| Trajectory data have the value of further analysis,processing and utilization because they contain a lot of spatio-temporal information.With the rapid development and popularization of location techniques,it is more easily for information collection of moving trajectory data.However,with the widespread use of trajectory acquisition terminals,the amount of raw trajectory data increases dramatically,and the redundancy is serious,which bring great pressure to the storage,transmission and further analysis of trajectory data.Therefore,it is significant to remove redundant information from trajectory data by compression.In this paper,under the premise of retaining main information,velocity,direction and position characteristics of trajectory data are fully considered to achieve multi-level ratio compression of complex trajectory data.Further,three off-line compression algorithms are proposed as follows:(1)Using thresholds of velocity change and angle,trajectories compression algorithm based on change of velocity and angle offset is suggested to reserve trajectory points,which have obvious changes in velocity and direction.(2)Covering all trajectory points with grids,calculating the time difference between the earliest points and other points in each grid,trajectories compression algorithm based on scale grid delete the trajectory points whose time difference is less than the preset time threshold.Then,the algorithm compresses points of adjacent grids through a method similar to chain encoding.(3)Using linearly varying curves to simulate the nonlinear changes of trajectory,trajectories compression algorithm based on locally weighted linear regression is presented to eliminate trajectory points deviated the original trajectory on the fitting curve with the preset distance threshold.These algorithms achieve the competitive performance compared with the state-of-the-art algorithms.The experimental results show the compression algorithms discussed in this paper are feasible and effective. |
PDF Full Text Request