Automated Detection and Counting of Pedestrians and Bicyclists Along an Urban Roadway

As urban transportation planning becomes more sophisticated, the accurate detection and counting of pedestrians and bicyclists becomes more important. In this project, two distinct camera-based approaches are integrated together to create a real-time pedestrian and bicyclist counting system which is regularly accurate to 85% and often achieves higher accuracy. The primary approach retasks a state-of-the-art traffic camera, the Autoscope Solo Terra, for pedestrian and bicyclist counting. Object detection regions are sized to identify multiple pedestrians moving in either direction on an urban sidewalk and bicyclists in an adjacent bicycle lane. Collected results are processed in real time, eliminating the need for video storage and postprocessing. Although this primary approach shows high accuracy, in some cases undercounting or overcounting of pedestrians can occur. To combat these issues, a second camera can be used to identify pedestrian heads and shoulders. Image recognition is then used to improve the accuracy of the count while still allowing the overall combined approach to operate in real time. In this report, results are presented for a pedestrian walkway for pedestrian flow up to 108 persons/min and the limitations of the implemented system are enumerated.