摘要：Road safety in primary school areas is a delicate issue due to the vulnerability of children as road users. They are inexperienced traffic participants and sometimes their behavior in traffic situations is unpredictable. This paper reports a safety analysis conducted using video processing of conflict trajectories of vehicles and active transportation users (AT users). The videos were collected using unmanned aerial vehicles (UAVs) as this technology does not affect the actual behavior of traffic participants. Due to their airborne position, UAVs can conveniently gather information about driving behaviors and the exact positions of various participants. The safety analysis was conducted using surrogate indicators suitable for evaluating the risk of potential collisions as they incorporate both spatial proximity and speed. Three conflict indicators were used in the safety analysis: (i) time-to-collision, (ii) post-encroachment time, and (iii) heavy braking. The methodology was tested in a primary school area in the city of Zagreb. With the applied methodology, a total of 43 potential conflicts were identified in the school area (14 time-to-collision, 23 post-encroachment time, and 6 heavy braking). Based on the determined potential conflicts, safety measures were proposed to decrease the number and severity of potential conflicts and to increase traffic safety near school areas.
