Cyclist Speed Research Articles

Computer Vision Techniques for the Automated Collection of Cyclist Data

One of the main challenges in the conduct of detailed analysis of cyclist behavior is the lack of reliable data. Collection of data through manual methods is a labor-intensive and time-consuming process. Two of the important areas of cyclist data collection are volume counts and average speed measurements. A volume count provides the basis for necessary exposure measures and conveys essential information about traffic patterns. Cyclist speed data are used for traffic control and safety studies. The application of computer vision (CV) techniques enables the collection of precise spatial and temporal measurements of road users in a resource-efficient way. This paper presents the use of a set of CV techniques for the automated collection of cyclist data. Cyclist tracks obtained from video analysis were used to perform screen line counts as well as cyclist speed measurements. The applications were demonstrated with the use of a real-world data set from a roundabout in Vancouver, British Columbia, Canada. Further analysis was conducted on the mean speed of cyclists with regard to several factors (e.g., travel path, helmet use, group size). The motivation for this research was to understand better cyclist behavior and how it varied under different conditions. Several conclusions could be drawn from the analysis of cyclist speed behavior. Group size, travel path, lane position, and helmet use were all found to affect the cyclist mean speed. Single cyclists had a slightly, but significantly, higher mean cycling speed than did group cyclists. The mean cycling speed was highest for those cyclists who used the road rather than the sidewalk. The mean cycling speed decreased for cyclists without helmets.