POI简化并行计算方法研究

发布时间：2018-03-23 09:10

本文选题：POI　切入点：地图综合　出处：《南京师范大学》2013年硕士论文　论文类型：学位论文

【摘要】：随着移动互联网的发展和智能设备的普及,用户越来越多的关注自己的位置信息,实时准确的位置信息正在成为用户最渴求的信息之一,用户在关注其位置信息的同时,也关注其周围的其它基础地理信息,各种基于LBS (Location Based Service)的应用软件广受用户的欢迎,手机地图也成为智能手机的必备软件,各IT信息服务商,也都推出其地图在线搜索服务。POI (Points of Interest),兴趣点,涵盖了餐饮、娱乐、金融机构、旅游景点、地标建筑、加油站、停车场等人们日常生活中最为经常的基础信息,POI数据的准确性,属性的丰富程度、表达的清晰度及其实时显示效率都将影响移动地图的质量与可用性效果。但当前POI表达存在许多问题,特别是在用户搜索特定信息时,由于查询结果数据量较大,会产造成某几个区域POI的重叠、压盖等,进而影响用户的对地理空间的认知,降低地图的可用性。针对此问题,本文提出了面向POI简化的并行算法。本文的研究内容与成果主要包括以下几部分：1、分析了当前网络地图中POI的表达特点,总结了当前POI表达常用的方法；结合点要素综合质量评价指标,提出了描述POI基本信息的指数与POI简化的质量评价指标。分析了点要素简化算法的效率与质量。结合POI对简化算法的需求,遴选了适合POI简化的算法。2、总结了常用的并行方法设计方法,改进了“圆”增长算法、居民地空间比率算法。提出“圆”增长算法、居民地空间比率算法对数据划分方法的要求,结合POI与道路的关系,提出基于道路网眼层次结构分解的POI简化并行计算方法。3、设计了POI简化并行的实验平台,基于此平台编程实现了基于道路网眼层次结构分解的POI简化并行算法,以及改进后“圆”增长算法、居民地空间比率算法的任务并行计算方法。使用不同规模的POI数据与道路数据对算法进行了测试、统计与分析,分析了算法的执行效率与综合质量,结果证明该数据并行计算方法满足了POI简化对效率与质量的需求。
[Abstract]:With the development of mobile Internet and the popularity of smart devices, users pay more and more attention to their location information. Real-time and accurate location information is becoming one of the most desired information. They also pay attention to other basic geographic information around them. Various applications based on LBS location Based Service are popular with users, and mobile maps have become the necessary software for smartphones, and various IT information service providers. They also launched their online map search service. POI points of interest, which covers restaurants, entertainment, financial institutions, tourist attractions, landmarks, gas stations, parking lots, and other basic information in people's daily lives, and the accuracy of POI data. The richness of attributes, clarity of expression and efficiency of real-time display will affect the quality and usability of mobile maps. However, there are many problems in POI expression, especially when users search for specific information. Because of the large amount of data of query results, the overlap and cover of POI in certain regions will be produced, which will affect the users' cognition of geographical space and reduce the usability of maps. In this paper, a parallel algorithm for POI simplification is proposed. The main contents and achievements of this paper include the following parts: 1, analyzes the characteristics of POI expression in current network maps, and summarizes the common methods of POI expression; The index describing the basic information of POI and the quality evaluation index of POI simplification are put forward. The efficiency and quality of the point element simplification algorithm are analyzed. Combined with the requirement of POI to simplify the algorithm, This paper selects the algorithm .2which is suitable for POI simplification, summarizes the design methods of common parallel methods, improves the "circle" growth algorithm, and proposes the "circle" growth algorithm. Considering the relationship between POI and road, this paper presents a simplified parallel computing method of POI based on decomposition of road mesh hierarchy, and designs an experimental platform for simplifying parallelism of POI. Based on this platform, a simplified parallel POI algorithm based on hierarchical decomposition of road mesh and an improved "circle" growth algorithm are implemented. Using different scale POI data and road data to test, statistics and analysis of the algorithm, the efficiency and comprehensive quality of the algorithm are analyzed. The results show that the parallel computing method meets the requirements of POI simplification for efficiency and quality.
【学位授予单位】：南京师范大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP338.6

【相似文献】