测绘通报 ›› 2023, Vol. 0 ›› Issue (8): 67-71.doi: 10.13474/j.cnki.11-2246.2023.0234

• 学术研究 • 上一篇    下一篇

融合Deep-ResUnet和PS-InSAR的城市道路网形变灾害监测方法——以合肥市为例

邹鑫1,2,3, 王磊1,2,3, 李靖宇4, 滕超群1,2,3, 黄金中1,2,3, 李忠1,2,3, 李世保1,2,3   

  1. 1. 安徽理工大学空间信息与测绘工程学院, 安徽 淮南 232001;
    2. 安徽理工大学矿山采动灾害空天地协同监测与预警安徽普通高校重点实验室, 安徽 淮南 232001;
    3. 安徽理工大学矿区环境与灾害协同监测煤炭行业工程研究中心, 安徽 淮南 232001;
    4. 中国矿业大学(北京)地球科学与测绘工程学院, 北京 100083
  • 收稿日期:2023-03-01 发布日期:2023-09-01
  • 通讯作者: 王磊。E-mail:austwlei@163.com
  • 作者简介:邹鑫(1998-),男,硕士生,主要从事遥感图像智能提取方面的研究工作。E-mail:17755457268@163.com
  • 基金资助:
    国家自然科学基金(52074010);安徽省优秀青年科学基金(2108085Y20)

A method of urban road settlement monitoring combining Deep-ResUnet and PS-InSAR:a case study of Hefei city

ZOU Xin1,2,3, WANG Lei1,2,3, LI Jingyu4, TENG Chaoqun1,2,3, HUANG Jinzhong1,2,3, LI Zhong1,2,3, LI Shibao1,2,3   

  1. 1. School of Spatial Informatics and Geomatics Engineering, Anhui Universiy of Scienceand Technology, Huainan 232001, China;
    2. Key Laboratory of Aviation-aerospace-ground Cooperative Monitoring and Early Warning of Coal Mining-induced Disasters of AnhuiHigher Education Institutes, Anhui University of Science and Technology, Huainan 232001, China;
    3. Coal Industry Engineering Research Center of Mining Area Environmental and Disaster Cooperative Monitoring, Anhui University of Science and Technology, Huainan 232001, China;
    4. College of Geoscience and Surveying Engineering, China University of Mining and Technology, Beijing 100083, China
  • Received:2023-03-01 Published:2023-09-01

摘要: 针对城市道路网变形监测存在的高分辨率影像获取难、道路人工提取效率低、传统变形监测工作量大等难题,本文提出了一种基于融合Deep-ResUnet和PS-InSAR的城市道路网形变监测方法。主要思路为: 首先对目标区的哨兵-1A(Sentinel-1A)影像数据进行伪彩色变换建立道路数据集;然后训练深度残差网络(Deep-ResUnet)模型并对道路网栅格进行提取;最后利用永久散射体干涉测量(PS-InSAR)技术获取PS点形变信息与道路网栅格融合。研究结果表明,Sentinel-1A影像经过伪彩色处理后,能提高城市道路网提取的完整性,交并比提高6%~9%,道路提取精度平均提高10%左右,得到的城市道路网形变信息专题图能为城市道路变形监测和健康状况评估提供科学依据。

关键词: 永久散射体雷达干涉, 语义分割, 道路提取, 道路形变, 哨兵-1A

Abstract: In view of the problems of deformation monitoring of urban road network, such as difficulty in obtaining high-resolution images, low efficiency of manual road extraction, and heavy workload of traditional deformation monitoring, this paper proposes a deformation monitoring method of urban road network based on fusion of Deep-ResUnet and PS-InSAR. The main idea is to first perform pseudo color transformation on Sentinel-1A image data in the target area to establish a road dataset, then train a Deep-ResUnet model and extract the road network grid. Finally, the permanent scatterer interferometry (PS-InSAR) technique is used to obtain PS point deformation information and fuse it with the road network grid. The research results show that after the Sentinel-1A image is processed with pseudo color, the integrity of urban road network extraction can be improved, the intersection and merge ratio can be improved by 6%~9%, and the accuracy of road extraction can be improved by about 10% on average. The thematic map of urban road network deformation information obtained can provide scientific basis for urban road deformation monitoring and health assessment.

Key words: PS-InSAR, semantic segmentation, road extraction, road deformation, Sentinel-1A

中图分类号: