Abstract: Aiming at the problem that how to improve the accuracy and speed of image registration, this paper presents a pyramid approach for image registration based on mutual information and Fourier-Mellin, which integrate the accuracy of calculating linear parameters in Fourier-Mellin-based method and the accuracy of calculating rotation and scale parameters in mutual-information-based method. Experimental results on remote sensing images demonstrate the presented method can decrease the registration time and improve the accuracy.

Key words: image registration, wavelet transform, Fourier-Mellin transform, mutual information

摘要： 针对如何提高图像配准的精度和速度的问题，利用基于Fourier-Mellin变换求解水平参量的准确性和基于互信息方法求解旋转及尺度变换参量准确性的优点，提出一种基于Fourier-Mellin和互信息的图像配准金字塔方法。对多幅遥感图像的实例仿真结果表明，该方法能在提高配准精度的同时减少配准时间。

关键词: 图像配准, 小波变换, Fourier-Mellin变换, 互信息

CLC Number: 

• TP391