空间频率估计在傅里叶变换轮廓术中的应用

Application of Spatial Frequency Estimation in Fourier Transform Profilometry

作为空域位相分析的主要技术之一,傅里叶变换轮廓术在图像中引入载波,通过傅里叶变换在频域中滤波提取条纹信息,但其中大位相变形可能增加条纹分量的带宽,导致难以获取其位相图。对此提出将空间频率估计应用于傅里叶变换轮廓术,在不规则图像中应用傅里叶变换,利用估计的空间频率设定较小频率搜索范围,通过图像边缘上的位相梯度估计出边界外像素点的强度,从而将非矩形光场扩展成矩形光场,以边界外推提高傅里叶变换轮廓术的精度。仿真实验结果表明,将空间频率估计算法应用于傅里叶变换轮廓术,能提高搜索效率,有效减小傅里叶变换轮廓术的边缘误差,提升三维面形重建的速度和精度。

As one of the main techniques of spatial phase analysis,Fourier transform profilometry(FTP)introduces carrier into the pattern,and obtains the fringe after filtering in the frequency domain by using Fourier transform(FT).With it,the very large phase deformation in a spatial carrier fringe pattern may increase the bandwidth of fringe component,leading to difficulties in retrieving its phase map.It is proposed that spatial frequency estimation algorithm is applied to FTP,and FT is applied to non-rectangular patterns.A small frequency search range is set by using the estimated spatial frquency,and the intensity of pixels outside the boundary is estimated through the phase gradient on the edge of the pattern so as to expand the non-rectangular aperture into a rectangular one,and the boundary of the aperture is extrapolated to improve the accuracy of Fourier transform profilometry.The results of simulation experiments show that the application of spatial frequency estimation algorithm to Fourier transform profilometry can improve the search efficiency,reduce the edge error of Fourier transform profilometry,and enhance the speed and accuracy of 3D surface reconstruction.