【摘要】 传统的线扫描空间相位展开算法可以实现快速相位展开,但容易受到噪声、离焦等引起的残余误差传播的影响,导致重建结果出错,众多的拉线误差甚至无法重建。由此,提出了一种尽可能抑制拉线误差的快速空间相位展开算法。所提算法在条纹投影轮廓术的基础上,将所获得的截断相位二值化,生成二值化的截断相位序列,规避了相位阈值对条纹级次的影响。此外,对该序列进行前后交叉的多锚点验证,寻找级次跳变点,确认条纹级次图。最后,由条纹级次图和原始的截断相位计算出展开相位和物体深度信息。实验结果表明了该算法的可行性和有效性。同时该算法继承了传统线扫描算法高速相位展开特性,且具有更高的测量精度和抗噪声性能。更多还原

【Abstract】 Traditional line scanning spatial phase unwrapping algorithms can achieve rapid phase unwrapping,however,it is susceptible to pull-line error propagation caused by factors such as noise and defocusing,which can lead to reconstruction errors or even failure. In light of these challenges,this paper proposes a spatial phase unwrapping algorithm with suppressing pull-line error as far as possible. The proposed algorithm,built upon the principles of fringe projection profilometry,binarizes the obtained wrapped phase to generate a binary wrapped phase sequence,thereby mitigating the influence of phase thresholds on fringe orders. Furthermore,it employs multi-anchor point validation by forward-backward crossings of this sequence to locate fringe jump points and confirm the fringe order map. Finally,the unwrapping phase and object depth information are computed from the fringe order map and the original wrapped phase. Experimental results demonstrate the feasibility and effectiveness of this algorithm. Moreover,it inherits the high-speed phase unwrapping capabilities of traditional line scanning algorithms while achieving higher measurement accuracy and noise robustness.更多还原