Dual Frequency Unwrapping Optimization Algorithm for High-Speed Fringe Projection System

Due The dual-frequency number-theoretic phase unwrapping approach is often used in high-speed FPP 3D measurements due to its small number of projected images. However, the traditional number-theoretic phase unwrapping method still produces phase unwrapping errors, which are caused by the complex surface reflectivity of the object and the noise in the camera sensor. These errors significantly affect the reconstruction quality of the system. In this paper, we propose a method to improve the accuracy of the dual-frequency number-theoretic phase unwrapping algorithm. First, we correct the error using the unique attribute of the fringe error of the number theory phase unwrapping algorithm. Then, we obtain the local continuous phase using the multi-frequency phase unwrapping algorithm. Based on the local continuous phase, the spatial phase unwrapping approach can rectify low-precision absolute phase, significantly reducing the effect of noise and enhancing phase unwrapping reliability. Our experiments effectively demonstrate that our method enhances the algorithm's performance without requiring more patterns.