广义条纹图序列编码的相位重建

Phase Reconstruction Based on Generalized Fringe Pattern Sequence Encoding

为解决拓扑复杂物体三维数字化过程中的绝对相位测量及相位重建问题,提出了一种广义变频条纹图序列的相位展开算法。简要综述了现有的各种时间维相位展开算法并指出这些算法存在的问题,在此基础上,详细分析比较提出方法与正指数条纹序列算法的噪声免疫力和时间复杂度,证明该方法有更高的噪声免疫力和更低的时间复杂度。特别是,当相邻编码条纹图的条纹数之比不等于2时,正指数条纹序列算法出现严重噪声,而提出的方法依然能够正常工作。此外,该算法还能够根据实际情况决定需要投影的条纹频率,从而能够使用相对较少的条纹编码图像进行有效的相位展开,进而提高系统的实时性。理论分析和实验结果都证明了提出算法的有效性。

A phase unwrapping algorithm with generalized fringe pattern sequence is proposed for absolute phase measurement and phase reconstruction of three-dimensional digital object surfaces with complex topography. Firstly, various phase reconstruction techniques based on temporal phase unwrapping are reviewed and the problems existing in these techniques are pointed out. On the basis of such analysis, the detailed comparison between the fringe sequence algorithm based on positive exponent and the algorithm proposed is made in terms of the noise immunity and computational complexity. It is shown that the proposed approach has better performance. In particular, the fringe sequence algorithm based on positive exponent shows severe noise in the case that the ratio of number of fringes of successive patterns does not equal to 2, whereas the proposed approach is able to work well as usual. Furthermore, the proposed approach can determine the number of projection desired fringes according to practical situation, leading to less encoded fringe patterns required for phase unrapping and decreased processing time. The theoretical analysis and experimental results show the validity of the proposed algorithm.