Movement Pattern Extraction Based on a Non-parameter Sub-trajectory Clustering Algorithm

He Ailin,Liu Zhong,Zhou Dechao

DOI: https://doi.org/10.1109/icbda.2019.8713239

2019-01-01

Abstract:This paper proposed a movement pattern extraction method based on trajectory clustering, including three phases: trajectory partitioning, clustering and movement extracting. Firstly, trajectories are partitioned at characteristic points whose accumulated turn angle exceeds the threshold value which is determined by Minimum Description Length (MDL) principle. Secondly, a non-parameter sub-trajectory clustering based on Density Peak Clustering (DPC) is proposed to acquire trajectory line segment clusters. Lastly, trajectory segments are used to extract the overall trajectory motion trend, that is, trajectory movement pattern. Experiment results demonstrate that our algorithm can correctly extract the movement pattern from real trajectories data.