提出一种基于克莱姆正则化分析法的三帧自标定相移阴影莫尔三维轮廓技术.该技术首先采用移动光栅的方法获得相移条纹图, 然后通过不同帧相移条纹图相减去除条纹图背景, 进而结合克莱姆正交化法和最小二乘法, 发展了一种相位解调方法, 提取了测量相位.以五步Harlharan算法为参考, 用不同算法对同一物体表面进行测量.结果表明, 相对于典型的三步相移法和主量分析方法, 提出的方法测量得到的相位误差最小(<0.5 rad), 且简化了测量过程.

A self-calibration three-dimensional object surface measurement method with three image frames based on Gram-Schmidt orthonormalization method was proposed. The phase shift fringe map was obtained by translation grating, and the fringe background was eliminated by subtracting phase shift fringe maps with different frames. Then a phase demodulation method using Gram-Schmidt orthonormalization method and the least-square method was developed, so that the measurement phase can be extracted. Taking five steps Harlharan algorithm as a reference, the object surface was measured by different algorithms. The results show that compared with three steps phase shift algorithm and principal component analysis algorithm, the proposed method is more precise with phase error less than 0.5 rad and simplifies the mensuration.