM method and almost all the systematic errors with the LM method.