A hand-held, high-resolution 3D shape measurement system using structured and unstructured illumination

摘要

This paper presents a high-resolution 3D shape acquisition method for developing hand-held 3D measurement system by projecting structured and unstructured patterns. The structured patterns consist of three-step phase-shifting fringe patterns, and the phase can be computed pixel by pixel to achieve high-resolution 3D measurement, and the phase-shifting method commonly requires many additional images to implement a phase unwrapping procedure for obtaining absolute phase value and removing the correspondence ambiguity, but it would slow down the measurement speed and limits its applications in fast 3D measurement. To address this problem, an unstructured pattern using random speckle is employed to reject wrong correspondence and promises a correct 3D result. Therefore, only four images are required to reconstruct one 3D point cloud, which is suitable for fast 3D measurement in hand-held systems. Moreover, the proposed method can be speed up by parallel computing technology for real-time data processing. The experiments verify the performance of the proposed method.