Research On The Method Of Extracting Stay Area Of Ship Based On Trajectory Data

With the universal application of automatic ship identification systems,a large number of ship trajectory data have been accumulated,which have important value for marine management and service.Among them,how to,use the ship trajectory data mining analysis to obtain the information of the ship’s stay area is an important issue in the port management business.The analysis and excavation of the specific ship stay area(including berth and anchorage)in the port can not only provide the necessary basic information for the safe docking of the ship,but also provide an important source of information for the navigation of the ship port.At the same time,it also provides the planning and construction of the port.Important decision support.However,because the ship trajectory data has a large amount of data,the ship trajectory data has the problems of collecting sparseness,trajectory drift,and large space-time span with respect to other trajectory data.At the same time,the anchorage and berth of the ship in the port also have different staying behaviors.Therefore,based on the characteristics of ship trajectory data and the characteristics of port stop area,this paper studies the port stop area extraction method based on ship trajectory data.The main work of this thesis includes:1.Combining the characteristics of ship trajectory data,the concept of ship trajectory staying section is defined,and the ship trajectory data is denoised,interpolated,etc.in order to identify the ship trajectory big data with low precision,sparseness and drift.Based on the processing operation,a ship trajectory extraction method under multi-constrained conditions is proposed,and a ship trajectory extraction algorithm based on MapReduce is designed.2.Based on the difference of the anchoring behavior and the mooring behavior characteristics of the ship’s trajectory in the port staying area,a machine learning classification model based on random forest is proposed.The ship’s dwelling trajectory is classified,and two different trajectories of anchorage and berth are obtained respectively.Stay data.3.Based on the trajectory retention data of anchorage and berths,the anchor polygon extraction model based on trajectory rasterization and the berth polygon extraction model based on cluster analysis are designed respectively.The berth regular polygon data information and anchor irregular polygon are realized.Extraction of data information.In addition,the subject also verified and evaluated the above research content based on the real ship trajectory data set.At the same time,the berth and anchorage information of multiple ports extracted have also been applied to the actual business of the relevant units.