Progressive spatial join for polygon data stream

TL;DR: The key idea of the method is to represent a polygon with multiple levels of detail (LODs) and process spatial join in progressive way from low to high LOD, which improves the storage utilization by filtering out non-overlapping polygons with low Lod, but also reduces the processing cost by simpler polygons of low LOD.

Abstract: Handling spatial data stream is a crucial requirement in data stream management systems (DSMSs). Especially, spatial join processing for non-point objects in DSMS is more complicated and expensive than point objects. Two critical issues arise in processing spatial join on non-point objects in DSMS. First, the size of a non-point object is larger than a point and it decreases the storage utilization of sliding windows in DSMS. It results in a low accuracy of spatial join processing. Second, the processing cost for spatial join on non-point spatial data stream increases due to the complexity of non-point objects. In this paper, we propose a method for spatial join processing on spatial data stream of polygon objects in DSMS. The key idea of our method is to represent a polygon with multiple levels of detail (LODs) and process spatial join in progressive way from low to high LOD. This strategy not only improves the storage utilization by filtering out non-overlapping polygons with low LOD, but also reduces the processing cost by simpler polygons of low LOD. Intensive experiments were carried out to analyze the performance of our methods with different parameters.