关键词: 时空多媒体分析, 时间约束, 谷歌地图, 拼车

Abstract: As the increasing number of private cars has caused serious traffic jams, carpooling as an environmentally-friendly way has become an important choice for people. In order to improve the quality of carpooling service, trajectory matching becomes a new research hotspot. Generally, there are two problems with the traditional trajectory matching based on Hausdorff distance：only the coordinate information of the point on the path is considered, and the user’s waiting time is ignored; calculating the Hausdorff distance of the entire path directly, which can’t reflect the influence of the special road segment on the matching metrics. This paper proposes two improvements for the above issues：a Hausdorff distance calculation method with time constraints is proposed; an optimized trajectory matching method is proposed：dividing the original path with turning points and refining the matching metrics to each segment. In order to evaluate the performance of the proposed framework, a large number of roadmaps in Wuhan City of the Hubei Province are downloaded from Google map. Empirical studies have shown that compared with the traditional Minimum Completion Time（MCT） online mode scheduling algorithm, the proposed method can help passengers find better matches and reduce waiting time, and thus reducing air pollution.

Key words: spatio-temporal multimedia analysis, time constraint, Google maps, carpool