基于SAR图像变化检测的滑坡提取研究
DOI:
CSTR:
作者:
作者单位:

作者简介:

通讯作者:

中图分类号:

P315

基金项目:

国家自然科学基金(42371404、41601479)、甘肃省科技计划项目自然科学基金面上项目(22JR5RA822)、甘肃省地震局创新团队专项(2019TD-01-02)、中国地震局地震预测研究所基本科研业务费专项(2021IESLZ4)、中国地震局地震应急与信息青年重点任务(CEAEDEM202322)共同资助


Research on Landslide Extraction Based on Change Detection of SAR Image
Author:
Affiliation:

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    合成孔径雷达(SAR)作为一种先进的遥感技术,广泛应用于环境监测、资源调查和灾害评估等多个领域,并取得了显著的成效。SAR影像变化检测技术凭借高分辨率回波信号对地物变化敏感的优势,通过分析地震灾害滑坡前后多时相强度特征,可以更快速地获得全面和准确的评估响应。然而,由于SAR图像中噪声成分丰富且有效信息提取不足,从而削弱了图像变化检测的准确性。为进一步提高变化检测的准确性,本文基于不同时相极化SAR强度信息的变化构建差异图,采用自适应高斯阈值算法计算出最优阈值,以多数投票法提取初步滑坡候选区,同时融合数字高程模型(DEM)计算的坡度值作为地形约束,有效滤除平原区伪变化信号,最终提取出滑坡范围。实验结果表明,该研究方法在复杂地形条件下检测滑坡展现出显著优势,模型能够成功识别出大多数滑坡事件,总体精度达90.7%,表现出较好的稳健性,为震后灾害评估和应急响应提供了可靠的技术支持。

    Abstract:

    Synthetic Aperture Radar(SAR),as an advanced remote sensing technology,is widely used in environmental monitoring,resource investigation,and disaster assessment,where it has demonstrated remarkable effectiveness. SAR image change detection is of great significance because it enables rapid and comprehensive damage assessment by analyzing multi-temporal intensity images acquired before and after earthquake-induced landslides. However,excessive noise and insufficient exploitation of image information in SAR data can adversely affect the accuracy of change detection. To further improve detection accuracy,this study constructs difference images based on variations in SAR intensity and applies an adaptive Gaussian thresholding algorithm to determine the optimal threshold. A majority voting method is then employed to extract preliminary landslide candidate areas,which are further refined by incorporating slope values derived from a Digital Elevation Model(DEM)as terrain constraints. Experimental results demonstrate the strong potential of this approach for landslide detection. Under complex terrain conditions,the proposed model successfully identified most landslide events when validated against field-based landslide inventory data,achieving an overall accuracy of 90.7% and showing robust performance. This method provides reliable technical support for post-earthquake disaster assessment and emergency response.

    参考文献
    相似文献
    引证文献
引用本文

杜建清,翟玮,杨刚玉.基于SAR图像变化检测的滑坡提取研究[J].中国地震,2026,42(1):73-89

复制
分享
相关视频

文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2025-04-24
  • 最后修改日期:2025-08-14
  • 录用日期:
  • 在线发布日期: 2026-05-28
  • 出版日期:
文章二维码
您是第位访问者
中国地震 ® 2026 版权所有
技术支持:北京勤云科技发展有限公司