Real-time detection of moving vehicles

Computer vision-based traffic flow monitoring is of major importance for enforcing traffic management policies. Information such as the number of vehicles passing on a road per time unit, or vehicles' turning rates at intersections are exploited by traffic management policies to supervise traffic-light timings. Computer vision-based traffic flow monitoring requires extraction of moving vehicles from traffic scenes in real time. To accomplish this task, efficient algorithms must be used and effective, low-cost hardware implementation must be pursued. This paper first describes the algorithms used in the VTTS (vehicular traffic tracking system) to achieve segmentation of moving vehicles. Then, hardware implementation on a re-programmable FPGA-based board is described in detail.