三维物体形貌检测图像处理中的小波数字滤波技术

傅里叶变换轮廓术和相移法是投影栅法测物体三维形貌的两种重要方法。相移法需要精密的移相装置,且需采集一幅以上的图像,在应用中受到一定限制。而傅里叶变换轮廓术只需一幅图像便可计算出各点相位值,测量精度高,测量范围广,因而得到广泛运用。最大测量范围受到高度分布h(x,y)在与光栅线垂直向上变化率的限制,即面形变化剧烈的物体易产生相邻谱区的混合,不易测准。本文提出的基于小波数字滤波的傅里叶变换法,可以将物体的可测梯度提高3倍。因为利用小波技术对变形栅线图进行处理后,滤去了直流分量和高频分量,得到的相对栅线图只剩下基频分量。采用小波数字滤波与傅里叶变换轮廓术相结合的方法,用小波变换去除直流成分和高频成分,从而提高了物体的可测精度,降低了对滤波器的设计要求,能较好地将物体的相位信息分离出来。本文给出了这种算法的理论分析和计算机模拟试验的结果。

Wavelet Digital Filter Method of Image Processing in 3D Measurement Profilometry

Fourier transform profilometry and Phase shifting are two important methods in 3-D measurement of body.Phase-shifting requires a precise Phase-shifting set and more than one image collected,therefore its application is limited.Fourier Transform Profilometry requires only one image and the phase values of each points can be calculated.The precision and range of measurements are higher and wider,therefor its application is very broad.The measured region is limited by the altitude distribution of the function h(x,y).This paper proposes a Fourier Transform Profilometry based on Wavelet Digital Filter,which can increase three times for the measured grads of a surface.Wavelet Digital Filter and Fourier Transform Profilometry are combined,higher spectrum and direction components are removed.Therefor the design demand is played down,dividing phase degree of precision is raised and the phase information of measured body is resolved.Analysis of the above calculus and the results of simulated experiments are presented.