Fast image registration for Sub-Pixel using pyramid edge images

Abstract

Although image registration has been widely used in many image processing-related fields, it still has difficulties to be solved yet. One of the difficulties is to fine the accurate information such as translation, rotation, and scaling factors between images. Besides, it needs high computational cost. In order to solve such problems, we propose a new image registration method base on Fourier Transform through pyramid edge images and a simple line fitting. The main advantages of the proposed algorithm are that 1) it can compute the accurate information at sub-pixel precision and 2) performs fast for image registration. Experimental results showed that the proposed scheme is more efficient than other baseline algorithms for the large images such as aerial images. Therefore, we believe that the proposed approach can be a useful tool for processing-related fields that requires high efficiency such as GIS, MRI, CT, image mosaicing, and weather forecasting. © 2013 International Information Institute.