Research Of Fast Electronic Map Generating Technology Based On Hadoop

With the rapid development of computer technology,there has been an explosion in the amounts of spatial data produced by satellites remote sensing and geographical mapping technology.Today electronic map is used widely because of the increasing usage of smart phones.A major need for all these applications is the ability to process such big data.So how to analyze and process spatial data fast and efficiently has become a big challenge.Prerendering,as one of the most important way to map visualization,has the advantage of imposing light burden on client software.Hadoop,as a cloud computing platform,provides a high-performance and easy-deployment scheme via Map Reduce which serves as a powerful distributed computing framework and HBase which serves as a rapid and random data accessing database.This demonstration presents fast electronic map generating technology based on Hadoop,an extensible Map Reduce-based system for processing big spatial data.It is under the precondition of a fully research on the map rendering technology and map tiles creation algorithm.This scheme has two main unique features that distinguish it from other techniques.First it uses pre-rendering technics which is friendly to a light client.Second it is capable of generating big images and then creates map tile pyramid speedily based on Map Reduce.There are three important modules in the scheme which are map style creation module,distributed map generating module and distributed map tile creation module.The map style creation phase creates different styles for different layer of map,realizes a rich diversities of the map form.The map generating phase constructs map elements and map layers,and creates partial images based on the study of space data related algorithms.The map tile creation phase produces map tiles pyramid and saves it into Hadoop directly.The specific works are as follows:(1)This thesis does thorough research on map pre-rendering technology,studies implementation details of SLD which defines the map style,and then creates different map style files for different map layers.(2)This thesis investigates the space data related algorithms,analyzes the technique to construct point,line,polygon and text label,analyzes the skill to draw different map layers,and then generates big images with high resolution based on Hadoop.Compared with singlemachine main-memory algorithms,the distributed one overcomes these limitations.(3)This thesis analyzes the Quadtree algorithm and presents map stripe algorithm,and generates the map tile pyramid according to each map layer based on Hadoop.Compared with single-machine main-memory algorithms,the distributed one shows a better performance.In summary,distributed electronic map generating technology can support the procedure form original spatial data to big images and then to map tile pyramid in a good manner.With some preliminary tests,the conclusion can be drawn that this demonstration has a better performance compared with single machine especially when the original data is extraordinary big.