Beyond phase error compensation: pixel mapping-based error correction for high-accuracy 3D surface measurement
Zaixing He,Peilong Li,Xinyue Zhao,Shuyou Zhang,Jianrong Tan
DOI: https://doi.org/10.1088/1361-6501/ab6675
IF: 2.398
2020-03-17
Measurement Science and Technology
Abstract:In optical-based phase-shifting 3D surface profilometry, phase errors are inevitable and lead to final measurement errors. Error compensation, based on pre-estimated phase errors, is an effective technique to alleviate the phase errors. However, the static and inaccurate nature of the pre-estimated errors lead to an implicit limit to the compensation accuracy. Furthermore, the procedures for error pre-estimation require additional projections, which reduces convenience in practice. In this paper, a novel pixel mapping-based error correction method is proposed. Different from compensation, with the proposed pixel-mapping errors inside each camera pixel can be directly corrected by online replacing with its corresponding ground-truth values input in the projector. This replacement technique is the key idea of our method, which establishes a shortcut to error-free ground-truth values, skipping the traditional laborious and imprecise compensation procedures. To achieve pixel-mapping, an optimization method is proposed and no additional projections are required. A phase error formulation is also introduced in the optimization method to eliminate the phase error distrubance when computing the pixel mapping. Extensive experimental results are shown to demonstrate that the phase accuracy can be obviously improved after utilizing the proposed method.
engineering, multidisciplinary,instruments & instrumentation