Video Stabilization for Micro Air Vehicles

Abstract: In Today�s world, Micro Air Vehicles (MAVs) equipped with digital cameras are increasing in number for real time applications. MAVs are widely used because of small, low weight, low cost and flexibility in the design. However, the video captured by MAVs will consist of unwanted translations and rotations due to atmospheric conditions. Hence, a video stabilization algorithm is needed to compensate the frames and then to avoid losing of any valuable information in video. This paper presents a real time video stabilization algorithm and mainly consists of three steps. Firstly, the key points from each frame are calculated by using Harris Corner Detection algorithm. Secondly, the descriptors between the two frames are used to calculate the affine transformation. Finally, this affine transformation is used to compensate the two successive frames in a video. This work mainly focuses on the digital image processing based video stabilization. The programming was primarily using MATLAB and later the algorithm was implemented on Raspberry Pi using Opencv-Python Programming.

Keywords: Micro Air Vehicles, Video stabilization, Harris Corner Detection, Affine transformation, Motion Compensation.