Search Results (637)

Building change detection and building damage assessment are two essential tasks in post-disaster analysis. Building change detection focuses on identifying changed building areas between bi-temporal images, while building damage assessment involves segmenting all buildings and classifying their damage severity. These tasks play a critical role in disaster response and urban development monitoring. Although supervised learning has significantly advanced building change detection and damage assessment, its reliance on large labeled datasets remains a major limitation. In contrast, self-supervised learning enables the extraction of meaningful data representations without explicit training labels. To address this challenge, we propose a self-supervised learning approach that unifies denoising autoencoders and contrastive learning, enabling effective data representation for building change detection and damage assessment. The proposed architecture integrates a dual denoising autoencoder with a Vision Transformer backbone and contrastive learning strategy, complemented by a Feature Pyramid Network-ResNet dual decoder and an Edge Guidance Module. This design enhances multi-scale feature extraction and enables edge-aware segmentation for accurate predictions. Extensive experiments were conducted on five public datasets, including xBD, LEVIR, LEVIR+, SYSU, and WHU, to evaluate the performance and generalization capabilities of the model. The results demonstrate that the proposed Denoising AutoEncoder-enhanced Dual-Fusion Network (DAEDFN) approach achieves competitive performance compared with fully supervised methods. On the xBD dataset, the largest dataset for building damage assessment, our proposed method achieves an F1 score of 0.892 for building segmentation, outperforming state-of-the-art methods. For building damage severity classification, the model achieves an F1 score of 0.632. On the building change detection datasets, the proposed method achieves F1 scores of 0.837 (LEVIR), 0.817 (LEVIR+), 0.768 (SYSU), and 0.876 (WHU), demonstrating model generalization across diverse scenarios. Despite these promising results, challenges remain in complex urban environments, small-scale changes, and fine-grained boundary detection. These findings highlight the potential of self-supervised learning in building change detection and damage assessment tasks. Full article

Floods are among the most devastating hydrometeorological natural disasters worldwide, causing massive infrastructure and economic loss in low-lying, flood-prone developing countries like Bangladesh. Effective disaster mitigation relies on organized and detailed flood damage information to facilitate emergency evacuation, coordinate relief distribution, and formulate an effective disaster management policy. Nevertheless, the nation confronts considerable obstacles due to insufficient historical flood damage data and the underdevelopment of near-real-time (NRT) flood monitoring systems. This study addresses this issue by developing a replicable methodology for flood damage assessment and NRT monitoring systems. Using the Google Earth Engine (GEE) platform, we analyzed flood events from 2019 to 2023, integrating geospatial layers such as roads, cropland, etc. Analysis of flood events over the five-year period revealed substantial impacts, with 21.60% of the total area experiencing inundation. This flooding affected 6.92% of cropland and 4.16% of the population. Furthermore, 18.10% of the road network, spanning over 21,000 km within the study area, was also affected. This system has the potential to enhance emergency response capabilities during flood events and inform more effective disaster mitigation policies. Full article

Landslides are among the most hazardous natural phenomena affecting Greater Abidjan, causing significant economic and social damage. Strategic planning supported by geographic information systems (GIS) can help mitigate potential losses and enhance disaster resilience. This study evaluates landslide susceptibility using logistic regression and frequency ratio models. The analysis is based on a dataset comprising 54 mapped landslide scarps collected from June 2015 to July 2023, along with 16 thematic predictor variables, including altitude, slope, aspect, profile curvature, plan curvature, drainage area, distance to the drainage network, normalized difference vegetation index (NDVI), and an urban-related layer. A high-resolution (5-m) digital elevation model (DEM), derived from multiple data sources, supports the spatial analysis. The landslide inventory was randomly divided into two subsets: 80% for model calibration and 20% for validation. After optimization and statistical testing, the selected thematic layers were integrated to produce a susceptibility map. The results indicate that 6.3% (0.7 km²) of the study area is classified as very highly susceptible. The proportion of the sample (61.2%) in this class had a frequency ratio estimated to be 20.2. Among the predictive indicators, altitude, slope, SE, S, NW, and NDVI were found to have a positive impact on landslide occurrence. Model performance was assessed using the area under the receiver operating characteristic curve (AUC), demonstrating strong predictive capability. These findings can support informed land-use planning and risk reduction strategies in urban areas. Furthermore, the prediction model should be communicated to and understood by local authorities to facilitate disaster management. The cost function was adopted as a novel approach to delineate hazardous zones. Considering the landslide inventory period, the increasing hazard due to climate change, and the intensification of human activities, a reasoned choice of sample size was made. This informed decision enabled the production of an updated prediction map. Optimal thresholds were then derived to classify areas into high- and low-susceptibility categories. The prediction map will be useful to planners in helping them make decisions and implement protective measures. Full article

