Visual surveillance system for monitoring of passenger flows at public transportation junctions

Abstract

The goal of the presented project Junction Flow is the development of an innovative pilot system for automated passenger flow analysis. Most critical transportation junctions are monitored using image understanding from visual surveillance and AI based trajectory analysis. With this mobile and continuously employable transport telematics system, the transfer and change behaviour of passengers at transportation nodes within the public transport system is automatically gathered and analyzed. This allows to continuously determine current and accurate passenger flow data with little effort, either regularly, on short term notice or over long periods of time. While the corresponding project Junction Flow is in its early stages we provide a thorough overview on the conceptual framework of the system, describe the functional system architecture and both already implemented and targeted interpretation capabilities within a cognitive vision system controlling a distributed sensor network of surveillance cameras. The preliminary outline of the computer vision system demonstrates first experimental results with promising performance by applying an innovative methodology for passenger tracking for the analysis of passenger junction flows.