Road traffic accidents are a leading cause of premature deaths and globally pose a severe threat to human lives. In particular, pedestrians crossing the road present a major cause of vehicle–pedestrian accidents in South Korea, but we lack dense behavioral data to understand the risk they face. This paper proposes a new analytical system for potential pedestrian risk scenes based on video footage obtained by road security cameras already deployed at unsignalized crosswalks. The system can automatically extract the behavioral features of vehicles and pedestrians, affecting the likelihood of potentially dangerous situations after detecting them in individual objects. With these features, we can analyze the movement patterns of vehicles and pedestrians at individual sites, and understand where potential traffic risk scenes occur frequently. Experiments were conducted on four selected behavioral features: vehicle velocity, pedestrian position, vehicle–pedestrian distance, and vehicle–crosswalk distance. Then, to show how they can be useful for monitoring the traffic behaviors on the road, the features are visualized and interpreted to show how they may or may not contribute to potential pedestrian risks at these crosswalks: (i) by analyzing vehicle velocity changes near the crosswalk when there are no pedestrians present; and (ii) analyzing vehicle velocities by vehicle–pedestrian distances when pedestrians are on the crosswalk. The feasibility of the proposed system is validated by applying the system to multiple unsignalized crosswalks in Osan city, South Korea.
Online Scheduling Technique To Handle Data Velocity Changes in Stream Workflows
