High-precision Real-Time 3D Shape Measurement Based on a Quad-Camera System
Tianyang Tao,Qian Chen,Shijie Feng,Yan Hu,Minliang Zhang,Chao Zuo
DOI: https://doi.org/10.1088/2040-8986/aa9e0f
IF: 2.1
2017-01-01
Journal of Optics
Abstract:Phase-shifting profilometry (PSP) based 3D shape measurement is well established in various applications due to its high accuracy, simple implementation, and robustness to environmental illumination and surface texture. In PSP, higher depth resolution generally requires higher fringe density of projected patterns which, in turn, lead to severe phase ambiguities that must be solved with additional information from phase coding and/or geometric constraints. However, in order to guarantee the reliability of phase unwrapping, available techniques are usually accompanied by increased number of patterns, reduced amplitude of fringe, and complicated post-processing algorithms. In this work, we demonstrate that by using a quad-camera multi-view fringe projection system and carefully arranging the relative spatial positions between the cameras and the projector, it becomes possible to completely eliminate the phase ambiguities in conventional three-step PSP patterns with high-fringe-density without projecting any additional patterns or embedding any auxiliary signals. Benefiting from the position-optimized quad-camera system, stereo phase unwrapping can be efficiently and reliably performed by flexible phase consistency checks. Besides, redundant information of multiple phase consistency checks is fully used through a weighted phase difference scheme to further enhance the reliability of phase unwrapping. This paper explains the 3D measurement principle and the basic design of quad-camera system, and finally demonstrates that in a large measurement volume of 200 mm x 200 mm x 400 mm, the resultant dynamic 3D sensing system can realize real-time 3D reconstruction at 60 frames per second with a depth precision of 50 mu m.