Ultra-robust spatial point-to-point phase unwrapping algorithm for severe interference signal in 3D sensing

Spatial phase unwrapping (SPU) is a thorny ill-posed problem in the field of three-dimensional(3D) sensing, because the influence of interference signal such as noise signal and aliasing signal has always restricted the stable unwrapping of wrapped phase map. Ingeniously, the simplest filtering method has been used to successfully solve the most complex problem of phase unwrapping that has plagued researchers for years, accompanied by an increase in measurement accuracy. And, a point-to-point phase unwrapping algorithm is proposed, which differs from all previous methods of pixel-by-pixel unwrapping of noise-wrapped phase map. Most importantly, a differential iterative model with less than 11/2 phase shift is established to solve the long-standing problem of noise-induced local fringe order inaccuracy in SPU. The proposed model is universal to all SPU methods and can achieve ultra-robust anti-noise ability. Without noise correction, the bi-staggered phase unwrapping method (BSPU) is simplified to determine the fringe order of the smooth wrapped phase more efficiently and accurately. Point-to-point inverse calculation of the noise wrapped phase for adaptive correction of errors can be achieved due to the smoothing characteristics of the phase. The phase in the unstable region is effectively estimated by the iterative filtering model to further improve the measuring accuracy. Experimental results show that the proposed method is ultra-robust and universal in 3D sensing.

Automated summary

Spatial phase unwrapping (SPU) is a challenging problem in 3D sensing due to interference signals. A simple filtering method is used to successfully solve the difficult problem of phase unwrapping and improve measurement accuracy. A point-to-point phase unwrapping algorithm is proposed, which differs from previous pixel-by-pixel methods. A differential iterative model is established to address the issue of noise-induced local fringe order inaccuracy. The proposed method shows ultra-robustness and universality in 3D sensing.