Visual Analysis Of Air Pollution Propagation Pattern Based On Particle Trajectory Tracking

Visual analysis is an effective means for analyzing big data,combining with visualization,human-computer interaction and automatic analysis technology.It aims to reveal the inherent laws of data in a graphical way,and helps users mine clear and structured information from complex data.With increasingly serious air pollution,how to deal with and analyze the air quality data effectively has become a hot issue.Air pollution in different regions may interact with each other,which is usually complex and has specific temporal and spatial characteristics.Previous studies on the spread of air pollution usually adopted complex regression models to simulate point-to-point pollution diffusion,failing to comprehensively consider the impact of surrounding areas on local air quality and its time-varying laws,and making it difficult for analysts to make spatial-temporal targeted decisions.Visual analysis of air pollution propagation patterns is helpful to understand the source information of air pollution and its time-varying laws,so as to promote the joint treatment of air pollution.In recent years,large-scale air quality data and meteorological data have been collected through air quality monitoring stations and sensor networks,enabling researchers to analyze air pollution from data scientific perspective.However,the multi-source heterogeneity,high dimensionality,time-varying,and large scale of the data pose great challenges to the analysis of air pollution propagation patterns.In this paper,an effective method is present to identify potential pollution sources and propagation patterns,and visualization technology is also introduced to deeply analyze air pollution problems from multiple perspectives.Considering the influence of meteorological factors on the diffusion and settlement of air pollutants,integrating multi-source heterogeneous spatial-temporal data,an air pollution propagation model is presented for time-series particle trajectory tracking.A method for identifying potential pollution sources is proposed,simultaneously considering the correlation among the pollutant sources influence frequency,transport value,and affected stations.The strong influence station group can be effectively detected.For each target station,based on the propagation information of the strongly affected station group,the time-series feature vector is constructed.The clustering algorithm with high accuracy and efficiency is adopted to extract the pollution propagation patterns based on the constructed time-series feature vector.An integrated visual analysis system is designed and developed.The well-designed visual view,rich multi-graph linkage technology,zoom-pan,filtering and other exploration and analysis functions will help users conduct in-depth analysis of air pollution problems in areas of interest from multiple perspectives.The effectiveness of the proposed method is demonstrated by real-world datas.