交通高维数据逻辑整合与降解研究

发布时间：2018-06-26 23:45

本文选题：交通 + 高维数据　；参考：《重庆交通大学》2015年硕士论文

【摘要】：一直以来,交通数据管理都是一个严重的社会问题,由于交通管理部门建立的信息系统大多处于“孤岛型”运作,导致交通信息零散、数据之间缺乏必要的内在联系等问题的大量存在,从而给交通信息的应用造成了极大的不便。因此,对高维交通数据进行逻辑整合及降解研究对交通数据的管理和应用具有重大的实际意义。本文首先对各种交通检测数据进行了有关处理,先对各数据进行了时空规范化处理,具体给出了路网上检测点位的编码方法;然后对规范的交通数据进行有效性判别,对无效的故障数据提出了具体的修复方法;最后对由于检测设备具有不同检测精度引起的交通数据冲突予以识别和消减。其次,本文以云计算技术为支撑,借助于分布式存储的思想,通过对交通数据进行详细分类,并对各类交通数据进行了维度分析,同时给出了各类交通数据的二维表结构。基于分布式存储的理念构建了以Hadoop为平台、以Hbase分布式数据库为存储工具的交通数据存储体系,并对各类交通数据创建了基于IP地址查询的多级索引,同时对存储在Hbase中的数据表赋予了相应的IP,在此基础上实现交通高维数据逐层存储的目标。这不仅实现了交通高维数据的逻辑整合,也为交通数据的查询或统计分析提供了方便。针对交通数据的检索,鉴于交通数据具有明显的时空特征,首先将所查数据的空间坐标与GIS图进行映射,得到与该空间位置对应的路网检测点位,然后对检测点位上所包含的交通信息进行检索,通过逐层遍历,得到存储该检测点位上所需信息的IP地址,通过访问IP地址,得到二维表,最后通过输入查询条件,完成对二维表的查询,得到所查交通数据。论文中的方法基于云平台,减少了因建立数据中心的巨大花费,同时也能较好的消除系统间的信息孤岛现象,使得信息能够更加高效的互通、共享。不仅为交通信息使用者提供了方便,也为交通管理者提供了决策依据,对交通数据的管理和应用具有一定的实用性和应用价值。
[Abstract]:Traffic data management has always been a serious social problem. Because most of the information systems set up by traffic management departments operate in "isolated islands", traffic information is scattered. There are a lot of problems such as the lack of necessary internal connection between the data, which cause great inconvenience to the application of traffic information. Therefore, the research on logical integration and degradation of high-dimensional traffic data is of great practical significance to the management and application of traffic data. In this paper, we first deal with all kinds of traffic detection data, first, we normalize the data in time and space, and then give the coding method of detecting points on the road network, and then we judge the validity of the standard traffic data. Finally, the traffic data conflict caused by the different detection accuracy of the detection equipment is identified and reduced. Secondly, with the support of cloud computing technology and the idea of distributed storage, the traffic data are classified in detail, and the dimension of traffic data is analyzed, and the two-dimensional structure of traffic data is given. Based on the idea of distributed storage, a traffic data storage system based on Hadoop and Hbase distributed database is constructed, and a multi-level index based on IP address query is created for all kinds of traffic data. At the same time, the corresponding IPs are given to the data table stored in Hbase, and the goal of storing traffic high-dimensional data layer by layer is realized on this basis. This not only realizes the logical integration of high dimensional traffic data, but also provides convenience for the query or statistical analysis of traffic data. For the traffic data retrieval, in view of the obvious space-time characteristics of the traffic data, the spatial coordinates of the data are mapped to the GIS map, and the road network detection points corresponding to the spatial location are obtained. Then the traffic information contained on the detection point is retrieved, and the IP address of the information needed on the detection point is obtained by traversing it layer by layer, and the two-dimensional table is obtained by visiting the IP address, and finally the query condition is input. Complete the query of the two-dimensional table and get the traffic data. The method in this paper is based on cloud platform, which reduces the huge cost of establishing data center, and can eliminate the phenomenon of information isolation between systems, so that the information can be exchanged and shared more efficiently. It not only provides convenience for users of traffic information, but also provides decision basis for traffic managers. It has certain practicability and application value for traffic data management and application.
【学位授予单位】：重庆交通大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U491

【参考文献】