Deep Learning Approaches for Wildland Fires Using Satellite Remote Sensing Data: Detection, Mapping, and Prediction

Fire, Journal Year: 2023, Volume and Issue: 6(5), P. 192 - 192

Published: May 7, 2023

Wildland fires are one of the most dangerous natural risks, causing significant economic damage and loss lives worldwide. Every year, millions hectares lost, experts warn that frequency severity wildfires will increase in coming years due to climate change. To mitigate these hazards, numerous deep learning models were developed detect map wildland fires, estimate their severity, predict spread. In this paper, we provide a comprehensive review recent techniques for detecting, mapping, predicting using satellite remote sensing data. We begin by introducing systems use wildfire monitoring. Next, methods employed tasks, including fire detection estimation, spread prediction. further present popular datasets used studies. Finally, address challenges faced accurately behaviors, suggest future directions developing reliable robust models.

Language: Английский

---

SWVR: A Lightweight Deep Learning Algorithm for Forest Fire Detection and Recognition

Forests, Journal Year: 2024, Volume and Issue: 15(1), P. 204 - 204

Published: Jan. 19, 2024

The timely and effective detection of forest fires is crucial for environmental socio-economic protection. Existing deep learning models struggle to balance accuracy a lightweight design. We introduce SWVR, new algorithm. Utilizing the Reparameterization Vision Transformer (RepViT) Simple Parameter-Free Attention Module (SimAM), SWVR efficiently extracts fire-related features with reduced computational complexity. It bi-directional fusion network combining top-down bottom-up approaches, incorporates Ghost Shuffle Convolution (GSConv), uses Wise Intersection over Union (WIoU) loss function. achieves 79.6% in detecting fires, which 5.9% improvement baseline, operates at 42.7 frames per second. also reduces model parameters by 11.8% cost 36.5%. Our results demonstrate SWVR’s effectiveness achieving high fewer resources, offering practical value fire detection.

Language: Английский

---

Deep artificial intelligence applications for natural disaster management systems: A methodological review

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

---

Automatic coastline extraction through enhanced sea-land segmentation by modifying Standard U-Net

International Journal of Applied Earth Observation and Geoinformation, Journal Year: 2022, Volume and Issue: 109, P. 102785 - 102785

Published: April 27, 2022

Sea-land segmentation (SLS) is an essential remote sensing task for various coastal and environmental studies such as coastline extraction, erosion, area monitoring, ship or iceberg detection. This study aims at improving the SLS performance by modifying Standard U-Net (SUN) model developing automatic extraction framework. SUN generally has acceptable in many applications. However, better outputs are needed reliable extraction. In our proposed framework, we firstly analyzed three different input images, including Red-Green-Blue (RGB), Normalized Difference Water Index (NDWI), Near-Infrared (NIR) images. Secondly, modified architecture to improve results. The main modifications using loss functions two fusion methods RGB NIR results were then passed into subsequent pipeline based on morphological operations pixel connectivity analysis. training testing steps accomplished utilizing a benchmark dataset of China's areas. Moreover, another consisting time series Landsat-8 imagery from southern Caspian Sea coastlines was collected evaluate efficiency. indicate that could effectively enhance SUN, with most significant improvement Intersection over Union (IoU) score being high 1.68% 8.95% China datasets, respectively, while outperforming other state-of-the-art models FC-DenseNet DeepLabV3+.

Language: Английский

---

Early Wildfire Detection Technologies in Practice—A Review

Sustainability, Journal Year: 2022, Volume and Issue: 14(19), P. 12270 - 12270

Published: Sept. 27, 2022

As fires grow in intensity and frequency each year, so has the resistance from their anthropic victims form of firefighting technology research. Although it is impossible to completely prevent wildfires, potential devastation can be minimized if are detected precisely geolocated while still nascent phases. Furthermore, automated approaches without human involvement comparatively more efficient, accurate capable monitoring extremely remote vast areas. With this specific intention, many research groups have proposed numerous last several years, which grouped broadly into these four distinct categories: sensor nodes, unmanned aerial vehicles, camera networks satellite surveillance. This review paper discusses notable advancements trends categories, with subsequent shortcomings challenges. We also describe a technical overview common prototypes analysis models used diagnose fire raw input data. By writing paper, we hoped create synopsis current state emergent area provide reference for further developments other interested researchers.

