Ecological Indicators, Journal Year: 2024, Volume and Issue: 163, P. 112067 - 112067
Published: May 6, 2024
Deep learning techniques through semantic segmentation networks have been widely used for natural disaster analysis and response. The underlying base of these implementations relies on convolutional neural (CNNs) that can accurately precisely identify locate the respective areas interest within satellite imagery or other forms remote sensing data, thereby assisting in evaluation, rescue planning, restoration endeavours. Most CNN-based deep-learning models encounter challenges related to loss spatial information insufficient feature representation. This issue be attributed their suboptimal design layers capture multiscale-context failure include optimal during pooling procedures. In early CNNs, network encodes elementary representations, such as edges corners, whereas, progresses toward later layers, it more intricate characteristics, complicated geometric shapes. theory, is advantageous a extract features from several levels because generally yield improved results when both simple maps are employed together. study comprehensively reviews current developments deep methodologies segment images associated with disasters. Several popular models, SegNet U-Net, FCNs, FCDenseNet, PSPNet, HRNet, DeepLab, exhibited notable achievements various applications, including forest fire delineation, flood mapping, earthquake damage assessment. These demonstrate high level efficacy distinguishing between different land cover types, detecting infrastructure has compromised damaged, identifying regions fire-susceptible further dangers.
Language: Английский
Citations
19Post Disaster Damage Assessment Using Ultra-High-Resolution Aerial Imagery with Semi-Supervised Transformers
Sensors, Journal Year: 2023, Volume and Issue: 23(19), P. 8235 - 8235
Published: Oct. 3, 2023
Preliminary damage assessments (PDA) conducted in the aftermath of a disaster are key first step ensuring resilient recovery. Conventional door-to-door inspection practices time-consuming and may delay governmental resource allocation. A number research efforts have proposed frameworks to automate PDA, typically relying on data sources from satellites, unmanned aerial vehicles, or ground together with processing using deep convolutional neural networks. However, before such can be adopted practice, accuracy fidelity predictions level at scale an entire building must comparable human assessments. Towards this goal, we propose PDA framework leveraging novel ultra-high-resolution (UHRA) images combined state-of-the-art transformer models make multi-class buildings. We demonstrate that semi-supervised trained vast amounts unlabeled able surpass generalization capabilities frameworks. In our series experiments, aim assess impact incorporating data, as well use different model architectures. By integrating UHRA models, results suggest overcome significant limitations satellite imagery traditional CNN leading more accurate efficient
Language: Английский
Citations
15Improving Landslide Susceptibility Prediction in Uttarakhand through Hyper-Tuned Artificial Intelligence and Global Sensitivity Analysis
Earth Systems and Environment, Journal Year: 2024, Volume and Issue: unknown
Published: Sept. 23, 2024
Language: Английский
Citations
5RESNET-34 DERİN ÖĞRENME ALGORİTMASI İLE DEPREM SONRASI YIKILAN YAPILARIN TESPİTİ: 6 ŞUBAT 2023 DEPREMİ, ANTAKYA ÖRNEĞİ
Mühendislik Bilimleri ve Tasarım Dergisi, Journal Year: 2025, Volume and Issue: 13(1), P. 49 - 63
Published: March 19, 2025
Bu çalışma, 6 Şubat 2023 Kahramanmaraş depreminin etkilediği Hatay ilinin Antakya ve Defne ilçelerinde ResNET-34 derin öğrenme algoritmasını kullanarak deprem sırasında yıkılan yıkılmayan yapıların tespit edilmesini incelemiştir. Çalışmada, Pleiades-1B uydu görüntüleri OpenStreetMap verileri kullanılarak hasar durumları analiz edilmiştir. Derin algoritması olarak mimarisi, bu verilerle eğitilmiş %85 doğruluk %91 F1 skoru elde yüksek oranı, yöntemlerinin afet sonrası yapı tespitinde ne denli etkili bir araç olduğunu ortaya koymuştur. Çalışmanın bulguları, tekniklerinin analizi güvenliği değerlendirmelerinde etkin kullanılabileceğini göstermiştir. Ayrıca yönetimi şehir planlaması süreçlerinde rol oynayabileceğine dair değerli bilgiler sunmuştur.
