Abnormality Behavior Recognition Method for Ship in Bridge Waters Considering Small-Sample Problem

Abstract

With the development of transportation, more and more bridges are built, and the safety of bridges is becoming more and more important. In the waters of the bridge area, more and more ships pass through, and the navigation environment is more complex than that of the ordinary channel. For the staff of the traffic management department and bridge maintenance, it is necessary to always pay attention to the behavior of passing ships and remind the ships with abnormal behavior to prevent accidents that damage the bridge. Firstly, by analyzing the collision risk of ship bridge and the abnormal behavior of ship itself, this paper combines the two to obtain the judgment standard of ship abnormal behavior in the waters of bridge area; then, considering the small sample problem of ship abnormal behavior in the actual data, a large number of ship trajectory data are generated by the method of ship motion equation and random data generation. Based on the generated ship trajectory data, the identification model of ship abnormal behavior in the bridge area is established by using BP neural network classification method to give early warning of ship abnormal behavior, Provide technical means for the safe passage of ships in the waters of the bridge area to reduce or avoid the losses caused to the bridge.