International Journal of Engineering and Technology

Detection of the motion pattern of the object under study is the initial and major task in aerial surveillance. Object detection and tracking is the primary approach to find the interactions between the objects. Here the background modeled mean shift tracking algorithm is proposed for target detection in videos taken by Unmanned Aerial Vehicle. Traditional Mean shift tracking technique works well for dynamic objects with static background but UAV videos are of dynamic nature. The target may increase or decrease in size invariantly in consequent frames. Thus the proposed technique involves updation of background model adaptively for objects with random size variation and movement. The MATLAB based Simulation for the sample frames of UAV Videos is done and the performance evaluation is done to justify the efficiency of the proposed technique