Study Of Automated Generalization Method For Mobile Trajectories Based On Mixed Spatiotemporal And Semantic Information

Moving trajectory data is an important part of geographic space-time big data.It also has the characteristics of time and space distribution and scale characteristics,and its characteristics of volume,diversification,real-time and network make data processing,transmission and analysis need more automatic integration and scale transformation support.Compared with other geospatial data,trajectory data has unique temporal,spatial and semantic features,and its information simplification and feature extraction have different automatic synthesis strategies.This research focuses on the trajectory data faced by the current positioning technology and location service.Taking the GPS trajectory as the object,the method of spatiotemporal and semantic information fusion is used to explore the space-time,semantic and scale characteristics of the moving trajectory by introducing the geographical environment of the moving trajectory.A time-space-semantic fusion-based moving trajectory expression model is established.Based on this,a segmentation synthesis strategy and adaptive synthesis method for moving trajectories are proposed,which provides basic support for mobile trajectory big data analysis and mining.To sum up,the research content and results of this paper mainly include the following five aspects:(1)The characteristics of the moving trajectory are analyzed.A moving track is not a type of geographic entity.It is a trace of a moving object that moves in a geospatial environment for a certain purpose.It is constrained by the geographical environment and has dynamic spatiotemporal information.Therefore,the moving trajectory data is a special kind of geographic spatiotemporal data,which has the characteristics of unstable data precision,integration of spatiotemporal information,rich semantic information and fuzzy spatial scale.This paper analyzes the temporal,spatial,semantic and scale features of the moving trajectory,and proposes the "four spatial features","four kinds of temporal features" and "four kinds of space-time fusion methods" of the moving trajectory.The connotation and classification of the semantic concept of the trajectory are studied,and four trajectory semantic categories such as moving object,geospatial environment,moving mode and collecting equipment are proposed.The time scale,spatial scale and semantic scale characteristics of the trajectory are discussed,and the transformation relationship between spatial scale and time scale is established,and the consistency relationship between the semantic scale and the space-time scale is discussed.(2)A trajectory expression model of spatiotemporal-semantic fusion is constructed.The spatiotemporal representation of the moving trajectory is usually a sequence consisting of spatiotemporal points.The semantic expression model of the moving trajectory is currently used in the Stop/Move model and the CONSTANT model.The two models each have their own applicable scenarios and there are deficiencies in spatiotemporal-semantic synchronization analysis and the scaled analysis.Based on this,this paper proposes two trajectory expression models of spatiotemporal-semantic fusion: the spatiotemporalsemantic integrated expression model based on trajectory points and the trajectory expression model of spatiotemporal-semantic correlation.The former directly adds semantic information to the temporal and spatial points of the trajectory,thus integrating expression;the latter uses semantic nodes to express semantic information,and then associates with space-time points.In addition,for the construction of trajectory expression model of spatiotemporal semantic association,a method of trajectory stay segment extraction based on clustering algorithm is proposed,and the trajectory semantic segment matching method is designed and verified by an example.(3)A trajectory segmentation synthesis method based on semantic constraints is proposed.The moving trajectory implies the semantic information of various geospatial environments such as road network semantics and location semantics.The purpose of the synthesis is not simply to compress the space-time trajectory,but to express and synthesize the information with semantic trajectory analysis and information mining as goals,therefore,it is necessary to consider the trajectory semantic information and its semantic synthesis.Based on the spatiotemporal-semantic association expression model,this paper constructs a trajectory synthesis framework of spatiotemporal-semantic fusion which is the trajectory segmentation synthesis based on semantic constraints,including the semantic synthesis method of trajectory,the simplification method of trajectory stay segment,and the simplification method of the trajectory moving segment.Through the personal GPS trajectory data test,the integrated method can provide a large compression ratio while maintaining good space-time precision,and can also obtain multi-scale semantic information,which can be applied to the analysis needs of different scenarios.(4)A trajectory simplification method for feature point queuing under road network constraints is proposed.For the trajectory of the vehicle,for example,due to its large amount of data and close correlation with the road network,the comprehensive target is to obtain high-precision and high-compression data,so that the compressed data maintains good temporal and spatial characteristics and road network semantic features.Therefore,the timespace-semantic integration is used to express the moving trajectory,and the trajectory points are used as the evaluation object.The temporal and spatial features of the trajectory points and the semantic features of the road network are evaluated respectively,thereby establishing the spatiotemporal feature queue of the trajectory points and the spatiotemporal feature queue considering the semantics of the road network.Finally,the trajectory is simplified according to a certain compression ratio.The verification of the taxi trajectory data at different time scales shows that the proposed method has different degrees of improvement in algorithm efficiency,compression accuracy and road network feature points.(5)An evaluation index for the comprehensive quality of the trajectory is established.Based on the original time-space feature accuracy evaluation index,combined with the track road network matching,two improved spatiotemporal feature accuracy evaluation indexes of network homomorphic distance error and matching homomorphic distance error are proposed.For the semantic precision evaluation,two evaluation indicators of the number and the correct rate of stay semantic segment is proposed.