测绘通报 ›› 2022, Vol. 0 ›› Issue (7): 93-99.doi: 10.13474/j.cnki.11-2246.2022.0210

• 学术研究 • 上一篇    下一篇

一种自身全局最优的道路网Stroke生成方法

吕峥, 孙群, 温伯威, 马京振   

  1. 信息工程大学地理空间信息学院, 河南 郑州 450001
  • 收稿日期:2021-08-17 出版日期:2022-07-25 发布日期:2022-07-28
  • 作者简介:吕峥(1996—),男,博士生,主要从事多源矢量数据融合及制图综合研究。E-mail:lvzheng_xd@163.com
  • 基金资助:
    河南省中原学者资助项目(202101510001);智慧中原地理信息技术河南省协同创新中心和时空感知与智能处理自然资源部重点实验室基金(212102)

A self-global-best-fit road network Stroke generation method

Lü Zheng, SUN Qun, WEN Bowei, MA Jingzhen   

  1. Institute of Geospatial Information, Information Engineering University, Zhengzhou 450001, China
  • Received:2021-08-17 Online:2022-07-25 Published:2022-07-28

摘要: Stroke在道路网研究中发挥着重要作用,目前Stroke连接策略大多限于局部,对Stroke的全局表现考虑不足。针对以上不足,本文提出了一种自身全局最优的道路网Stroke生成方法,以起始路段为根结点,以连接规则为约束构建Stroke树,并通过随机森林算法投票选出每棵树的最优Stroke。经试验验证,自身全局最优的道路网Stroke生成方法在视觉认知和网络功能方面的表现均优于自身适合策略、自身最大适合策略及每对最大适合策略。

关键词: 道路网, 自身全局最优, 随机森林, Stroke树

Abstract: Stroke plays an important role in road network research. At present, most stroke connection strategies focus on the local and don't consider global performance. The paper presents a self-global-best-fit strategy for generating road network Stroke. Firstly, the Stroke tree is built under the constraints of connection rules with the selected road as the root node. Then, the best Stroke of each tree is selected by the random forest algorithm. The results show that the performance of the self-global-best-fit strategy is better than the self-fit strategy, self-best-fit strategy, and every-best-fit strategy in visual cognition and network function.

Key words: road network, self-global-best-fit, random forest, Stroke tree

中图分类号: