Photogeometric Structured Light: A Self-Calibrating and Multi-Viewpoint Framework for Accurate 3D Modeling

Daniel G. Aliaga,Yi Xu

DOI: https://doi.org/10.1109/cvpr.2008.4587709

2008-01-01

Abstract:Structured-light methods actively generate geometric correspondence data between projectors and cameras in order to facilitate robust 3D reconstruction. In this paper, we present photogeometric structured light whereby a standard structured light method is extended to include photometric methods. Photometric processing serves the double purpose of increasing the amount of recovered surface detail and of enabling the structured-light setup to be robustly self-calibrated. Further, our framework uses a photogeometric optimization that supports the simultaneous use of multiple cameras and projectors and yields a single and accurate multi-view 3D model which best complies with photometric and geometric data.