AI-Based Image and Data Analysis for Automated Assessment of Residential Damage in Seismic Regions

Rapid post-earthquake damage assessment using patch-level CNNs and VLM

This study explores AI-driven image classification to expedite damage evaluation by identifying damaged buildings from post-disaster photos much faster than conventional methods, providing a more detailed understanding of structural integrity across affected areas.

Rapid Post-earthquake Damage Detection Using UAV Imagery and YOLO: A Case Study from the Kahramanmaras Earthquake

AI-Driven Visual Intelligence for Disaster Damage Assessment and Emergency Response

The increasing frequency and severity of natural disasters have created a growing demand for intelligent systems capable of rapidly assessing disaster impacts and supporting emergency response operations. This study presents a comprehensive analysis of recent artificial intelligence (AI) techniques applied to disaster damage assessment using satellite imagery, unmanned aerial vehicle (UAV) data, remote sensing products, and geospatial information. The reviewed studies encompass machine learning, deep learning, hybrid and transfer learning, and ensemble learning approaches across diverse disaster scenarios, including floods, landslides, earthquakes, wildfires, cyclones, and sinkholes. The analysis reveals that deep learning and ensemble learning techniques generally achieve superior predictive performance for disaster classification, detection, and damage mapping, while machine learning approaches offer advantages in interpretability and computational efficiency. The study further identifies key challenges, including dataset imbalance, limited geographical diversity, poor cross-regional generalization, high computational requirements, insufficient multimodal data integration, and difficulties in real-time deployment. Based on these findings, a conceptual framework for next-generation AI-driven disaster assessment is proposed, emphasizing multimodal data fusion, explainable AI, adaptive learning, and real-time decision support. The study provides comparative insights into existing methodologies, highlights critical research gaps, and outlines future directions for developing scalable, interpretable, and operationally effective disaster management systems.

An artificial intelligence model for rapid health assessment of earthquake-affected shear wall buildings: a disaster mitigation study

Rapid post-earthquake health assessment of shear wall buildings is critical for effective disaster mitigation and swift rehabilitation, particularly in earthquake‑affected urban areas. This study developed a novel artificial intelligence (AI)-based approach to predict key structural health parameters—maximum displacement, maximum inter‑storey drift (MISD), base shear, and maximum overturning moment—without requiring physical measurements of structural response. Instead of instrumented monitoring, the model relies solely on earthquake ground motion records that are readily available online, enabling fast and reliable predictions. A robust AI model was trained on a curated dataset generated from validated finite element simulations, attaining a high regression accuracy of 97.528%. The approach captures a broad spectrum of ground motion frequencies and structural response parameters, ensuring adaptability to diverse seismic conditions and urban settings. The results demonstrate that the AI model can efficiently assess structural integrity for disaster mitigation, providing a cost‑effective and non‑intrusive post‑earthquake assessment solution. Furthermore, the methodology can be extended to other building types, paving the way for a generalized AI‑driven structural health assessment system. This research highlights the transformative potential of AI to enhance the speed, accuracy, and practicality of seismic risk assessment for high‑rise shear wall buildings in disaster mitigation.

Artificial neural network-based approach for structural health monitoring of earthquake-affected industrial structures

The purpose of this study is to develop a reliable and rapid method for seismic damage detection in industrial structures using artificial neural networks (ANN). Industrial buildings often lack traditional seismic-resistant elements like shear walls, making them more vulnerable to earthquake-induced damage. Due to the difficulty of visually identifying internal structural damage, this research aims to offer a robust AI-based solution for detecting and evaluating such damage efficiently. The proposed approach facilitates post-earthquake safety assessment and structural integrity evaluation through rapid prediction of critical design parameters. This study employs a feed-forward multi-layer ANN model trained on earthquake ground motion data and structural response outputs obtained via finite element simulations of a reinforced concrete (RC) industrial building. The ANN maps real-time seismic input parameters to essential structural design variables, such as inter-storey drift and maximum displacement. The model is designed to provide quick and accurate estimations of structural damage, enabling post-earthquake evaluations. The approach integrates numerical modeling and AI to build a predictive tool for structural health monitoring in seismic environments. The ANN model demonstrated high accuracy and efficiency in predicting key structural responses under seismic loading. The trained network successfully identified critical post-earthquake parameters, such as maximum displacement and inter-storey drift, indicating its suitability for rapid seismic health assessments. Results confirmed the model's capability to localize and quantify potential damage in industrial structures that lack traditional seismic reinforcements. The approach is both time-efficient and cost-effective, showing potential for large-scale implementation in structural health monitoring systems for industrial infrastructure. This research introduces a novel ANN-based framework tailored specifically for seismic damage detection in industrial structures – a domain where SHM applications are limited. Unlike conventional methods, this approach enables real-time monitoring and rapid post-earthquake evaluation without reliance on labor-intensive inspections. By combining AI techniques with numerical simulations, the model provides a practical solution for enhancing seismic resilience and safety in industrial facilities. Its ability to predict critical structural parameters with minimal computation time offers significant value for emergency response, structural retrofitting decisions and disaster preparedness strategies.

A Deep Learning Framework for Rapid Building Damage Detection through Multimodal Data Fusion: Application to the 2025 Myanmar Earthquake

Abstract. Rapid and reliable assessment of building damage after major earthquakes is essential for effective emergency response and recovery planning. This study formulates post-disaster building damage detection (BDD) as a binary image classification task (damaged vs. undamaged buildings) using multimodal satellite data and a unified ResNet-18 backbone to enable a controlled comparison of fusion strategies. The analysis focuses on the Mw 7.7 Myanmar earthquake of 28 March 2025 and integrates post-event COSMO-SkyMed Second Generation (CSG) dual-polarization (HH, HV) SAR imagery, Maxar optical data, OpenStreetMap (OSM) building footprints, and UNOSAT damage annotations. Three fusion paradigms are evaluated: Early Fusion (EF), Late Fusion (LF), and a novel Middle Fusion (MF) approach. The proposed MF framework introduces a Footprint-Guided Cross-Attention (FGCA) mechanism that uses building geometry as a spatial prior to guide feature-level interaction between SAR and optical representations. Five-fold cross-validation results show that MF consistently outperforms EF and LF, achieving higher precision, F1-score, and robustness across modality configurations. By jointly exploiting SAR structural sensitivity, optical detail, and footprint-based spatial context, the proposed Footprint-Guided Middle Fusion (FGMF) framework enables accurate and scalable building damage mapping from heterogeneous Earth Observation (EO) data.