测绘通报 ›› 2022, Vol. 0 ›› Issue (2): 62-66.doi: 10.13474/j.cnki.11-2246.2022.0044

• 学术研究 • 上一篇    下一篇

城市尺度典型地表要素综合提取方法研究

孙琴琴1, 蔡国印1,2, 杨柳忠3, 张宁3, 杜明义1,2   

  1. 1. 北京建筑大学测绘与城市空间信息学院, 北京 100044;
    2. 北京建筑大学未来城市设计高精尖中心, 北京 100044;
    3. 住建部遥感应用中心, 北京 100835
  • 收稿日期:2021-02-22 修回日期:2021-05-21 发布日期:2022-03-11
  • 作者简介:孙琴琴(1996-),女,硕士生,主要研究方向为城市遥感。E-mail:3207326985@qq.com
  • 基金资助:
    国家重点研发计划(2017YFB0503900-4-3);北京建筑大学市属高校基本科研业务费重点科研平台学术骨干项目(X20078)

Research on urban land use extraction at the metropolitan scale

SUN Qinqin1, CAI Guoyin1,2, YANG Liuzhong3, ZHANG Ning3, DU Mingyi1,2   

  1. 1. School of Geomatics and Urban Spatial Informatics, Beijing University of Civil Engineering and Architecture, Beijing 100044, China;
    2. Beijing Advanced Innovation Center for Future Urban Design, Beijing University of Civil Engineering and Architecture, Beijing 100044, China;
    3. Remote Sensing Application Center, Ministry of Housing and Urban-Rural Development of the People's Republic of China, Beijing 100835, China
  • Received:2021-02-22 Revised:2021-05-21 Published:2022-03-11

摘要: 国产高分卫星分辨率的不断提高,使其可以从几何形态、纹理结构及光谱信息等不同侧面实现对城市地表要素的精细描述。与面向对象分类技术相比,深度学习技术的快速发展,使得城市建筑物提取的精度不断提高。然而,由于道路两旁高大建筑物及树木的遮挡,城市道路的提取精度依然有限。本文在利用卷积神经网络提取建筑物的基础上,利用OSM面状道路数据及城市边界数据,结合植被指数和水体指数,借助空间图层叠加,使得城市建筑物、道路、植被和水体提取总体精度优于90%,为国产高分影像辅助城市精细化管理和应用提供了有效解决方案。

关键词: 地表要素, GF-2, OSM, 深度学习

Abstract: With the improvement of the high resolution of domestic satellites, the detailed description of urban surface elements can be described from different aspects such as geometric form, texture structure, and spectral information. Compared with object-oriented classification method, the rapid development of deep learning technology has continuously improved the accuracy of urban building extraction. However, due to the high buildings and trees on both sides of the road, the extraction accuracy of urban roads is still limited. Based on the use of convolutional neural networks to extract buildings, this paper uses OSM polygon road data and urban boundary data, combined with normalized vegetation index and water index, and with the help of spatial layer overlay and error-tolerant processing, high-precision extraction of urban buildings, roads, vegetation and water bodies within urban built-up areas has been achieved. This study provides an effective approach in supporting for domestic high-resolution image-assisted urban fine management and application.

Key words: earth surface elements, GF-2, OSM, deep learning

中图分类号: