Abstract
High quality video data is a core component in emerging remote toweroperations as it inherently contains a huge amount of information on which anair traffic controller can base decisions. Various digital technologies alsohave the potential to exploit this data to bring enhancements, includingtracking ground movements by relating events in the video view to theirpositions in 3D space. The total resolution of remote tower setups withmultiple cameras often exceeds 25 million RGB pixels and is captured at 30frames per second or more. It is thus a challenge to efficiently process allthe data in such a way as to provide relevant real-time enhancements to thecontroller. In this paper we discuss how a number of improvements can beimplemented efficiently on a single workstation by decoupling processes andutilizing hardware for parallel computing. We also highlight how decoupling theprocesses in this way increases resilience of the software solution in thesense that failure of a single component does not impair the function of theother components.