Research And Preliminary Practice Of Moving Target Spatiotemporal Big Data Analysis And Mining Technology

In the process of moving,the coordinate points which change with time are called trajectories,which describe the historical activities of the moving target for a long time,and record the space-time information of its activities in a more accurate and detailed way.The spatial-temporal big data analysis of moving objects can help us better understand and master the intention,purpose,operation rule and change trend of moving objects.The trajectory of moving object belongs to the space-time data,including time and space dimensions.According to the task of mining,spatiotemporal trajectory data mining can be divided into spatiotemporal clustering mining(mining tracks with high similarity among many moving objects,classifying those with similar motion into one category),spatiotemporal pattern mining(mining the data with the highest frequency from spatiotemporal data set,such as finding out the range,behavior and activity with high frequency of target occurrence in the activity track of moving objects And so on),spatiotemporal prediction and classification(starting from the historical moving track of the target,inferring its future moving trend,intention or behavior purpose,etc.)and spatiotemporal outlier detection(mining all moving tracks of a target,finding out the tracks with very low similarity with most other tracks),etc.This paper mainly studies the mining of the behavior rules of moving objects,which belongs to the category of temporal and spatial pattern mining.1)TDTR algorithm is used to compress the track data of moving target;2)for the index access of track data,the global segmentation framework block concept and coding design based on the idea of "turning the ball into the square" must support the organization,storage,index and access technology of actual target data quickly and effectively,and adopt efficient access technology Ask the general method of data in the space block to access the data acquired and stored in the space block efficiently;3)finally,mine and analyze the data,count the number of track points in each area according to the space block,filter the area with less track points according to the set threshold value,and use the DBSCAN clustering algorithm based on density to judge the density accessibility between the track points,which can not be eliminated Reachable active outliers and noise information.Finally,the frequent active space block is output to generate the frequent active area of the moving object.In the experiment part of this paper,the distributed technology is used to obtain the AIS data of the sea target ship from the open source Internet as the experimental data,then access the AIS data stored in the space block,count the AIS data in each space block area,generate the frequent active area of the sea target ship,and mine out its motion law and purpose.