Research On Ship Trajectory Compression Based On Threshold-guided Sampling Method

With the mandatory installation of the Automatic Identification System(AIS)on ships,the scale of its collected trajectory data has grown exponentially.Massive AIS data will bring severe challenges to data mining.Efficient trajectory compression technology can reduce trajectory data and help data mining.The existing ship trajectory compression methods lack comprehensive consideration of the key information of each trajectory,and lack of processing the trajectory data whose AIS data disappears for a long-time during compression.Finally,most of the methods are only for trajectory compression methods,and do not combine the actual applications after trajectory compression.Therefore,the compression of AIS trajectory data is still a challenge in the field of maritime traffic.To solve the problems in the field of ship trajectory compression,this thesis has done the following work:1)A threshold-guided sampling method considering multiple trajectory attributes is proposed to compress the trajectory.For the traditional trajectory compression method,only the individual attributes of the ship trajectory are considered,and other key information of the ship trajectory is ignored.Based on the direction preservation algorithm,this thesis proposes a threshold guided sampling method that combines ship motion characteristics and AIS data characteristics to compress ship trajectories.2)Determine the judgment conditions of each trajectory attribute and the judgment priority of the algorithm.First,accurately segment the trajectory of the ship by the angle of the ship trajectory time difference,identify the staying point and the trajectory data that has disappeared for a long time,and ensure that the trajectory conforms to the navigation habits;then,the azimuth angle difference,heading difference and speed difference of the trajectory are characterized by the azimuth angle Point,angle feature point and speed feature point are determined to determine the feature points in the trajectory;the determination order of all feature points is determined by the method of entropy value reduction,and all the feature points are retained by this determination order,and finally all the feature points are output.It is a compressed track.3)Use the threshold guided sampling method to compress the trajectory of the ship.By selecting the AIS data of Xiamen Port as the trajectory data source,and then using the algorithm proposed in this thesis and the direction preservation algorithm to compress the trajectory respectively,the compression results of the two are compared and analyzed.Experimental results show that the proposed algorithm can identify the trajectory feature points of different attributes of trajectory data and is superior to the direction preservation algorithm in compression efficiency,computing time and compression error rate.4)After compressing the ship’s trajectory,use DBSCAN’s trajectory clustering algorithm to verify the practicability.By selecting the AIS data of Xiamen Port,using the direction retention algorithm and the compression algorithm proposed in this thesis to compress the trajectory,and then use the DSSCAN clustering method to cluster the trajectory.The experimental results show that the results of trajectory compression using this method can not only be applied to trajectory clustering,but also the clustering effect is in line with expectations.Compared with the clustering effect compressed by the direction preservation algorithm,the result of trajectory clustering is also more ideal.In this thesis,AIS information is used as the support,and the proposed algorithm considers the multiple characteristic attributes of the ship’s trajectory to achieve a more comprehensive compression of the ship’s trajectory.The method of trajectory clustering verifies that the proposed algorithm can efficiently compress the trajectory while still retaining the key information of the original trajectory of the ship and can accurately identify and process the trajectory of the AIS information with a large time span and the trajectory of the ship at the stop point.Finally,after using the proposed algorithm to compress the trajectory,the compressed trajectory can be applied to the cluster analysis of the trajectory,and the ideal trajectory clustering result is obtained.