基于图像二维小波多尺度分解傅里叶变换轮廓术

应用傅里叶变换轮廓术测量物体三维面形时,当被测物体形状复杂或是被噪声严重污染时,导致频谱分布展宽,发生频谱混叠现象,基频提取困难,无法准确恢复物体的三维面型.提出了基于小波分解的傅里叶变换轮廓术,采用小波变换的方法对变形条纹图进行二维多尺度分解,重构被测物的背景图像,滤出图像的零频成分,得到相对变形条纹.运用小波变换与傅里叶变换轮廓术相结合的方法,只需拍摄一幅变形条纹图,将被测物体与背景分离,不受背景成分的影响,且易于基频信息的提取,降低了对滤波器的要求.实验证明该方法较好地防止了频谱的混叠问题,提高了测量范围与解相精度.

Fourier Transform Profilometry Based on Two-Dimensional Wavelet Transform of Image Processing

When the object shape is complex or polluted, it is easy to cause spectrum distribution spreading and spectrum mixed. It is unable accurately to restore the 3D shape by using Fourier Transform Profilometry to measure the shape. Fourier transform profilometry based on wavelet transform is proposed. 2D multiscale analysis of the distorted fringe is performed by using wavelet transform. The background image of the object measured is reconstructed. The zero-frequency component of image is filtered out to obtain the relative distorted fringe. It requires only one distorted fringe image by using the method with the combination of wavelet transform and Fourier transform. The object measured is separated from the background and it is not affected by the background component. It is easy to extract the fundamental frequency and reduces the requirement of filter. The experiment comfirms that the method measurement scope and resolution accuracy. prevents the spectrum mixing problem and improves the