Trajectory Voting and Classification Based on Spatiotemporal Similarity in Moving Object Databases

We propose a method for trajectory classification based on trajectory voting in Moving Object Databases (MOD). Trajectory voting is performed based on local trajectory similarity. This is a relatively new topic in the spatial and spatiotemporal database literature with a variety of applications like trajectory summarization, classification, searching and retrieval. In this work, we have used moving object databases in space, acquiring spatiotemporal 3-D trajectories, consisting of the 2-D geographic location and the 1-D time information. Each trajectory is modelled by sequential 3-D line segments. The global voting method is applied for each segment of the trajectory, forming a local trajectory descriptor. By the analysis of this descriptor the representative paths of the trajectory can be detected, that can be used to visualize a MOD. Our experimental results verify that the proposed method efficiently classifies trajectories and their sub-trajectories based on a robust voting method.