Research on Motion Pattern Mining and Visualization of Moving Objects Based on Trajectory Data

Abstract

In recent years, with the popularity of the global positioning System, we have obtained a large amount of trajectory data due to the driving trajectory and the personal user's check-in information on the social network. Trajectory data can be understood as a curve formed in time sequence on a two-dimensional map during the movement of a moving object. Researchers analyze the location, destination and driving conditions of the moving object through trajectory data mining technology. However, it is difficult to analyze the group characteristics of moving objects only through the motion trajectory of a single moving object, so the concept of pattern mining is proposed. This paper focuses on the accuracy and efficiency of mobile object aggregation pattern mining. A clustering pattern mining method based on locus clustering is proposed. Firstly, the density-based clustering algorithm is used for spatial clustering of moving objects, and then the moving object clustering is used to form the spatiotemporal graph. Finally, according to the spatiotemporal graph, the clustering pattern mining algorithm is used to find the moving object clustering set that satisfies the spatiotemporal constraints accurately. This method introduces a spatiotemporal graph data model, which is composed of clusters of moving objects. Each node in the graph contains not only the information of the moving objects that make up the cluster, but also the formation time and position of the corresponding cluster, and each edge records the spatiotemporal relationship between the two clusters. The spatiotemporal graph can accurately reflect the spatiotemporal change characteristics of moving object clustering, and accurately analyze the clustering pattern. Based on the spatiotemporal graph, the algorithm GR is proposed to mine the aggregation pattern of moving objects, which can mine the aggregation pattern according to time and space information. The experimental results show that the proposed method is more accurate than the existing methods.