Video Sequence Stabilization for Real-Time Remote Sensing Applications

Video sequence stabilization permits an accurate analysis of the video originate from sources as video cameras or IR sensors. In fact, a system of video stabilization allows the availability of aligned image sequences and, consequently, to improve the operator analysis of the scenario. In this paper different image processing methods are analyzed to remove the camera unwanted motions and reconstruct a stabilized sequence for real time applications. In particular three methodologies are compared in term of computational load and of the expected scenarios to determine the best strategies to be applied in a real time system.