Using computer vision and machine learning to identify bus safety risk factors

• DOI: 10.1016/j.aap.2023.107017
• Publication Date: 2023-03-06T17:48:26Z
• Authors: Becky P.Y. Loo; Zhuangyuan Fan; Ting Lian; Feiyang Zhang
• Keywords: Crowding

In road safety research, bus crashes are particularly noteworthy because of the large number passengers involved and challenge that it puts to network (with closure multiple lanes or entire roads for hours) public health care system injuries need be dispatched hospitals within a short time). The significance improving is high in cities heavily relying on buses as major means transport. recent paradigm shifts design from primarily vehicle-oriented people-oriented urge us examine street pedestrian behavioural factors more closely. Notably, environment highly dynamic, corresponding different times day. To fill this research gap, study leverages rich dataset - video data dashcam footage identify some high-risk estimating frequency crashes. This applies deep learning models computer vision techniques constructs series factors: exposure factors, jaywalking, stop crowding, sidewalk railing, sharp turning locations. Important risk identified, future planning interventions suggested. particular, administrations devote efforts improve along streets with volume pedestrians, recognise importance protection railing protecting pedestrians during serious crashes, take measures ease crowding prevent slight injuries.