Improving the performance of GIS polygon overlay computation with MapReduce for spatial big data processing

As one of the important operations in Geographic Information System (GIS) spatial analysis, polygon overlay processing is a time-consuming task in many big data cases. In this paper, a specially designed MapReduce algorithm with grid index is proposed to decrease the running time. Our proposed algorithm can reduce the times of calling intersection computation by the aid of grid index. The experiment is carried out on the cloud framework based on Hadoop built by ourselves. Experimental results show that our algorithm with spatial grid index consumes less time than its peer without spatial index. Moreover, the proposed algorithm has an upward speed-up ratio when more nodes of Hadoop framework are used. Nevertheless, with the increase of nodes, the upward trend of speed-up ratio slows down.