Language: Английский

---

A deep learning model using geostationary satellite data for forest fire detection with reduced detection latency

GIScience & Remote Sensing, Journal Year: 2022, Volume and Issue: 59(1), P. 2019 - 2035

Published: Nov. 16, 2022

Although remote sensing of active fires is well-researched, their early detection has received less attention. Additionally, simple threshold approaches based on contextual statistical analysis suffer from generalization problems. Therefore, this study proposes a deep learning-based forest fire algorithm, with focus reducing latency, utilizing 10-min interval high temporal resolution Himawari-8 Advanced Himawari Imager. Random (RF) and convolutional neural network (CNN) were utilized for model development. The CNN accurately reflected the approach adopted in previous studies by learning information between adjacent matrices an image. This also investigates contribution spatial to two machine techniques combining input features. Temporal factors contributed reduction latency false alarms, respectively, could be most effectively detected using both types information. overall accuracy, precision, recall, F1-score 0.97, 0.89, 0.41, 0.54, best scheme among RF-based schemes 0.98, 0.91, 0.63, 0.74, that CNN-based schemes. indicated better performance attributed its pattern training data augmentation. all test within average 12 min, one case was 9 min earlier than recording time. Moreover, proposed outperformed recent operational satellite-based algorithms. Further generality results showed had reliable robust under varied environmental conditions. Overall, our demonstrated benefits geostationary monitoring.

Language: Английский

---

Oil spills detection from SAR Earth observations based on a hybrid CNN transformer networks

Marine Pollution Bulletin, Journal Year: 2023, Volume and Issue: 190, P. 114834 - 114834

Published: March 17, 2023

Language: Английский

---

Advancements in remote sensing for active fire detection: A review of datasets and methods

The Science of The Total Environment, Journal Year: 2024, Volume and Issue: 943, P. 173273 - 173273

Published: May 31, 2024

Language: Английский

---

A Review of Technologies for the Early Detection of Wildfires

ASME Open Journal of Engineering, Journal Year: 2025, Volume and Issue: 4

Published: Jan. 1, 2025

Abstract Wildfires have become a persistent and growing global risk, causing increasing financial, human, environmental damage. By all accounts predictions, they will continue to rise in frequency intensity throughout the 21st century. This paper begins by analyzing physics of fire outlines why detecting wildfires their incipient stages is most effective way manage them. We review various architectures approaches adopted for wildfire detection, including spaceborne, airborne, fixed cameras, sensor networks. The further analyzes pros cons each approach reviews recent deployments published research. In particular, it focuses on significant role that Artificial Intelligence (AI) Deep Learning (DL) play improving effectiveness aforementioned architectures. It examines algorithms models detection platforms compares effectiveness. study suggests solutions combine elements mentioned architectures, integrating different sensors look signatures, coupling them with sophisticated DL maximize sensitivity while minimizing false alarms. An important trend advancement low-power high-performance hardware enabling real-time operation an edge device limited memory processing resources. As seconds minutes can significantly impact our ability effectively suppress wildfire, process data, at network edge, even remote, unpredictable, fragile environment crucial.

Language: Английский

---

A self-adaptive wildfire detection algorithm by fusing physical and deep learning schemes

International Journal of Applied Earth Observation and Geoinformation, Journal Year: 2024, Volume and Issue: 127, P. 103671 - 103671

Published: Jan. 27, 2024

Currently, the spectra-based physical models and deep learning methods are frequently used to detect wildfires from remote sensing data. However, algorithms mainly rely on radiative transfer processes, which limit their effectiveness in detecting small weak fires. On other hand, usually lack mechanism constraints, thus generally resulting false alarms of bright surfaces. It is promising combine advantages them correspondingly reduce inherent error a single algorithm. To this end, paper, both local contextual global index method based mechanisms optimized, simultaneously, new U-Net model also establish accurately Moreover, YOLO v5 incorporated for first time extract remove objects with high exposure. Based above series novel works, self-adaptive fusing algorithm finally proposed. Our results reveal that: (1) Short-wave infrared band about 2.15 μm crucial fire detection data moderate-to-high resolutions. Taking Landsat 8 as an example, combinations 7, 6, 2(SWIR + VI), 5(SWIR NIR), 5, 3(SWIR VI NIR) show reasonable accuracy, recall rate greater than 81 %. The thermal can be assist general location serve alternative choice extreme cases. (2) optimized predict more accurate positions. (3) very effective introduce framework exposure urban suburban regions. (4) proposed fusion integrates various schemes, proving its better performance terms robustness, stability generality compared any method. Even situations such Gobi Desert, thin cloud edges, mountain shadow areas, still works well. tests Sentinel-2A, WorldView-3, SPOT-4 potential applicability newly algorithm, especially fine spatial spectral

Language: Английский

---