Citations
0Leveraging convlstm and satellite imagery for predictive modeling of floods, landslides, and earthquakes
i-manager’s Journal on Future Engineering and Technology, Journal Year: 2025, Volume and Issue: 20(2), P. 60 - 60
Published: Jan. 1, 2025
This study combines the spatial data from satellite imagery with temporal learning capabilities of convolutional long short-term memory (ConvLSTM) networks to improve both prediction accuracy and processing efficiency. By utilizing diverse spectral bands resolutions, model captures a wide range environmental features. Preprocessing steps, such as normalization noise reduction, are applied refine input enhance performance ConvLSTM network. The architecture is carefully structured balance dependencies, ensuring effective integration satellite-derived data. framework optimized identify complex relationships within dataset, enabling precise forecasts upcoming disasters. It has been tested on various natural events, including hurricanes, floods, wildfires, achieving higher shorter lead times compared traditional techniques. aims strengthen early warning systems, disaster preparedness, reduce economic social damage affected regions.
Language: Английский
Citations
0Innovative Deep Learning Image Technologies
Advances in computational intelligence and robotics book series, Journal Year: 2025, Volume and Issue: unknown, P. 145 - 180
Published: March 7, 2025
The chapter gives an overview of the applications deep learning and image processing in different industries medicine, automobiles, entertainment, security. Multiple advanced techniques such as CNN, GAN, ViT that have become handy analysis processing. From medical diagnostics to autonomous vehicles, environmental monitoring, surveillance, its show impact on accuracy efficiency. It also discusses critical ethical issues, data privacy, model biases, energy consumption, points out some possible solutions reduce those effects. In general, this contribution provides a advances related by potential for further innovative developments wide range applications.
Language: Английский
Citations
0ChatGPT in transforming communication in seismic engineering: Case studies, implications, key challenges and future directions
Earthquake Science, Journal Year: 2024, Volume and Issue: 37(4), P. 352 - 367
Published: July 13, 2024
Language: Английский
Citations
3Rapid seismic damage assessment of building portfolio based on fusion of surrogate model and monitoring data
International Journal of Disaster Risk Reduction, Journal Year: 2025, Volume and Issue: unknown, P. 105293 - 105293
Published: Feb. 1, 2025
Language: Английский
Citations
0Autonomous Decision-Making Enhancing Natural Disaster Management through Open World Machine Learning: A Systematic Review
Human-Centric Intelligent Systems, Journal Year: 2025, Volume and Issue: unknown
Published: April 25, 2025
Abstract Natural disasters can be fatal and necessitate critical thinking in decision-making processes; thus, simulations of multiple scenarios must evaluated to prepare for such imminent events. An Open World Machine Learning (OWML) framework has been proposed as a focal tool improve natural disaster management. To better understand formulate comprehensive foundation, systematic review different cases was conducted. In this study, four types disasters-earthquakes, floods, wildfires, hurricanes-were analyzed using machine learning techniques within the OWML framework. The essence approach lies its ability accumulate knowledge from various sources adapt new, unknown event types. Moreover, by explicitly incorporating both qualitative quantitative characteristics, enhances predictive accuracy adaptability. results demonstrate that models effectively process large geospatial datasets respond looming threats with greater precision. Albeit some limitations, data quality model complexity, findings suggest serve foundational governments reduce expenditure evacuation strategies. evolutionary nature allows continuous adaptation, which is crucial escalate frequency intensity due climate change. essence, study provides significant contribution management demonstrating potential frameworks enhancing processes.
Language: Английский
Citations
0Post-disaster damage and loss assessment in the Iranian healthcare sector: a qualitative interview study
BMC Public Health, Journal Year: 2024, Volume and Issue: 24(1)
Published: Sept. 4, 2024
Language: Английский
Citations
1