在相位测量型光学3维传感技术中,为了正确恢复物体的真实相位,采用了一种基于调制度轮廓线连接极点的相位展开算法。该算法将调制度分布图与分支阻断法相结合,利用图像增强技术中拉普拉斯算子检测边缘,从被测物体调制度分布中提取物体的轮廓线,用于指导极点的连接和截断线的设置,最后结合菱形算法快速展开被测物体的截断相位,能够得到较为真实的相位展开结果。实验结果表明,新方法结合了调制度分布图展开相位可靠和分支阻断法计算时速快的优点。在面形存在陡变和凹陷的复杂情况下,新算法甚至比可靠度排序法更能够有效地恢复出物体的真实相位,保证了测量的精度,而且运算速度比可靠度排序算法快了3倍。

In order to retrieve the nature phase of the tested objects in optical three dimensional sensing based on phase measuring profilometry,a new phase unwrapping method,which connects the residue pairs by branch cuts based on the outline of modulation of the measured object,was presented.The new algorithm employs Laplacian edge detection to extract the outline of modulation,used to guide the branch cuts setting.Then,adopting diamond phase-unwrapping algorithm,the wrapped phase of the measured object can be retrieved accurately.The experimental results show that this new method,combining with the advantages of the reliability of modulation and the rapidity of branch cut algorithm,can effectively remain the accuracy of the measurement and fast retrieve the nature phase of the tested objects under the condition of a steep-shape.It is three times faster than the reliability-guided phase unwrapping algorithm.