Rapid and comprehensive assessment of building damage caused by earthquakes is essential for effective emergency response and rescue efforts in the immediate aftermath. Advanced technologies, including real-time simulations, remote sensing, and multi-sensor systems, can effectively enhance situational awareness and structural damage evaluations. However, most existing methods rely on isolated time snapshots, and few studies have systematically explored the continuous, time-scaled integration and update of building damage estimates from multiple data sources. This study proposes a stepwise framework that continuously updates time-scaled, single-damage estimation outputs using the best available multi-sensor data for estimating earthquake-induced building damage. We demonstrated the framework using the 2024 Noto Peninsula Earthquake as a case study and incorporated official damage reports from the Ishikawa Prefectural Government, real-time earthquake building damage estimation (REBDE) data, and satellite-based damage estimation data (ALOS-2-building damage estimation (BDE)). By integrating the REBDE and ALOS-2-BDE datasets, we created a composite damage estimation product (integrated-BDE). These datasets were statistically validated against official damage records. Our framework showed significant improvements in accuracy, as demonstrated by the mean absolute percentage error, when the datasets were integrated and updated over time: 177.2% for REBDE, 58.1% for ALOS-2-BDE, and 25.0% for integrated-BDE. Finally, for stepwise damage estimation, we proposed a methodological framework that incorporates social media content to further confirm the accuracy of damage assessments. Potential supplementary datasets, including data from Internet of Things-enabled home appliances, real-time traffic data, very-high-resolution optical imagery, and structural health monitoring systems, can also be integrated to improve accuracy. The proposed framework is expected to improve the timeliness and accuracy of building damage assessments, foster shared understanding of disaster impacts across stakeholders, and support more effective emergency response planning, resource allocation, and decision-making in the early stages of disaster management in the future, particularly when comprehensive official damage reports are unavailable. Full article

In October 2024, a DANA (Isolated Depression at High Levels) triggered torrential rains across the Valencian Community, causing 227 deaths, severe infrastructure damage, and economic losses estimated at €17.8 billion. In this context of crisis, startups, despite having fewer resources and less experience than large corporations, played a significant role in crisis communication, shaping public perception and operational continuity. This study explores the communication strategies adopted by startups during and after the disaster, focusing on their activity on Instagram, TikTok, and Facebook between October 2024 and January 2025. Using a mixed-methods approach, we conducted a quantitative analysis of digital discourse through the Fanpage Karma tool, assessing metrics such as engagement, reach, and posting frequency. Sentiment analysis was performed using GPT-4, an advanced natural language processing model, and in-depth interviews with startup representatives provided qualitative insights into reputational impacts. The findings reveal that startups which aligned their discourse with the social context, prioritizing transparency and emotional proximity, enhanced their visibility and credibility. These results underscore how effective crisis communication not only mitigates reputational risk but also strengthens the local entrepreneurial ecosystem through trust-building and social responsibility. Full article

Rockfalls, among the most common natural disasters, pose risks such as traffic congestion, casualties, and substantial property damage. Guizhou Province, with China’s fourth-longest highway network, features mountainous terrain prone to frequent rockfall incidents annually. Consequently, assessing highway rockfall risks in Guizhou Province is crucial for safeguarding the lives and travel of residents. This study evaluates highway rockfall risk through three key components: susceptibility, hazard, and vulnerability. Susceptibility was assessed using information content and logistic regression methods, considering factors such as elevation, slope, normalized difference vegetation index (NDVI), aspect, distance from fault, relief amplitude, lithology, and rock weathering index (RWI). Hazard assessment utilized a fuzzy analytic hierarchy process (AHP), focusing on average annual rainfall and daily maximum rainfall. Socioeconomic factors, including GDP, population density, and land use type, were incorporated to gauge vulnerability. Integration of these assessments via a risk matrix yielded comprehensive highway rockfall risk profiles. Results indicate a predominantly high risk across Guizhou Province, with high-risk zones covering 41.19% of the area. Spatially, the western regions exhibit higher risk levels compared to eastern areas. Notably, the Bijie region features over 70% of its highway mileage categorized as high risk or above. Logistic regression identified distance from fault lines as the most negatively correlated factor affecting highway rockfall susceptibility, whereas elevation gradient demonstrated a minimal influence. This research provides valuable insights for decision-makers in formulating highway rockfall prevention and control strategies. Full article

