Aircraft ground monitoring with high performance computing multicore enabled video tracking

Safety of aircraft requires resiliency from onboard and off-board threats that could come from natural or man-made disturbances. Maintaining situation awareness requires monitoring and coordination of threats from the air and the ground. Threat detection of people, vehicles, and person-vehicle interactions of possible harm to an aircraft operations is a difficult problem due to the complexity of the coverage area, varying sensor capabilities (e.g., resolutions), and cultural factors of disruption. Methods and techniques can be incorporated to aid analysts (e.g., airport traffic controllers) to track and identify entities using modern large scale visual sensors such as the Wide Area Motion Imagery (WAMI) systems. Such systems typically produce an overwhelmingly large amount of information. The lack of computationally efficient algorithms has become a bottleneck for utilizing WAMI data in surveillance. To facilitate the application of such surveillance system development for safe operations, in this paper, three different strategies are implemented based on the on-board multicore technology to speed up video tracking algorithms. A complete tool chain to implement the video tracking, such as registration, detection, and multiple target association, is presented. Experimental results are illustrated. We demonstrate aircraft ground monitoring using the Columbus Large Image Format (CLIF) dataset to show the performance improvement by using the proposed high performance computing enabled video tracking algorithm to facilitate safe air travel.