| With the development of the world economy,the transportation task of water traffic is becoming increasingly heavy,and marine traffic routes are intertwined and becoming increasingly complex,gradually forming a route network.The study of the method of constructing a route network in a specific water area is of great significance in understanding the changes in the route in that water area,improving the safety of ships and enhancing safety management.This study combines data mining techniques such as trajectory compression,segmentation,feature point identification and cluster analysis,and proposes a framework for constructing a route network based on AIS data and information entropy.Using the ship AIS trajectory data,a route network was constructed for the Yangtze River estuary waters and waters near the Maao port area,and the constructed route network were analyzed to verify the effectiveness of the route network construction framework.The specific work includes the following three aspects.Firstly,to address the difficulty in determining the compression threshold of TD-TR algorithm nowadays,a TD-TR compression algorithm with adaptive compression threshold is proposed based on the calculation of the compression threshold by adaptive score.The compression results of the adaptive threshold and the manually determined threshold are compared for the raw ship trajectories in the Yangtze estuary waters.The results show that the improved algorithm can remove a large number of redundant trajectory points,and the compression threshold is adaptively determined according to the characteristics of each trajectory,which is not affected by human factors.Finally,TD-TR compression with adaptive thresholding was performed on the raw ship trajectories within the Yangtze estuary waters.Secondly,to solve the problem that segmentation and feature point recognition algorithms require artificial determination of course thresholds,speed thresholds and even time thresholds,this study proposes an adaptive segmentation and feature point recognition algorithm for trajectories based on information entropy considering AIS dynamic information.The basis of the route network construction is the identification of nodes.Although the nodes of the route network can be generated based on the arrival and departure records of ships,these records are usually extracted from AIS static messages and the data quality is much lower than that of AIS dynamic messages.Based on the compressed ship trajectories in the investigated waters,the algorithm is used to identify key feature points such as speed feature points,course feature points and trajectory start and end points of the ship,which lays the foundation for generating route network nodes and constructing route networks.Finally,based on the OPTICS clustering algorithm and graph theory on the basis of the previous study,the route network of the Yangtze estuary waters and waters near the Maao port area was constructed.Specifically,the key feature points of ships are abstracted into nodes by OPTICS clustering algorithm,and the ship route network is formed based on graph theory with the voyage traces existing between the nodes as edges.The route network is analyzed in relation to the real environment within the investigated waters.This paper proposes a network construction framework that integrates ship information such as course,speed and geographic location,which can discover the distribution pattern of ship trajectories among the scattered and chaotic ship trajectories.The research results are of certain application value for the relevant departments to grasp the water traffic situation and optimize the adjustment of marine routes. |