To explore the influence of mine roof on the damage and failure of sandstone surrounding rock under different pressure rates, mechanical experiments with different strain rates were carried out on sandstone rock samples. The strength, deformation, failure, energy and damage characteristics of rock samples with different strain rates were also discussed. The research results show that with the increases in the strain rate, peak stress, and elastic modulus show a monotonically increasing trend, while the peak strain decreases in the reverse direction. At a low strain rate, the proportion of the mass fraction of complete rock blocks in the rock sample is relatively high, and the shape integrity is good, while rock samples with a high strain rate retain more small-sized fragmented rock blocks. This indicates that under high-rate loading, the bifurcation phenomenon of secondary cracks is obvious. The rock samples undergo a failure form dominated by small-sized fragments, with severe damage to the rock samples and significant fractal characteristics of the fragments. At the initial stage of loading, the primary fractures close, and the rock samples mainly dissipate energy in the forms of frictional slip and mineral fragmentation. In the middle stage of loading, the residual fractures are compacted, and the dissipative strain energy keeps increasing continuously. In the later stage of loading, secondary cracks accelerate their expansion, and elastic strain energy is released sharply, eventually leading to brittle failure of the rock sample. Under a low strain rate, secondary cracks slowly expand along the clay–quartz interface and cause intergranular failure of the rock sample. However, a high strain rate inhibits the stress relaxation of the clay, forces the energy to transfer to the quartz crystal, promotes the penetration of secondary cracks through the quartz crystal, and triggers transgranular failure. A constitutive model based on energy damage was further constructed, which can accurately characterize the nonlinear hardening characteristics and strength-deformation laws of rock samples with different strain rates. The evolution process of its energy damage can be divided into the unchanged stage, the slow growth stage, and the accelerated growth stage. The characteristics of this stage reveal the sudden change mechanism from the dissipation of elastic strain energy of rock samples to the unstable propagation of secondary cracks, clarify the cumulative influence of strain rate on damage, and provide a theoretical basis for the dynamic assessment of surrounding rock damage and disaster early warning when the mine roof comes under pressure. Full article

Concrete is a fundamental material in structural engineering, widely used in critical infrastructure such as bridges, nuclear power plants, and dams. These structures may be subjected to extreme dynamic loads resulting from natural disasters, industrial accidents, or missile impacts. Therefore, a comprehensive understanding of concrete behavior under high strain rates is essential for safe and resilient design. Experimental investigations, particularly spalling tests, have highlighted the strain-rate sensitivity of concrete in dynamic tensile loading conditions. This study presents a macroscopic 3D discrete element model specifically developed to simulate the dynamic response of concrete subjected to extreme loading. Unlike conventional continuum-based models, the proposed discrete element framework is particularly suited to capturing damage and fracture mechanisms in cohesive materials. A key innovation lies in incorporating a physically grounded strain-rate dependency directly into the local cohesive laws that govern inter-element interactions. The originality of this work is further underlined by the validation of the discrete element model under dynamic tensile loading through the simulation of spalling tests on normalstrength concrete at strain rates representative of severe impact scenarios (30–115 s⁻¹). After calibrating the model under quasi-static loading, the simulations accurately reproduce key experimental outcomes, including rear-face velocity profiles and failure characteristics. Combined with prior validations under high confining pressure, this study reinforces the capability of the discrete element method for modeling concrete subjected to extreme dynamic loading, offering a robust tool for predictive structural assessment and design. Full article

