Generating urban road intersection models from low-frequency GPS trajectory data

ABSTRACT

Detailed real-time road data are an important prerequisite for navigation and intelligent transportation systems. As accident-prone areas, road intersections play a critical role in route guidance and traffic management. Ubiquitous trajectory data have led to a recent surge in road map reconstruction. However, it is still challenging to automatically generate detailed structural models for road intersections, especially from low-frequency trajectory data. We propose a novel three-step approach to extract the structural and semantic information of road intersections from low-frequency trajectories. The spatial coverage of road intersections is first detected based on hotspot analysis and triangulation-based point clustering. Next, an improved hierarchical trajectory clustering algorithm is designed to adaptively extract the turning modes and traffic rules of road intersections. Finally, structural models are generated via K-segment fitting and common subsequence merging. Experimental results demonstrate that the proposed method can efficiently handle low-frequency, unstable trajectory data and accurately extract the structural and semantic features of road intersections. Therefore, the proposed method provides a promising solution for enriching and updating routable road data.

KEYWORDS:

• Low-frequency trajectory data
• road intersections
• spatial clustering
• principle curve fitting
• traffic rules inference

Acknowledgments

We would like to thank the editors and the anonymous reviewers for their valuable comments and Yunfei Zhang owes special thanks to Prof. Bisheng Yang at Wuhan University.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by the National Natural Science Foundation of China [41601495, 41730105, 41771492, 41501442]; the National Key Research and Development Program of China [2017YFB0503500]; Natural Science Foundation of Hunan Province [2018JJ3525]; Open Research Fund of State Key Laboratory Of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University [17S01]; Open Research Fund of State Key Laboratory of Resources and Environmental Information System and Scientific Research Foundation of Hunan Education Department [17B004].

Notes on contributors

Min Deng

Min Deng is currently a professor with Central South University and Dean of Geo-informatics department. His research interests are map generalization, spatio-temporal data analysis and mining.

Jincai Huang

Jincai Huang is a Ph.D. student at Central South University and his research interest focus on information extraction from trajectory data.

Yunfei Zhang

Yunfei Zhang receives the Ph.D. degree in Wuhan University. She is currently a lecturer with Changsha University of Science & Technology and her research interests include spatial data matching, digital map conflation, and trajectory data analysis.

Huimin Liu

Huimin Liu is an associate professor with Central South University and works in the area of map generalization and multiple representation.

Luliang Tang

Luliang Tang is a professor with Wuhan University and works in the area of space time GIS, GIS for transportation and change detection.

Jianbo Tang

Jianbo Tang receives the Ph. D. degree in Central South University and works in the area of road data updating.

Xuexi Yang

Xuexi Yang is a Ph. D. student at Central South University and works in the area of spatio-temporal data mining.