Automated Video Incident Detection (AVID) System
项目名称: Automated Video Incident Detection (AVID) System
摘要: Detection and verification of incidents on major freeways is one of the most critical functions for incident response. Studies show that every seven minute delay in detection results in one additional mile of queue in the system. Therefore, early detection and verification of an incident results in less congestion and faster restoration of traffic flow. One of the methods that Transportation Management Centers (TMCs) use to verify an incident is to manually monitor Closed Circuit Television (CCTV) cameras once the incident is identified. The cameras are typically in idle mode over 95% of the time, performing no functions when not being used for incident verification. The Department lacks an automated method to detect incidents, so existing CCTV cameras are only used passively for manual verification after the TMC is informed of an incident. Currently, TMC operators use CCTV cameras to verify incidents upon either receiving calls from commuters, monitoring California Highway Patrol (CHP) logs and news media, or viewing the Advanced Transportation Management System (ATMS) map, which presents freeway vehicle speeds through the vehicle detection systems that feed information to the TMC. Once informed of an incident, they monitor the CCTV cameras to verify it prior to taking appropriate action in responding to and removing the incident. Therefore, time is lost during the period between the operators being informed of the incident and manually identifying its location using the existing CCTV cameras. This time loss will add a few minutes to the verification time of an incident, resulting in additional total delay. If the Department had a method to automatically detect incidents using existing CCTV cameras, which are currently only used for manual incident verification, TMC operators could detect and respond to incidents more rapidly.
状态: Active
资金: 0
资助组织: California Department of Transportation
项目负责人: Slonaker, John
主要研究人员: Recker, Will
开始时间: 20150112
预计完成日期: 20170111
实际结束时间: 0
检索历史
应用推荐