Image Referencing Using Fast Fourier Transform

Abstract

The simpler case of multi-D data processing is the case when we have directly measured 2-D and 3-D data. This is the case with satellite images, aerial images, and many others; in computer science, the corresponding data sets are usually called images even when they do not represent a visual image. In such situations, the corresponding intensities are directly measured. The only information that sometimes is lacking - and that we need to recover by processing the data-is the information about the exact spatial locations in which these intensities are measured. In other words, in such situations, one of the main computational problems is to reference the corresponding image (or images). Usually, when we have a new image that requires referencing, we already have referenced previous images of the same area. So, a natural way to reference an image is to compare it with the previous images of the same area, and extract the spatial coordinates of the new points from this comparison. Image referencing is one of the main problems that we address in this paper.