Image registration is a technique, which helps in comparing and combining two images, which are taken at same or different times, with same or different cameras. The resultant image, which is obtained after the combination, is useful for the analysts as it gives complete information in one image, which is earlier in two different images. For doing image registration one can use feature-based methods or similarity measures can be used for the optimization purpose. In this paper feature based and intensity based registration methods are compared using various transformation functions. The results clearly showed that geometric based measures outperformed intensity based transformations in terms of execution time. In addition, Global transformations outperformed local transformations in terms of execution time.
Image registration, feature based registration, intensity based registration, similarity metric, piecewise linear transformation.