Divide and conquer: high-accuracy and real-time 3D reconstruction of static objects using multiple-phase-shifted...

Area of Science:

• Computer Vision
• Optical Metrology
• 3D Imaging

Background:

• Structured light illumination is crucial for high-accuracy 3D reconstruction of static objects.
• Traditional methods struggle with real-time phase computation, limiting dynamic applications.
• Existing algorithms can be sensitive to nonlinear distortions in structured light systems.

Purpose of the Study:

• To develop a real-time phase computation method for multiple-phase-shifted structured light 3D reconstruction.
• To improve the speed and robustness of 3D reconstruction algorithms.
• To enable faster and more efficient 3D imaging of static objects.

Main Methods:

• A divide-and-conquer strategy is proposed, categorizing N=KM images into M groups of K phase-shifted images.
• Phases are computed for each group and then averaged to obtain the final phase.
• Look-up tables (LUTs) storing pre-computed arctangent values are utilized for K=3, 4, or 6, enabling rapid phase retrieval via table lookups.

Main Results:

• The proposed method achieves real-time phase computation, significantly improving speed.
• Accuracy is comparable to traditional algorithms, with robustness against nonlinear distortions.
• For K=4, M=3, phase computation speed is enhanced by a factor of 6x compared to classical 12-step methods using a single CPU core.

Conclusions:

• The novel approach enables real-time 3D reconstruction through efficient phase computation.
• The method offers a robust and accelerated alternative to conventional phase measuring profilometry.
• This advancement is particularly beneficial for applications requiring rapid and accurate 3D imaging.