Name
Abstract | ||
|---|---|---|
With more and more vehicles equipped with GPS tracking devices, there is increasing interest in building and updating maps using vehicular GPS traces. But commodity GPS devices have lower accuracy and lower sampling frequency, which made it more difficult to infer road network than most existing approaches that using highprecision and high-frequency GPS devices. As a key component of road network, intersection plays the role of transport hub. So, if the intersections are detected in advance, the road network can be then constructed conveniently by connecting the intersections. In this paper, we propose a novel algorithm for recognizing intersections with coarse-grained GPS traces based on data preprocessing and clustering. The algorithm first prune low quality GPS points, then find out the turning points around intersections and the converging points in the preprocessing step, and finally cluster these converging points to find out the cluster centers, i.e. the intersection positions. In addition, we introduce a simple road network construction algorithm based on the identified intersections. We evaluate our method using GPS data gathered from 2,827 taxis in Shenyang, Liaoning, China. Evaluation results demonstrate that our algorithm is able to find most of the road intersections effectively. |
Year | DOI | Venue |
|---|---|---|
2013 | 10.4304/jcp.8.11.2959-2965 | JOURNAL OF COMPUTERS |
Keywords | Field | DocType |
GPS traces, road network, intersection, clustering | Computer security,Computer science,Artificial intelligence,Global Positioning System,Statistics,Cluster analysis,Machine learning | Journal |
Volume | Issue | ISSN |
8 | 11 | 1796-203X |
Citations | PageRank | References |
12 | 0.64 | 9 |
Authors | ||
4 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Junwei Wu | 1 | 13 | 0.98 |
| Yunlong Zhu | 2 | 519 | 40.65 |
| Tao Ku | 3 | 18 | 1.93 |
| Liang Wang | 4 | 1567 | 158.46 |