Video-Based Pedestrian-Vehicle Conflict Analysis in Mixed Traffic Conditions

With dramatic development of image processing technology, a growing number of traffic flow detection and analysis studies have been conducted by using video data. Vehicle-only collision/conflict has been very well studied. Time to collision (TTC) and post-encroachment time (PET) are two major parameters to indicate vehicle-only collision severity. However, micro-level pedestrian-vehicle collisions/conflicts are less studied because they are hard to observe or record. This paper tries to extract traffic object locations from video data, to define the parameter time-difference-to-collision (TDTC) to fit the pedestrian involved potential collisions, to collect 188 groups of pedestrian-vehicle conflict data from two kinds of traffic intersections, to analyze the parameter trends between different pedestrian crossing choices, to compare the TDTC parameters among different safety levels, and finally validate the TDTC parameter as an indicator for pedestrian safety. The result shows that pedestrian speed, vehicle speed and TDTC parameters scatter more at the intersection with signal control; about 80% cases categorized by the TDTC parameter can have the same result with the independent observation; the shorter a TDTC is, the more dangerous a pedestrian may be, and a pedestrian-vehicle conflict with continuous short TDTC series can be dangerous. Further research will fours on more kinds of mixed traffic conditions.