基于指数尺度间隔连续小波变换的相位提取算法

Phase Retrieval Method with Continuous Wavelet Transform Based on Exponent Spacing Scales

只需要一幅调制图像的光栅投影测量方法主要有傅里叶变换轮廓术(FTP)、小波变换轮廓术(WTP)等。采用基于指数尺度间隔的连续小波变换与重构方法,提取调制图像的瞬时相位。针对指数尺度间隔连续小波变换,指出了足够大的噪声能够改变小波变换脊的位置,并且该脊向上移动的概率最大。因此,为了重构载频信号,选择脊及其紧邻的较大的那个尺度所对应的小波系数,采用灰度图像阈值分割中最大类间方差法(OTSU),剔除掉幅值较小的系数;针对斑点噪声的影响,对OTSU算法的结果进行了修正;使用修正后的系数集合重构载频信号,并计算该信号的瞬时相位。理论分析和实验结果表明算法有效且具有稳健性。

The methods which only need one modulated pattern for fringe projection measurement mainly have Fourier transform profilometry, wavelet transform profilometry, and so on. The continuous complex Morlet wavelet based on exponent scale interval and reconstruction is used to retrieve the instantaneous phase of fringe pattern. For the continuous wavelet transform, the great enough noise can change the position of wavelet transform ridge, the probability that the ridge moves up with the influence of noise is greatest. So, the wavelet coefficients corresponding to the ridge and the closest larger scale are chosen, and the maximum between-class variance methord （OTSU） is used to prevent the interference caused by low frequency components~ in order to overcome the speckle noise, the result of OTSU is modified and at last, the modified result is used to reconstruct the analytic carrier-frequency signal, and compute the intantaneous phase of the signal. The theoretical analysis and experimental results illustrate that the method is valid and robust.