稀疏交通轨迹数据的可视分析及系统开发

发布时间：2018-01-26 03:51

本文关键词： 可视分析稀疏交通轨迹套牌车识别用户行为模式张量分解　出处：《浙江大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着城市的快速发展,城市生活更加多元化和复杂化,人们丰富的出行活动带来了大量的移动轨迹数据。轨迹数据包含时空和语义信息,有助于城市规划和人们行为活动的理解,已成为一大研究热点。本文采用开车数据和公共交通数据对用户出行模式进行了的相关可视分析,实现了相应的可视分析技术。本文以稀疏交通卡口数据研究开车出行行为模式。系统实现了交通卡口宏观流量探索;基于套牌车识别算法设计了可视查询模型;通过空间气泡图、CirFlow图、往返时间分布图等分析可视查询结果在全局、单卡口、卡口对上的时空分布。通过案例分析了交通卡口流量分布、套牌车行为的识别过程等,验证了本文方法能够有效分析卡口通行数据。本文基于公共自行车租赁数据研究公共交通出行模式。本文建立了时间-空间-用户属性三维张量模型,采用非负张量分解算法将张量分解为多个基本模式。在时间、空间和用户属性维度上,分别采用时间曲线图、热力图和年龄仪表图进行可视编码和展示,支持用户交互式探索不同模式间的时间、空间、用户年龄的分布情况。最后,利用纽约和芝加哥两个城市的公共自行车租赁数据案例,验证了用户行为模式分析的有效性。
[Abstract]:With the rapid development of cities, urban life becomes more diversified and complicated. People's abundant travel activities bring a lot of moving trajectory data, which contain space-time and semantic information. It is helpful to understand urban planning and people's behavior, and has become a hot research topic. In this paper, we use driving data and public transportation data to analyze the user travel patterns visually. The corresponding visual analysis technology is realized. In this paper, the driving behavior pattern is studied with sparse traffic bayonet data. The macroscopic flow exploration of traffic bayonet is realized systematically. The visual query model is designed based on the identification algorithm. The spatial bubble chart CirFlow chart and round-trip time distribution map are used to analyze the spatial and temporal distribution of the visual query results in the global single bayonet and the upper side of the bayonet. The traffic flow distribution of the bayonet is analyzed by a case study. The process of identifying the behavior of a licensed car, etc. It is verified that this method can effectively analyze the traffic data of the bayonet. Based on the public bicycle rental data, this paper studies the public transport travel mode. In this paper, a three-dimensional Zhang Liang model of time-space-user attributes is established. The non-negative Zhang Liang decomposition algorithm is used to decompose Zhang Liang into several basic patterns. In the dimension of time, space and user attributes, time curve diagram, thermal diagram and age meter chart are used to encode and display them visually. Support users to interactively explore the distribution of time, space, and user age between different models. Finally, use the public bicycle rental data from New York and Chicago. The validity of user behavior pattern analysis is verified.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP311.52;U491

【参考文献】