面向交通轨迹的数据流可视化方法  

The Visualization Approach Based on Data Flow for Traffic Trajectory

作　　者：赵韦鑫 谭博友 周锐 王桂娟 陈华容[1] 吴亚东 Zhao Weixin;Tan Boyou;Zhou Rui;Wang Guijuan;Chen Huarong;Wu Yadong(School of Computer Science and Technology,Southwest University of Science and Technology,Mianyang 621010;School of Computer Science and Engineering,Sichuan University of Science and Engineering,Zigong 643002)

机构地区：[1]西南科技大学计算机科学与技术学院,绵阳621010 [2]四川轻化工大学计算机科学与工程学院,自贡643002

出　　处：《计算机辅助设计与图形学学报》2022年第5期768-776,共9页Journal of Computer-Aided Design & Computer Graphics

基　　金：国家自然科学基金(61872304);四川省杰出青年基金(19JCQN0108)。

摘　　要：为通过视觉图形交互式创建交通轨迹可视化,提出一种面向交通轨迹的数据流可视化方法.使用向导推荐辅助构建数据流图,完成交通轨迹聚类分析、相关性分析和异常检测.首先,开发数据流系统用于交互式构建交通轨迹可视化;其次,使用基于数据流异构网络的向导推荐模型辅助构建数据流图.该推荐模型将向导推荐作为图搜索问题,将用户关于向导推荐结果的评分反馈至数据流异构网络以更新节点关系.基于真实交通轨迹数据的案例评估和关于用户选择倾向的里克特量表评估结果表明,所提方法是有效的.To create traffic trajectory visualization through visual graphics interactively,the traffic trajectory-oriented data flow visualization method is proposed.The method uses wizard recommendations to assist in data flow construction for traffic track clustering,correlation analysis and anomaly detection.Firstly,the data flow system is developed to interactively create traffic trajectory visualization.Secondly,the wizard recommendation model based on heterogeneous networks of data flow is proposed to assist in data flow construction.The recommendation model treats wizard recommendations as a graph search problem and feeds user ratings on wizard recommendations to a heterogeneous network of data flow to update node relationships.The results of the case evaluation based on real traffic trajectory data and the evaluation of the Likert scale on users’propensity to choose indicate the validity of the approach.

关 键 词：交通轨迹 可视分析 交互式可视化 向导推荐 数据流 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]