Image registration using phase correlation in wavelet domain

Summary form only given. This paper presents an image registration technique in the frequency domain. We apply phase correlation to find the transition between images. However, phase correlation can find only translation on the x-y axis. We have to apply another technique before performing phase correlation to find an angular transition. Therefore, the Hough transform technique is implemented to overcome this problem. In this paper, images are pre-processed using the wavelet transform. The multiresolution property of the wavelet transform can help reduce processing time. Experiments are conducted to show the proposed structure of the wavelet; such as number of levels, selected sub-band, and other parameters suitable for image registration.