Dissertation > Excellent graduate degree dissertation topics show

The Research and Realization of the Parallel Spatial Operation in a Simple Feature Model

Author: YeZuo
Tutor: XieZhong
School: China University of Geosciences
Course: Cartography and Geographic Information Engineering
Keywords: Parallel computing GIS spatial operations Spatial data is divided Dynamic Load Balancing Polygon clipping
CLC: P208
Type: Master's thesis
Year: 2009
Downloads: 269
Quote: 4
Read: Download Dissertation

Abstract

Has long been the evolution of the architecture and application model of geographic information system (GIS) has been picked up the pace with the development of computing environments, from the host - Terminating GIS-based C / S mode to component GIS the WebGIS, GIS gradually toward more complex transition and the development of efficient computing model. As the hot issues of recent years, the field of computer research, grid computing with its high degree of sharing, flexible combination of resources and convenient architecture known as the 3rd wave of IT, GIS toward complicated the task of efficient performance of resource sharing provides the technical background and the opportunity for development. At the same time, the as a GIS soul space analysis of the of mature computing research and product Research and Implementation of High-performance GIS provides a solid foundation for operator also proposed urgent needs of high-performance environment GIS spatial computing. Parallel computing technology is a high-performance core of GIS technology aimed at by providing a parallel massive spatial data storage, parallel query, parallel search, parallel processing, and other key technologies to enhance the efficient processing and management of traditional GIS massive spatial data parallel performance parallel processing capabilities for all types of high-performance GIS applications, grid GIS provides the the theoretical reserves and core technology solutions. Parallel computing technology applied to research in the field of Spatial Information Science began in the actual 1990s, in particular aspects of remote sensing and aerial survey data, which is feasible to determine the parallel task due to the gate of the remote sensing image data format organized sex. 3D terrain display, parallel computing its efficiency for real-time three-dimensional image rendering rendering technical support. In traditional GIS field parallel spatial databases, the parallel indexing technology research focus on international application problems are mostly concerned about the location analysis, shortest path analysis. Parallel computing module as the core functionality of the GIS space research progress has been minimal, This article looks at the GIS in computing basic space parallel processing strategy and related technology, its goal is to own characteristics and spatial computing applications demand data from the vector space proposed a the massive vector space data distributed parallel spatial operations strategy and mechanism in order to achieve efficient processing of massive spatial data in distributed parallel computing environment. Text in accordance with the theory of traditional GIS system, the organization model of spatial data, spatial computing system and basic operator parallelization spatial operations related strategies designed Finally, the actual space operations, to discuss the design of simple feature model parallel spatial computing applications implementation, deployment and testing of distributed parallel computing environment. The paper first clarifies the background of knowledge, that is, parallel computing and high-performance GIS system analyzes and discusses the related progress of the thesis on the basis of clear definition and scope of, and has laid a good theoretical basis for the thesis provide a solid technical background, also makes it clear that the papers important to solve the actual problem, that simple factor model parallelization space computing core technologies and key strategies designed to achieve. The paper compared two common data organization model in GIS analysis on the feasibility and necessity of the simple factor model is applied to parallelize spatial analysis operation. To clear parallel the spatial computing system basic operator functional scope and feasibility of parallelization, papers GIS spatial analysis of the spatial relationships in the computing system operational operators and spatial analysis computing articulate operator algorithm design and function definitions. The paper core part of the parallel spatial computing strategy in a distributed environment. Spatial data partitioning strategy is the first step in spatial computing tasks in parallel, the paper analyzes the impact of spatial data by several key factors, clear a space object load metrics, the range of spatial objects, spatial range of load metrics and space data partitioning method and other critical factors in the prototype system to deploy the different characteristics of the optional implementation and a variety of ways. Overlay operations through a combination of space the actual use case of the pipeline superimposed, the superposition and block data parallel superposition of three spatial data partitioning strategy as well as the pros and cons of. Finally, the innovative debate the boundary polygons processing spatial computing parallel computing method, and comparative analysis of copy - removed method, extended boundary method, cutting - merge method characteristics of three different programs. The parallel the spatial computing tasks scheduling policy to space computing parallel processing load balancing and introduced. Task scheduling first discussed from a theoretical point of view, to evaluate the the different tasks transmission overhead way model, proposed alternative of polygons entity object ID number of polygons task passed, and then combined with the characteristics of the GIS spatial operations, respectively, describe the shared pool The three types of load balancing method, synchronous rotary distribution method and asynchronous rotation allocation method scheduling policy for the mode of operation of the GIS spatial computing tasks. In basic spatial computing strategy technical analysis based on the introduction of the parallel computing architecture based on the Master-Worker space operations, and Master-Worker parallel computing architecture workflow. Actual distributed parallel environment in order to verify the feasibility and correctness of the theoretical basis of the above research and key technologies, the paper used the vector polygon space clipping issues as the research object, discuss the parallel design of large-scale vector polygon clipping operations. Combined with the characteristics of the parallel environment, parallel computing framework was designed based on the space operations of the two-layer load balancing mechanism, shared data replication strategy, based on the load data partitioning strategy and dynamic task scheduling mechanism to achieve definition and algorithm design process from running angle operator parallelization space designed by the theoretical model to the ways and means of conversion. Gives a practical demonstration of dynamic data sharing and dynamic task scheduling strategy parallel spatial computing performance and results analysis algorithm model deployed in distributed parallel computing environment, experimental analysis on each of the core factors, granularity task block transfer . The paper studies the key technologies and strategy with generalized characteristics Simple Feature Model study parallelization strategy design and implementation of spatial computing method can be applied to various types of distributed parallel hardware and software environment for high-performance parallel GIS implementation provides the core technology system design.

Related Dissertations

CLC: > Astronomy,Earth Sciences > Surveying and Mapping > General issues > Mapping database and information system
© 2012 www.DissertationTopic.Net Mobile