An image processing-based hybrid MCDM framework for urban seismic vulnerability analysis and post-earthquake resource allocation: A case study of Dhaka

In resource-constrained and densely populated cities like Dhaka, emergency response capabilities following a major earthquake are likely to face significant limitations. Dense built environments, narrow roads, and highly congested neighborhoods can become inaccessible to emergency vehicles, further worsening the challenges and making prioritization in resource allocation crucial. To address the issue, this study identifies the structural, demographic, and systemic factors that are particularly relevant to the post-earthquake context of Dhaka and proposes a practical framework to rank residential areas by seismic vulnerability. The relative importance of the influencing factors was determined using the Fuzzy Analytic Hierarchy Process (Fuzzy AHP). This study employs CNN-based image processing to automatically generate criterion-specific structural scores from satellite imagery for evaluating the accessibility and feasibility of post-earthquake rescue operations. The predicted criterion-specific scores, together with scores of demographic and service-related indicators obtained from secondary databases, are then used as inputs in the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) algorithm to rank areas by their post-earthquake vulnerability. The proposed framework can act as a practical decision-support tool for policymakers and authorities responsible for emergency management to prioritize rescue operations and allocate limited resources more effectively after a major earthquake.