To reduce traffic accidents caused by abnormal pedestrian behaviors, the method to identify five dangerous abnormal pedestrian behaviors is studied in the paper. The algorithm contains pedestrian recognition, pedestrian tracking, and abnormal behavior detection. The paper places great emphasis on analyses and research of the relation between pedestrian trajectory and the road. Then the abnormal pedestrian behavior model is built based on the relation between the pedestrian trajectory and the road. Finally, the five abnormal pedestrian behaviors are detected separately based on the logical relationship of the abnormal pedestrian behaviors. Experiments show that the method can distinguish and detect the abnormal pedestrian behavior effectively, and it is suitable to use in real-time monitoring.
A Method of Abnormal Pedestrian Behavior Detection Based on the Trajectory Model
Fourth International Conference on Transportation Engineering ; 2013 ; Chengdu, China
ICTE 2013 ; 2861-2867
09.10.2013
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Vehicle abnormal behavior detection method and system based on trajectory data
Europäisches Patentamt | 2024
|Navigation trajectory feature-based ship abnormal behavior identification method
Europäisches Patentamt | 2022
|