Real time video stabilization methods in IR domain for UAVs — A review

Abstract

In applications that involve a camera attached to a moving platform, such as UAV, Car, Drones etc., mechanical vibrations of the moving platform cause undesired shaking motion in videos known as Video Jittering. This causes severe performance degradation in applications such as object detection and tracking for applications such as surveillance and remote sensing. Video stabilization is required as preprocessing step to remove undesired motion while preserving the desired ones. Video Stabilization is an active research area for past decade and most of work done is on offline methods for stabilizing video in visible range. However stabilizing Infrared videos and that too for real time applications is still an open research area. This is because of the difficulty in detecting and tracking good features in IR videos due to lower resolution/quality videos. In this paper various existing methods provided in literature are compared in detail to focus on method that is more feasible for real time IR stabilization implementation.