The Research Of Surface Area Algorithm And Distributed Parallel Processing Based On Statistics Of National Geographic Conditions

The census of geographical conditions requires rapid and accurate statistical analysis of the census results.And basic statistics is an important content of national geographic conditions of the census,in these statistics,the calculation of surface coverage of the surface area is the most time-consuming.It needs to deal with the largest amount of data and has the most intermediate data,so we propose to optimize the statistical model of surface area.Because of the topographic re lief it is difficult to fit the mathematical function directly.Now,there are many methods to calculate the surface area of three dimensions,but these methods are too complicated to deal with the cell grid.There are too many judgment conditions,thus,it takes a great deal of time to calculate the huge amount of data.On the basis of comparing the various methods,we put forward a simple method to calculate surface area and steps to improve the computational efficiency.With the development of new hardware architecture(single-machine multi-core,cluster,etc.),traditional data processing technology and serial computing technology cannot meet the demand of massive data processing,so we need to make full use of computing resources to improve the actual efficiency of data processing.We combine the dynamic development of distributed system infrastructure(Hadoop platform)to build clusters,and realize the distributed parallel computing of geographical conditions.This paper discussed all kinds of ground surface area algorithm,and proposed a new method for calculating the surface area based on the characteristics of the knowledge of computer graphics.The author give the concrete calculation steps and implemented the programming.At the same time,we installed the open source Hadoop platform and set up a small cluster for distributed parallel computing.We designed and relized the HDFS and map-reduce program respectively.Finally we selected dongying city of shandong province of geographic conditions survey pilot area to experiment.We give detailed task partitioning strategy,and the four factors that influence efficiency of cluster computing,and put forward feasible Suggestions to the overall optimization model.