Precise Image Registration of airborne images is important and challenging because complex geometric distortions occur within the images during image acquisition for various applications. In this thesis an approach is proposed implementing automatic near real time Image Registration for remote sensing application concentrating on area based methods for control point detection and various outlier detection methods. The algorithm is generated using various techniques present and modifying them by adding new features for precise image registration of the data set used. The proposed algorithm was applied on various image data sets collected at different time intervals and results were tested for various methods and modifications. The modified schemes are also compared to evaluate the improvement over previous methods.
