交互式套牌车可视识别与分析  

Interactive Visual Analysis of Fake Plate Vehicles Detection

作　　者：任水林 陶煜波[1] 林海[1] Ren Shuilin;Tao Yubo;Lin Hai(State Key Laboratory of CAD&CG, Zhejiang University, Hangzhou 310058)

机构地区：[1]浙江大学CAD&CG国家重点实验室,杭州310058

出　　处：《计算机辅助设计与图形学学报》2016年第11期1887-1898,共12页Journal of Computer-Aided Design & Computer Graphics

基　　金：国家自然科学基金(61472354);国家"八六三"高技术研究发展计划(2012AA12A404)

摘　　要：交通卡口数据能够有效地用于识别套牌车行为,已有的方法通常基于规则查询套牌车记录,但较少提供规则的含义及其对查询结果的影响、套牌车行为模式的分析.为此提出一个基于交通卡口数据的交互式套牌车可视识别和分析系统.基于规则识别套牌车,设计可视查询与条件筛选,利用地图和时间线展示不同车型的时空分布,并分析套牌车主要出现的卡口和时间;通过设计Cir Flow图符和分布图展示单卡口和卡口对的统计信息,分析套牌车识别算法中的参数调整对识别结果的影响.最后,通过案例验证了系统的可用性.Sparse traffic trajectory data can be used to detect fake plate vehicles effectively. Most existing techniquesare based on rules to search for fake plate vehicles records, but they generally don’t provide intuitivemeanings of rules and their impacts on results. Moreover, they provides less analysis on the behavior patterns offake plate vehicles. This paper presents a visual analysis system to interactively detect and analyze fake plate vehiclesbased on sparse traffic trajectory data. We first design a visual query model based on the rule-based fakeplate vehicles detection algorithm, then use maps and timelines to display temporal spatial distribution of differenttypes of fake plate vehicles, especially for cell locations and times that fake plate vehicles appear multiple times.The statistical information on a single cell and a cell pair is presented by CirFlow and histograms, to visuallyanalyze the impacts of different rules on the results. Finally, Case studies show the effectiveness of our system.

关 键 词：套牌车识别 可视查询 轨迹数据 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]