Stationary vehicle detection in aerial surveillance with a UAV

In this paper we propose an efficient static vehicle detection framework on aerial range data provided by the unmanned aerial vehicle (UAV) which is installed with a camera. Our system consists of three modules: moving vehicle detection, road area vehicle detection and post processing. First of all, we detect moving vehicles by clustering the singular points obtained after the motion estimation. Then, in the road area vehicle detection module, we detect all vehicles in the road area. For this purpose we propose an efficient method of automatic extraction of the road area. In order to detect both moving and static vehicles, we describe a method to detect blobs which are comprised of clusters of vehicle points. Finally, in post-processing module, we compare the results of the two modules mentioned above to distinguish the moving vehicles and the static vehicles with the help of different image coordinates. We evaluate our method on real aerial data and the experiments demonstrate the effectiveness of our approach.