Batched Trajectory Compression Algorithm Based on Hierarchical Grid Coordinates

Abstract

In the visual analysis application of large-scale off-line trajectory data, users need fast user-interface response with low latency. For such demand, we established a hierarchical grid coordinate system and proposed a batched trajectory compression algorithm. We projected the original track points into the discrete grid coordinates at the corresponding layer according to the trajectory visualization needs. Then, the original trajectory points were clustered by using the approximation of transformation to form a preliminary data set of compressed trajectory points in the corresponding layer. The data set was processed by the Douglas-Poker algorithm to get the final compressed trajectory data; we also accelerated the algorithm by using GPU. The analysis and experiments show that our algorithm can quickly generate the compressed trajectory data and improve the displaying efficiency of massive trajectory data on the basis of maintaining the invariance of the trajectory data visualization.