Increasing frequency and severity of urban flooding, driven by climate change and urban population growth, present major challenges. Traditional flood control infrastructure alone cannot fully prevent flood damage, highlighting the need for a comprehensive and multi-dimensional disaster management approach. This study proposes the Flood Risk Index for Building (FRIB)—a building-level assessment framework that integrates vulnerability, hazard, and exposure. FRIB assigns customized risk levels to individual buildings and evaluates the effectiveness of a multi-dimensional method. Compared to traditional indicators like flood depth, FRIB more accurately identifies high-risk areas by incorporating diverse risk factors. It also enables efficient resource allocation by excluding low-risk buildings, focusing efforts on high-risk zones. For example, in a case where 5124 buildings were targeted based on 1 m flood depth, applying FRIB excluded 24 buildings with “low” risk and up to 530 with “high” risk, reducing unnecessary interventions. Moreover, quantitative metrics like entropy and variance showed that as FRIB levels rise, flood depth distributions become more balanced—demonstrating that depth alone does not determine risk. In conclusion, while qualitative labels such as “very low” to “very high” aid intuitive understanding, FRIB’s quantitative, multi-dimensional approach enhances precision in urban flood management. Future research may expand FRIB’s application to varied regions, supporting tailored flood response strategies. Full article

Flash floods are natural disasters that have significant impacts on human life and economic damage. The detection of flash floods using remote-sensing techniques provides essential data for subsequent flood-risk assessment through the preparation of flood inventory samples. In this research, a new deep-learning approach for bi-temporal flash-flood detection in Synthetic Aperture Radar (SAR) is proposed. It combines a U-Net convolutional network with a Transformer model using a compact Convolutional Tokenizer (CCT) to improve the efficiency of long-range dependency learning. The hybrid model, namely CCT-U-ViT, naturally combines the spatial feature extraction of U-Net and the global context capability of Transformer. The model significantly reduces the number of basic blocks as it uses the CCT tokenizer instead of conventional Vision Transformer tokenization, which makes it the right fit for small flood detection datasets. This model improves flood boundary delineation by involving local spatial patterns and global contextual relations. However, the method is based on Sentinel-1 SAR images and focuses on Erbil, Iraq, which experienced an extreme flash flood in December 2021. The experimental comparison results show that the proposed CCT-U-ViT outperforms multiple baseline models, such as conventional CNNs, U-Net, and Vision Transformer, obtaining an impressive overall accuracy of 91.24%. Furthermore, the model obtains better precision and recall with an F1-score of 91.21% and mIoU of 83.83%. Qualitative results demonstrate that CCT-U-ViT can effectively preserve the flood boundaries with higher precision and less salt-and-pepper noise compared with the state-of-the-art approaches. This study underscores the significance of hybrid deep-learning models in enhancing the precision of flood detection with SAR data, providing valuable insights for the advancement of real-time flood monitoring and risk management systems. Full article

Natural disasters can cause severe damage to critical infrastructure such as road networks, significantly delaying rescue and recovery efforts. Conventional road damage assessments rely heavily on manual inspection, which is labor-intensive, time-consuming, and infeasible in large-scale disaster-affected areas. This study aims to propose a deep learning-based framework to automatically detect and quantitatively assess road damage using high-resolution pre- and post-disaster satellite imagery. To achieve this, the study systematically compares three distinct change detection approaches: single-timeframe overlay, difference-based segmentation, and Siamese feature fusion. Experimental results, validated over multiple runs, show the difference-based model achieved the highest overall F1-score (0.594 ± 0.025), surpassing the overlay and Siamese models by approximately 127.6% and 27.5%, respectively. However, a key finding of this study is that even this best-performing model is constrained by a low detection recall (0.445 ± 0.051) for the ‘damaged road’ class. This reveals that severe class imbalance is a fundamental hurdle in this domain for which standard training strategies are insufficient. This study establishes a crucial benchmark for the field, highlighting that future research must focus on methods that directly address class imbalance to improve detection recall. Despite its quantified limitations, the proposed framework enables the visualization of damage density maps, supporting emergency response strategies such as prioritizing road restoration and accessibility planning in disaster-stricken areas. Full article

