Vehicle flow detection in real-time airborne traffic surveillance system

Compared with the ground-based traffic surveillance systems, the airborne platform takes advantage of convenience and low cost. In most of the existing aerial-based systems, the airborne platform is only responsible for data collection and the ground station is in charge of all data processing for vehicle detection. However, such architecture has a low real-time performance unless with the support of specifically constructed wireless networks. This paper proposed an airborne traffic surveillance system, which can detect vehicles in real time with only a general narrowband wireless network support. In our system, the airborne platform not only gathers video information but also carries out data pre-processing, and only the useful data is transferred to the ground via a GPRS/EDGE network. Moreover, more accurate vehicle flow detection with a false-alarm reduction module is performed on the ground station. Experimental results preliminarily show that the system can achieve real-time vehicle flow detection with a high detection rate and a low false-alarm rate.