基于相位相关的小波域图像配准方法研究

小波变换是一种多尺度信号分析方法，近几年在图像处理领域受到广泛关注，它克服了傅立叶变换的固定分辨率的弱点，既可分析信号概貌，又可分析信号的细节。相位相关是一种频率域的图像配准参数估计方法，是利用傅立叶变换的平移、旋转等特性进行参数估计的。在研究多尺度小波分析和相位相关理论的基础上，提出基于小波系数的像素级相位相关图像配准方法：首先对待配准图像进行小波分解，获得低频小波系数后，再对小波系数应用相位相关进行配准参数估计。实验结果表明了该方法的可行性和有效性。

Image registration approach based on phase correlation in wavelet domain

Wavelet decomposition of the image is recommended for the pyramidal approach due to its inherent multi-resolution character. The advantage of this decomposition is that signal trends at different scales are isolated and studied because wavelet theory is the decomposition of signals into components at different scales or resolutions. Phase correlation method is a kind of registration method in frequency domain, and it is based on the well-known Fourier property. Multi-resolution wavelet analysis theory and phase correlation are introduced, and novel phase correlation image registration approach based on wavelet coefficients is presented. Firstly, wavelet decomposition of the images are performed. Secondly, low frequency wavelet coefficients are obtained. Finally, phase correlation is performed using wavelet coefficients. Experimental results show that satisfactory effect is obtained by applying the method.