本文選題：交通軌跡數(shù)據(jù)　切入點：數(shù)據(jù)處理　出處：《蘭州交通大學》2017年碩士論文

【摘要】：隨著社會的發(fā)展,城市化進程的加快,急劇增多的交通出行導致城市交通擁堵現(xiàn)象日益嚴重。伴隨著相關技術的快速發(fā)展,包括浮動車數(shù)據(jù)在內(nèi)越來越多的交通數(shù)據(jù)被收集存儲。交通數(shù)據(jù)具有海量、維度高,包含時間與空間標簽等特點,是典型的時空數(shù)據(jù),需要合理、高效的分析手段支持才能發(fā)現(xiàn)其背后隱含的規(guī)律。智能交通的興起加速了交通數(shù)據(jù)研究領域的發(fā)展,可視化作為數(shù)據(jù)分析的重要手段逐漸被應用到交通數(shù)據(jù)研究領域,可直觀呈現(xiàn)多維時空軌跡數(shù)據(jù),探索發(fā)現(xiàn)城市交通運行規(guī)律,為城市交通調(diào)度和道路建設規(guī)劃提供決策支持�，F(xiàn)有數(shù)據(jù)可視化解決方案在可視分析交通軌跡數(shù)據(jù)時存在對技術要求、通用性及可視化效果不能同時兼顧的問題�；诖�,本文設計并實現(xiàn)交通時空數(shù)據(jù)可視化平臺,通過統(tǒng)一的數(shù)據(jù)模式,對多源數(shù)據(jù)集成、轉(zhuǎn)換、數(shù)據(jù)清洗及地圖匹配,并完成基于交互的任務邏輯設計,平臺實現(xiàn)在提高通用性同時可對時空數(shù)據(jù)分析提供多樣化可視化方法。論文的主要內(nèi)容如下:本文給出數(shù)據(jù)處理方法。地圖數(shù)據(jù)是地圖可視化的基礎,本文采用ArcGIS Desktop完成原始地圖數(shù)據(jù)處理過程,為后期分析任務提供可用的地圖數(shù)據(jù);原始GPS數(shù)據(jù)存在噪聲數(shù)據(jù),根據(jù)GPS數(shù)據(jù)的特點制定數(shù)據(jù)的標準表示方法及數(shù)據(jù)過濾機制,將原始數(shù)據(jù)統(tǒng)一化表示并對噪聲等問題數(shù)據(jù)清洗過濾;定位位置與實際運行位置間存在偏差,地圖匹配可有效解決此問題,傳統(tǒng)匹配算法應用到低頻采樣浮動車數(shù)據(jù)時精度較低,基于此本文提出了一種改進的地圖匹配算法,采用改進的地圖網(wǎng)格劃分方法快速確定候選路段集,引入最短路徑與軌跡方向信息輔助確定最優(yōu)匹配路段。結(jié)合數(shù)據(jù)處理過程及時空軌跡數(shù)據(jù)特點,從功能及屬性兩個維度分析了平臺的需求,確定了不同層次的任務分析模型,采用分層的架構思想設計平臺架構,結(jié)合需求分析等給出數(shù)據(jù)導入模塊、數(shù)據(jù)處理模塊、數(shù)據(jù)可視化模塊及功能拓展模塊四個功能模塊的詳細設計,為平臺實現(xiàn)奠定基礎�；谠O計基礎,實現(xiàn)平臺各功能模塊功能,包括數(shù)據(jù)導入模塊、數(shù)據(jù)處理模塊、數(shù)據(jù)可視化模塊及功能拓展模塊各模塊功能,并對系統(tǒng)性能進行測試分析,滿足設計要求。
[Abstract]:With the development of society and the acceleration of urbanization, the phenomenon of urban traffic congestion is becoming more and more serious.With the rapid development of related technology, more and more traffic data, including floating vehicle data, are collected and stored.Traffic data is a typical spatio-temporal data with the characteristics of large volume, high dimension, including time and space tags, and it needs reasonable and efficient analysis means to find the hidden law behind it.The rise of intelligent transportation accelerates the development of traffic data research field. As an important means of data analysis, visualization is gradually applied to the field of traffic data research, which can visually present multidimensional space-time trajectory data.This paper explores and finds out the rules of urban traffic operation and provides decision support for urban traffic scheduling and road construction planning.The existing data visualization solutions have some problems such as the technical requirements, generality and visualization effect when analyzing the traffic track data visually.Based on this, this paper designs and implements the traffic space-time data visualization platform, through the unified data pattern, the multi-source data integration, the transformation, the data cleaning and the map matching, and completes the interactive task logic design.The platform can provide a variety of visualization methods for spatiotemporal data analysis.The main contents of this paper are as follows: this paper gives the method of data processing.Map data is the basis of map visualization. In this paper, ArcGIS Desktop is used to complete the processing process of original map data, which provides the available map data for the later analysis task, and the original GPS data has noise data.According to the characteristics of GPS data, the standard representation method of data and the mechanism of data filtering are established, the original data is unified and the noise data is cleaned and filtered, and there is a deviation between the location position and the actual running position.Map matching can solve this problem effectively. The accuracy of traditional matching algorithm is low when it is applied to the low-frequency sampling floating vehicle data. Based on this, an improved map matching algorithm is proposed in this paper.An improved map grid method is used to quickly determine the candidate section set, and the shortest path and trajectory direction information are introduced to help determine the optimal matching section.According to the characteristics of data processing process and space-time track data, this paper analyzes the demand of the platform from the two dimensions of function and attribute, determines the task analysis model of different levels, and designs the platform architecture with the idea of layered architecture.The detailed design of four functional modules, such as data import module, data processing module, data visualization module and function extension module, is given in combination with requirement analysis, which lays a foundation for the realization of the platform.Based on the design foundation, the functions of each functional module of the platform are realized, including the data import module, the data processing module, the data visualization module and the function expansion module. The system performance is tested and analyzed to meet the design requirements.
【學位授予單位】：蘭州交通大學
【學位級別】：碩士
【學位授予年份】：2017
【分類號】：TP311.52;U491

【參考文獻】

相關期刊論文 前10條

1 毛嘉莉;金澈清;章志剛;周傲英;;軌跡大數(shù)據(jù)異常檢測:研究進展及系統(tǒng)框架[J];軟件學報;2017年01期

2 楊旭華;彭朋;;基于條件隨機場和低采樣率浮動車數(shù)據(jù)的地圖匹配算法[J];計算機科學;2016年S1期

3 孫健;張穎;張純;;基于駕駛?cè)寺窂竭x擇偏好的OD行程時間預測方法[J];交通運輸工程學報;2016年02期

4 曾U，

本文編號：1709670