A rising application area for intuVision's video analytics in the traffic domain is tracking near-miss incidents between pedestrians and vehicles at intersections or other pedestrian crosswalks, as well as detecting non-compliant driver behavior. A near-miss incident may occur when a vehicle does not stop at a crosswalk when there are pedestrians, or pedestrians proceed to cross the street during a red light while there are vehicles passing through. Driver non-compliance around the crosswalks involve not slowing down approaching a crosswalk, or not stopping despite the traffic signage. Studying near-miss incidents that have the potential to cause, but do not actually result in human injury, provide data for planning safer crosswalks and signage for safety of pedestrians.
Detecting and keeping a record of near-miss incidents, or non-compliant driver behaviors are very important for preventing future accidents. intuVision VA's near-miss detection criteria aims to ensure the safety of pedestrians using crosswalks during their allowed times, either when allowed by crossing signage or following right-of-way in instances where traffic signals are not present. Factors such as the locations, time periods, and travel directions all represent important data in understanding the nature of these events. To facilitate later analysis, intuVision VA keeps a detailed log of all detected events with time stamps, descriptions, and image snapshots. This information can be used to cross-reference with incident reports, to track risky behavior around crosswalks, and to drive decisions around placement of ticketing agents, adjustments to signage, or other enforcement.
In the video below, intuVision VA Traffic module is shown in action at an intersection crosswalk, alarming for vehicles putting the safety of people at risk and turning into the crosswalk while the pedestrians are crossing. However, vehicles travelling parallel to the crosswalk or safely stopping before the crosswalk do not trigger an alarm.
The analytics also include options to tie into existing traffic signaling. This allows for creating of meta-reports, including separating incidents where the pedestrians had right of way from those when vehicles had right of way and pedestrians were crossing when not allowed.
Other potential uses for this near-miss detection are:
If you are interest in learning more about analytics for detecting traffic incidents or for any other unique application of video analytics, please contact us to learn more.