Effective post-disaster damage assessment is crucial for guiding emergency response and resource allocation. This study introduces DamageScope, an integrated deep learning framework designed to detect and classify building damage levels from post-disaster satellite imagery. The proposed system leverages a convolutional neural network trained exclusively on post-event data to segment building footprints and assign them to one of four standardized damage categories: no damage, minor damage, major damage, and destroyed. The model achieves an average F1 score of 0.598 across all damage classes on the test dataset. To support geospatial analysis, the framework extracts the coordinates of damaged structures using embedded metadata, enabling rapid and precise mapping. These results are subsequently visualized through an interactive, web-based platform that facilitates spatial exploration of damage severity. By integrating classification, geolocation, and visualization, DamageScope provides a scalable and operationally relevant tool for disaster management agencies seeking to enhance situational awareness and expedite post-disaster decision making. Full article

The rapid development of underground engineering contributes significantly to achieving China’s “dual carbon” strategic goals. However, during the construction and operation phases, this engineering project faces diverse risks and challenges related to disasters. Consequently, enhancing the evaluation capability for underground engineering resilience is imperative. Based on the characteristics of resilience evaluation and enhancement in underground engineering, this study defines the concept and objectives of resilience evaluation for underground space engineering and analyzes corresponding enhancement methods. By considering aspects such as the magnitude of collapse disaster risk in underground engineering, its vulnerability, resistance capacity, adaptability to disasters, recovery ability, and economic feasibility, a comprehensive index system for evaluating the resilience of collapse disaster risks in underground engineering has been established. This research suggests that disaster risk management should shift from passive to active prevention. Through resilience evaluation case applications, it is possible to improve the design objectives of underground engineering towards “structural recoverability”, “ease of damage repair”, and “controllable consequences after a disaster”. The integration of intelligent static assessment models based on artificial intelligence algorithms can effectively enhance the accuracy of resilience evaluations. Furthermore, dynamic assessments using multiple data fusion techniques combined with numerical simulations represent promising directions for improving the overall resilience of underground engineering. Full article

Urban water supply pipelines play a critical role in ensuring the continuous delivery of water, and their failure during earthquakes can result in significant societal disruptions. This study proposes a seismic damage prediction method for urban buried water supply pipelines affected by seismic wave propagation, grounded in empirical data from past earthquake events. The method integrates key influencing factors, including pipeline material, diameter, joint type, age, and soil corrosivity. To enhance its practical applicability and address the challenge of quantifying soil corrosivity, a simplified classification approach is introduced. The proposed model is validated using observed pipeline damage data from the 2008 Wenchuan earthquake, with predicted results showing relatively good agreement with actual failure patterns, thereby demonstrating the model’s reliability for seismic risk assessment. Furthermore, the model is applied to assess potential earthquake-induced damage to buried pipelines in the city center of Ganzhou, and the corresponding results are presented. The findings support earthquake risk mitigation and the protection of urban infrastructure, while also providing valuable guidance for the replacement of aging pipelines and the enhancement of urban disaster resilience. Full article

As climate change intensifies, assessing vulnerability at territorial levels such as cities, countries, and regions is essential for effective adaptation planning. This study evaluates the applicability of the United Nations Environment Programme and South Pacific Applied Geoscience Commission’s Environmental Vulnerability Index (EVI) for coastal regions in South Korea. By adapting and localizing 50 international indicators and a Geographic Information System framework, this research developed a Korean Coastal Vulnerability Index and used spatial regression analysis to compare results to historical water-related disaster data from 2010 to 2019. The findings reveal that contrary to South Korea’s global classification of “extremely vulnerable”, most coastal counties appear relatively resilient when viewed through the localized model. Sub-index analyses indicate that ecological and anthropogenic damage factors show the strongest correlation with past disasters among the hazard, resistance, and damage categories. While the model’s explanatory power was modest (R² = 0.017), the regression nonetheless provides meaningful insight into how global indices can reflect local vulnerability patterns. The regression results confirm that based on historical hazard records, the international model effectively predicts Korean coastal vulnerability. It demonstrates the potential of scaling down global models to fit national contexts, offering a replicable approach for countries lacking localized vulnerability frameworks. It advances climate adaptation research through methodological innovation, policy-relevant spatial analysis, and theoretical insights into the multidimensional nature of vulnerability. The results support more precise, data-driven resilience planning and promote international collaboration in climate risk management. Full article