Spatio-Temporal Vessel Trajectory Smoothing Based on Trajectory Similarity and Two-Dimensional Wavelet Transform

Abstract

Motivated by the increasing utilization of AIS-based vessel trajectory data used in both academic research and solving practical issues, we present a new method by combining trajectory similarity with two-dimensional wavelet transform to cope with the inevitable noise in the data. Using the smoothing scheme proposed, trajectory similarity is firstly introduced to cluster several sub-trajectories obtained by reasonable segmentation. In particular, Euler distance is able to act as a measure of trajectory similarity. The sub-trajectories are matched into groups according to their similarities, and two-dimensional wavelet transform is developed to conduct the decomposition and reconstruction of trajectory groups, thus achieving the denoising of them. The final high-quality vessel trajectory is obtained by extracting and combining the data from the groups. Extensive experimental results have shown that the proposed method achieves excellent effect in trajectory denoising under the premise of retaining the original characteristics of the trajectory, in terms of both qualitatively and quantitatively.