An Automatic Road Network Construction Method Using Massive GPS Trajectory Data

Automatically acquiring comprehensive, accurate, and real-time mapping information and translating this information into digital maps are challenging problems. Traditional methods are time consuming and costly because they require expensive field surveying and labor-intensive post-processing. Recent...

Full description

Saved in:
Bibliographic Details
Published inISPRS international journal of geo-information Vol. 6; no. 12; p. 400
Main Authors Zhang, Yongchuan, Liu, Jiping, Qian, Xinlin, Qiu, Agen, Zhang, Fuhao
Format Journal Article
LanguageEnglish
Published Basel MDPI AG 01.12.2017
Subjects
cartography
Construction
Construction methods
Data
Data acquisition
Data points
digital map
Digital mapping
Digital maps
Digital technology
Global positioning systems
GPS
GPS trajectory
Graph theory
Graphical representations
Labour
Mathematical models
Open systems
Positioning systems
Post-production processing
Road construction
road network construction
Roads
Satellite navigation systems
Surveying
Taxis
Traffic speed
Traffic volume
Trajectories
Online AccessGet full text

Cover

More Information
Summary:Automatically acquiring comprehensive, accurate, and real-time mapping information and translating this information into digital maps are challenging problems. Traditional methods are time consuming and costly because they require expensive field surveying and labor-intensive post-processing. Recently, the ubiquitous use of positioning technology in vehicles and other devices has produced massive amounts of trajectory data, which provide new opportunities for digital map production and updating. This paper presents an automatic method for producing road networks from raw vehicle global positioning system (GPS) trajectory data. First, raw GPS positioning data are processed to remove noise using a newly proposed algorithm employing flexible spatial, temporal, and logical constraint rules. Then, a new road network construction algorithm is used to incrementally merge trajectories into a directed graph representing a digital map. Furthermore, the average road traffic volume and speed are calculated and assigned to corresponding road segments. To evaluate the performance of the method, an experiment was conducted using 5.76 million trajectory data points from 200 taxis. The result was qualitatively compared with OpenStreetMap and quantitatively compared with two existing methods based on the F-score. The findings show that our method can automatically generate a road network representing a digital map.
ISSN:2220-9964
2220-9964
DOI:10.3390/ijgi6120400