测绘学报 ›› 2022, Vol. 51 ›› Issue (1): 104-114.doi: 10.11947/j.AGCS.2022.20200241

• 地图学与地理信息 • 上一篇    下一篇

三维Hilbert曲线层级演进模型与编码计算

吴宇豪, 曹雪峰, 余岸竹, 孙万忠   

  1. 信息工程大学地理空间信息学院, 河南 郑州 450001
  • 收稿日期:2020-06-13 修回日期:2021-03-16 发布日期:2022-02-15
  • 通讯作者: 曹雪峰 E-mail:CAO_Xue_Feng@163.com
  • 作者简介:吴宇豪(1989-),男,硕士,助理研究员,研究方向为全球离散网格。E-mail:514536575@qq.com
  • 基金资助:
    国家自然科学基金(41401465)

Hierarchical evolution model and coding calculation of three-dimensional Hilbert curve

WU Yuhao, CAO Xuefeng, YU Anzhu, SUN Wanzhong   

  1. Institute of Geospatial Information, Information Engineering University, Zhengzhou 450001, China
  • Received:2020-06-13 Revised:2021-03-16 Published:2022-02-15
  • Supported by:
    The National Natural Science Foundation of China (No. 41401465)

摘要: 格网单元的编码计算是全球离散网格系统的核心,支撑着网格快速索引及应用分析的高效计算。Hilbert曲线具有聚簇性高、连续性强的特点,是研究设计全球离散网格编码的重要工具。利用Hilbert曲线进行格元编码实现了坐标等效降维表达,但是对Hilbert曲线不同层级之间的变换关系、一维Hilbert码如何刻画格元多维空间结构与关系等网格编码基础理论问题的研究尚不完备。本文以八叉树立体网格中三维Hilbert曲线层级演进关系为突破口,使用状态矩阵与演进矩阵构建层级演进模型,进而分别设计笛卡儿坐标至Hilbert码计算以及邻近格元Hilbert码计算方法。与现有算法对比,本文算法以层级演进模型为理论基础,避免了烦琐迭代步骤以及转换步骤,算法流程简明直接。试验结果表明,本文笛卡儿坐标至Hilbert码计算效率较迭代算法提高为7%~23%,邻近格元Hilbert码计算效率较转换算法提高4.0~4.5倍。

关键词: 全球离散网格, Hilbert曲线, 编码计算, 邻近

Abstract: The coding calculation of grid cells is the core of the discrete global grid system, which supports efficient calculation of indexing and analysis. Hilbert curve has the characteristics of high clustering and strong continuity, and is an important tool for global discrete grid coding. The lattice element coding using Hilbert curve realizes the coordinate equivalent dimensionality reduction expression, but the research on the basic theoretical problems of grid coding still not complete. In this paper, the hierarchical evolution relationship of the three-dimensional Hilbert curve in the octree three-dimensional grid is used as a breakthrough. The state matrix and the evolution matrix are used to construct a hierarchical evolution model, and then the calculation methods of Cartesian coordinates to Hilbert codes and adjacent grid Hilbert codes are designed respectively. Compared with the existing algorithms, the algorithm in this paper is based on a hierarchical evolution model, which avoids cumbersome iteration steps and conversion steps, and the algorithm flow is simple and straightforward. The experimental results show that the calculation efficiency of Cartesian coordinates to Hilbert codes in this paper is "7%~23%" higher than that of the iterative algorithm, and the calculation efficiency of neighbor grid elements Hilbert codes is 4.0~4.5 times higher than that of the conversion algorithm.

Key words: discrete global grid, Hilbert curve, coding calculation, neighbor

中图分类号: