Development of passenger safety system based on the moving object detection and test result on the real vehicle

Due to motorcycles or vehicles, sudden accidents often occur when passengers disembark from buses at bus stops. These crashes may often cause risks to passengers. In this paper, we present a vision-based threat object (TO) detection and passenger safety system that helps bus drivers and ensures passengers safety when getting off and on buses. In these systems, the fast detection of threat objects is the main priority. Therefore, we also propose a fast motion analysis algorithm based on subtracted images using the change values of the brightness difference of the pixels corresponding to the boundary of a segmented object. We configure a secondary camera on the back of the bus to detect oncoming objects, and a primary camera up to the door to analyze objects' status and track TOs. TO information is transmitted to the control unit, and the control unit controls the door's status or triggers alarms with an on-board speaker to provide a danger signal to passengers. We set up our system on a bus and verified it in a real road environment.