| With the transfer of economic policies to inland areas,the state has vigorously developed inland waterway shipping.In order to improving the transport capacity and efficiency of inland rivers,the number of ships continues to increase,and ships are developing in the direction of large-scale and high-speed.But along with it,the traffic flow density of inland waterway is also increasing,and the navigation environment of inland waterway is becoming more and more complex,and in recent years bridges have been building at close distance in many waters,which significantly increases the probability of inland water traffic accidents.Water traffic accidents,especially ship collision accidents,often cause serious human death,property losses and environmental pollution.Therefore,the study of water traffic problems has attracted much attention.Automatic Identification System is a general automatic identification system in the navigation field,which can automatically exchange important information such as ship position,speed,course,ship name,call number and so on.Since AIS has been widely used,a large amount of AIS ship trajectory data has been generated.AIS data provide real-time navigation conditions of ships,which is acquisited conveniently,so that it can be effectively applied to the study of water traffic safety issues.Because the navigation environment of the multi-bridge waters is particularly complex and the probability of ship collision accidents is high,the following work has been done to find potential collision risk areas in multi-bridge waters through AIS data.(1)Referring to the concept of "traffic conflict" in the field of road traffic,defining ship collision conflict events,According to the principle of judging collision events,this paper presents a method of identifying collision danger points based on ship domain model using AIS data.Then,by analyzing the inland ship domain model and the characteristics of the multi-bridge waters,the ship domain model in the multi-bridge waters is constructed.Next the environmental factor coefficients of the ship domain model are determined by using the entropy and weight-fuzzy comprehensive evaluation model.(2)In this paper,2-d gaussian kernel density estimation is used to determine the parameter of DBSCAN algorithm adaptively.Then the parallel DBSCAN algorithm is designed and implemented on Spark platform.(3)Wuhan Port is a typical multi-bridge waters.This paper takes Wuhan Port as an example.Firstly,the entropy and weight-fuzzy comprehensive evaluation model is used to determine the environmental impact coefficients of long and short axes of ship domain in Wuhan Port,which are 0.714 and 0.625 respectively.Then,the collision danger points are identified by using the established ship domain model of Wuhan Port.Finally,parallel DBSCAN algorithm on spark is used to cluster the location of ship collision danger points,and 16 ship collision danger areas are obtained.In this paper,AIS ship trajectory data are adopted to identify caution areas of ship collision.This is of great significance for ensuring the safe navigation of ships,because the ship pilots will drive carefully in caution areas of ship collision and the maritime authorities also will strengthen the supervision of caution areas of ship collision. |
