基于条纹投影的傅里叶变换轮廓术中高斯滤波窗口的研究

Research on Gaussian FilterWindow in Fourier Transform Profilometry Based on Fringe Projection

傅里叶变换轮廓术(FTP)是结构光投影三维测量中的一项重要技术,只需要一副经过高度调制的变形条纹,通过二维傅里叶变换,提取基频信息,进行逆傅里叶变换,相位展开和高度映射就可得到物体的三维数据信息.由于物体的高度信息主要调制在基频信号中,提取高质量的基频信号对测量的精度有重要的影响,另外还会受噪声的影响,滤波方法的研究尤为重要,在滤波中,一般采用窗函数滤波法如矩形窗,而本文主要采用高斯窗,通过计算机仿真分析在窗口大小一定的情况下,可以选择合适的截止频率,可以使高度均方差降低到0.1443mm,在实验中,通过选择截止频率可以提高物体模型相位展开的准确度.

In the field of 3D measurement of structural light projection,Fourier Transform Profilometry(FTP)is an important technique.With merely a highly modulated deformation fringe,fundamental frequency information can be extracted by 2D Fourier transform,and the 3D data information of objects can be obtained by inverse Fourier transform,phase unwrapping and height mapping.Since the height information of objects is mainly modulated in fundamental frequency signals,extracting fundamental frequency signals with high quality is of great importance to the accuracy of measurement.Besides,due to the interference of noise,the research of filtering method is particularly important.In the filtering process,window function filtering method like the rectangular window is often used.This paper mainly adopts Gaussian window.If the window is of a certain size,computer simulation analysis can choose appropriate cut-off frequency to reduce the mean variance of height to 0.1443mm.Experiments show that the selection of cut-off frequency can improve the phase unwrapping accuracy of the model object.