中文摘要：

兴趣点，又称POI（points of interest）是网络地图、导航地图中重要的表达要素，包括餐饮、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使用的信息。其数据的准确性、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地图的服务质量。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，造成POI的叠置、压盖等，这一问题严重影响了用户对POI信息的查询与检索。地图综合提供了大量的算子算法以实现点或点群要素的选取，但是它们在综合效率方面亟待提高。面向矢量数据处理的并行计算，其数据划分不仅需要满足负载均衡、划分算法高效等要求，而且对于划分后各部分数据在计算前后拓扑关系的保持也显得尤为重要。兴趣点与路划网络是密切相关的要素，两者之间存在着相互依存的空间关系。本文提出基于路划网眼划分兴趣点的方法，既能保证兴趣点的划分效率，又能保证不同分区内POI选取计算的正确性。选择点选取算法中的“圆”增长算法，采用典型试验区域的路划网眼数据，基于不同节点数划分兴趣点数据，实现兴趣点选取并行计算。试验证明，该方法不仅保证了兴趣点划分的均衡性，而且可以提高兴趣点选取计算效率。通过这一研究，对面向矢量数据的地理信息分析、地图制图综合等复杂算法的并行计算具有重要意义。

英文摘要：

POl（points of interest） is an important feature of web map and navigation map, including catering, entertainment, financial institutions, tourist attractions, landmarks, gas stations, parking lots and etc., which is often used in daily life. The accuracy of the data, the abundance of the attribute information, clarity of visualization and real-time display efficiency of POI will affect the quality of map service. The current POI visualization exists many problems in the querying results when the user searches for the specific information. Due to the large amount of data, the overlay and congesting of the POI, this problem has seriously affected the user query for POI information retrieval. Map generalization provides a lot of algorithms to achieve the point or point cluster selection, but the efficiency of them needs to be improved. For parallel computing of vector data processing, the data decomposition is not only needs to satisfy the load balancing and the decomposition algorithm should be efficient but also is especially important for the topological relations maintaining between each part of the data before and after decomposition. Points of interest and stroke network is closely related features and the spatial interdependent relationship exists between the two. This paper puts forward the POI decomposition method based on the stroke mesh, which can not only ensure the decomposition efficiency but also the validity of POI selection of different part of data. This paper selected the ＂circle growth＂ algorithm, using the typical experimental regional stroke mesh data, divided the POI based on different number of nodes and realized parallel computing. The experiment proved that this method not only guarantees the equilibrium of POI decomposition, but also can improve the calculation efficiency of POI selection. The study bears substantial significance to the parallel computing of the vector data oriented geographic information analysis, map generalization and other complex algorithms for computing.