A Phase Compensation Method for Surface Reconstruction of Moving Object

Abstract

Recently, the surface reconstruction of moving objects based on phase shifting profilometry (PSP) becomes increasingly important in the academic and industrial fields. Accurate phase of fringe patterns on captured images should be compiled to yield high quality reconstruction results. However, it is still hard to ensure the quality of phase map when the object is moving. Besides, moving object can introduce phase error to PSP, which will lead to reconstruction errors, for instance, waviness errors on the surface. The existing methods of reducing phase error in dynamic scene have certain disadvantages. Such as highspeed photography that requires enormous expenses and advanced hardware, phase calibration which is inexpensive but computationally intensive. This paper introduces a phase compensation method, which strikes a balance between computational complexity and reconstruction accuracy. Experiments demonstrate that, the proposed method can significantly reduce computational complexity comparing with current method, while achieving adequate accuracy.