Feature based image registration using steerable filters and Harris algorithm

Image registration is the process of mapping similarity between two or more images of the same scene taken at different times, from different viewpoints and/or by different sensors. The accuracy of a registration process is highly dependent on the feature detection and matching. In this paper, we use steerable filters and Harris corner detection algorithm to detect features, which are invariant to rotation, scaling and noise. Then initial matching is computed using Euclidean distance and mismatch features between point pairs are eliminated using RANSAC. Performance of image registration measured in terms of matching Score. The experimental result shows that the automatic registration algorithm is correct and effective.