Real-time high dynamic range 3D measurement using fringe projection.

ABSTRACT

Fringe projection profilometry (FPP) is a widely used technique for real-time three-dimensional (3D) shape measurement. However, it tends to compromise when measuring objects that have a large variation range of surface reflectivity. In this paper, we present a FPP method that can increase the dynamic range for real-time 3D measurements. First, binary fringe patterns are projected to generate grayscale sinusoidal patterns with the defocusing technique. Each pattern is then captured twice with different exposure values in one projection period. With image fusion, surfaces under appropriate exposure are retained. To improve the real-time performance of high dynamic range (HDR) 3D shape measurements, we build a binocular fringe projection profilometry system that saves the number of patterns by geometry constraint. Further, to ensure the accuracy and robustness of HDR 3D measurements, we propose a mixed phase unwrapping method that can reduce phase unwrapping errors for dense fringe patterns. Experiment results show that the proposed method can realize accurate and real-time 3D measurement for HDR scenes at 28 frames per second.