Detail-Preserving Trajectory Summarization Based On Segmentation And Group-Based Filtering

With the development of science and technology,the deployment of location acquisition equipment such as global positioning system(GPS)is growing rapidly.At present,huge trajectory data have been generated,which contain a lot of valuable information and have been used in many practical applications,such as urban computing and intelligent transportation systems.The analysis and processing of these trajectory data become the key.This paper proposes a novel trajectory summary method based on trajectory segmentation.It is a multi-angle analysis method that combines trajectory anomaly detection,trajectory clustering and trajectory segmentation.The trajectory summary framework based on segmentation in this paper includes five stages.First,trajectory anomaly detection based on the relative distance ratio is performed to remove anomaly values.Secondly,track segmentation is carried out by using information entropy to quantify the features on which segmentation is based.Third,sub-tracks are combined into sub-track similar groups by using the spatial proximity and shape constraints limited by the search window,and sub-tracks in the same group are resampled so that each sub-track in the group has the same number of sampling points.Fourth,a non-local filtering method based on wavelet transform is performed for each sub-trajectory group.Fifth,the filtered sub-trajectories from the same trajectory are connected together to give a summary result.Meanwhile,the whole process is iterated to obtain multi-granularity trajectory summary results.In this paper,through a large number of experiments,using a variety of objective evaluation indicators to analyze the data results,it is shown that the segmentation-based trajectory summary method not only has a good clustering effect,but also due to the non-segmented trajectory summary method in retaining details.