地球信息科学学报

地球信息科学学报 ›› 2011, Vol. 13 ›› Issue (5): 701-706.doi: 10.3724/SP.J.1047.2011.00701

• 模型与算法应用 • 上一篇    下一篇

线状空间数据传输的几何相似性度量算法与实验分析

孙金礼1,2, 陈杰2, 邓敏3   

  1. 1. 山西煤炭运销集团大同有限公司, 大同 037004;
    2. 山西大同大学矿业工程系, 大同 037003;
    3. 中南大学测绘与国土信息工程系, 长沙 410083
  • 收稿日期:2011-05-03 修回日期:2011-08-31 出版日期:2011-10-25 发布日期:2011-10-25
  • 作者简介:孙金礼(1964-),男,汉族,山西原平人,硕士,副教授,现主要从事空间地物相似性度量和地理空间数据传输理论与方法。E-mail: chenjie_301@126.com

The Algorithms of Geometry Similarity Measurement and Experimental Analysis for Linear Spatial Data Transmission

SUN Jinli1,2, CHEN Jie2, DENG Min3   

  1. 1. Shanxi Coal Tonsportation and sales Group Datong Co., LTD. Datong 037004, China;
    2. Department of Mining Engineering, University of Shanxi Datong, Datong 037003, China;
    3. Department of Surveying and Geo-informatics, Central South University, Changsha 410083, China
  • Received:2011-05-03 Revised:2011-08-31 Online:2011-10-25 Published:2011-10-25

PDF (PC)

赞

被引次数

7

摘要： 同一地物在不同比例尺或者不同来源的地图上通常存在着相似性,对于图形几何形似性度量方法的研究有利于地图编制、查询、匹配、更新。线状地物要素在地图中占有很大的比例,因此,本文以线状空间目标为例,在前人的基础上给出了线状空间数据的几何图形相似性度量模型:(1)以差异距离作为相似性特征的位置相似度;(2)以分形维数作为相似性特征的形状相似度;(3)以长度或者周长作为相似性特征的大小相似度。相对于(3)大小相似度而言,(1)位置相似度、(2)形状相似度综合考虑了几何图形整体统计的方法和局部几何特征结构。完成多尺度传输的线状空间数据几何相似性度量实验,并对数据传输量与几何相似性度量方法进行比较,实验结果表明:基于广义Hausdorff距离模型的中位数Hausdorff距离的位置相似性对于空间数据渐进性传输具有稳定性和可行性。最后,总结了本文的研究成果,并展望了该方向进一步研究的若干问题。

关键词: 几何相似性, 分形维数, 广义Hausdorff距离

Abstract: Transmission and expression of multi-scale spatial data is of greatly significance for the quality of the evaluation of progressive transmission of spatial information by geometric similarity measurement, which has plagued GIS researchers for a long time. The same feature from different sources or different scales is usually similar on the map, so the similarity measurement of geometry shape is conducive to the preparation, query, match and update of the map. The line features on the map are in many high proportions, so this paper proposes the line features geometry similarity measurement models of spatial data on the basis of the predecessors. And then, a reclassification of geometric similarity measurement is made from the consideration of distance, shape, length. The corresponding geometric similarity measurements are further developed, including: (1) differences in distance as the similarity, considering shape differences and relative positions of spatial objects; (2) the fractal dimension as a similarity of shape similarity, to a large extent, being able to express morphological characteristics of spatial objects; (3) the length or girth size as the similarity. Relative to (3) length similarity, (1) position similarity and (2) shape similarity are both considered the overall statistical methods and local geometry structure. To complete the linear multi-scale geometric similarity measurement experiments of spatial data, and compare the relationship of different transmission volume and geometric similarity measurement by multi-scale linear features as the experimental data, the experimental results show that location similarity based on generalized Hausdorff distance model is feasibility to progressive transmission of spatial data. The change for location similarity measurement is smaller than those with other methods based on the same amounts of data transmission, indicating that the algorithm (generalized Hausdorff distance model) solves the similarity problem of the same feature. Finally, this paper summarizes main findings, and highlights further research directions in the near future, such as match, query, update and other issues for multi-scale spatial data.

Key words: generalized Hausdorff distance, geometric similarity, fractal dimension