Scholarlink

<p>Urban seismic resilience assessment is essential for spatial risk management and sustainable urban planning. However, existing studies are constrained by incomplete indicator systems, insufficient spatial information from Geographic Information Systems (GIS) and remote sensing, and limited capacity to capture the complex, nonlinear relationships among multiple factors and resilience outcomes. This study aims to address these gaps by proposing a spatially explicit evaluation framework for urban site seismic resilience at the city scale. A comprehensive four-level evaluation index system was constructed by integrating remote sensing data, GIS spatial layers, engineering geology information, and seismic hazard data, in accordance with 12 national and industry standards. Eighteen core indicators were systematically selected and quantified, covering site conditions, seismic response characteristics, secondary-disaster susceptibility, human-engineering disturbance, and post-earthquake recovery capacity. An improved entropy method was employed to determine objective indicator weights, which were then incorporated into a Deep Belief Network (DBN) model to enhance the nonlinear mapping from spatial indicators to seismic resilience levels. A GIS-based evaluation system was developed to automate data management, spatial analysis, resilience computation, and the generation of visual outputs. Case studies on six representative Chinese cities demonstrated that the proposed framework effectively captured spatial heterogeneity in urban seismic resilience, with comprehensive scores ranging from 48.2 to 82.3. The improved entropy-DBN model outperformed conventional methods in both accuracy and convergence efficiency, reducing MAE, MSE, and RMSE to below 0.1 after 300 iterations. The system processed a single site in 12.28s and batch-processed ten sites in 42.75s, with core factor identification accuracy reaching 98.5%. Integrating GIS- and remote sensing-derived spatial information with advanced machine learning provides robust decision support for urban seismic risk mitigation, land-use planning, and disaster prevention at the regional scale. The proposed framework offers a transferable methodology that can be adapted to other regions.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Spatial evaluation of urban seismic resilience using GIS and remote sensing

  • 期刊论文

  • 发表日期：

    2026年8月3日

查看全文
  • Jian Xi,
  • Xinyu Lu

摘要

Urban seismic resilience assessment is essential for spatial risk management and sustainable urban planning. However, existing studies are constrained by incomplete indicator systems, insufficient spatial information from Geographic Information Systems (GIS) and remote sensing, and limited capacity to capture the complex, nonlinear relationships among multiple factors and resilience outcomes. This study aims to address these gaps by proposing a spatially explicit evaluation framework for urban site seismic resilience at the city scale. A comprehensive four-level evaluation index system was constructed by integrating remote sensing data, GIS spatial layers, engineering geology information, and seismic hazard data, in accordance with 12 national and industry standards. Eighteen core indicators were systematically selected and quantified, covering site conditions, seismic response characteristics, secondary-disaster susceptibility, human-engineering disturbance, and post-earthquake recovery capacity. An improved entropy method was employed to determine objective indicator weights, which were then incorporated into a Deep Belief Network (DBN) model to enhance the nonlinear mapping from spatial indicators to seismic resilience levels. A GIS-based evaluation system was developed to automate data management, spatial analysis, resilience computation, and the generation of visual outputs. Case studies on six representative Chinese cities demonstrated that the proposed framework effectively captured spatial heterogeneity in urban seismic resilience, with comprehensive scores ranging from 48.2 to 82.3. The improved entropy-DBN model outperformed conventional methods in both accuracy and convergence efficiency, reducing MAE, MSE, and RMSE to below 0.1 after 300 iterations. The system processed a single site in 12.28s and batch-processed ten sites in 42.75s, with core factor identification accuracy reaching 98.5%. Integrating GIS- and remote sensing-derived spatial information with advanced machine learning provides robust decision support for urban seismic risk mitigation, land-use planning, and disaster prevention at the regional scale. The proposed framework offers a transferable methodology that can be adapted to other regions.