Wildfire Susceptibility Mapping in Baikal Natural Territory Using Random Forest

Forests, Journal Year: 2024, Volume and Issue: 15(1), P. 170 - 170

Published: Jan. 13, 2024

Wildfires are a significant problem in Irkutsk Oblast. They caused by climate change, thunderstorms, and human factors. In this study, we use the Random Forest machine learning method to map wildfire susceptibility of Oblast based on data from remote sensing, meteorology, government forestry authorities, emergency situations. The main contributions paper following: an improved domain model that describes information about weather conditions, vegetation type, infrastructure region context possible risk wildfires; database wildfires 2017 2020; results analysis factors cause assessment form fire hazard mapping. paper, collected visualized influencing their occurrence: meteorological, topographic, characteristics vegetation, activity (social factors). Data sets describing two classes, “fire” “no fire”, were generated. We introduced classification according which probability each specific cell territory can be determined built. allowed us achieve following accuracy indicators: accuracy—0.89, F1-score—0.88, AUC—0.96. comparison with earlier ones obtained using case-based reasoning revealed application approach considered initial stage for deeper investigations more accurate forecasting.

Language: Английский

---

Fire images classification based on a handcraft approach

Expert Systems with Applications, Journal Year: 2022, Volume and Issue: 212, P. 118594 - 118594

Published: Aug. 23, 2022

Language: Английский

---

Artificial neural networks for assessing forest fire susceptibility in Türkiye

Ecological Informatics, Journal Year: 2023, Volume and Issue: 75, P. 102034 - 102034

Published: Feb. 26, 2023

Language: Английский

---

Analysis of Spectral Separability for Detecting Burned Areas Using Landsat-8 OLI/TIRS Images under Different Biomes in Brazil and Portugal

Forests, Journal Year: 2023, Volume and Issue: 14(4), P. 663 - 663

Published: March 23, 2023

Fire is one of the natural agents with greatest impact on terrestrial ecosystem and plays an important ecological role in a large part surface. Remote sensing technique applied mapping monitoring changes forest landscapes affected by fires. This study presents spectral separability analysis for detection burned areas using Landsat-8 OLI/TIRS images context fires that occurred different biomes Brazil (dry ecosystem) Portugal (temperate forest). The research based fusion indices automatic classification algorithms scientifically proven to be effective as little human interaction possible. index (M) Reed–Xiaoli anomaly classifier (RXD) allowed evaluation thematic accuracy tested (Burn Area Index (BAI), Normalized Burn Ratio (NBR), Mid-Infrared (MIRBI), 2 (NBR2), Burned (NBI), Thermal (NBRT)). parameters were spatial dispersion validation data, commission error (CE), omission (OE), Sørensen–Dice coefficient (DC). results indicated exclusively SWIR1 SWIR2 bands showed high degree more suitable detecting areas, although it was observed characteristics soil performance indices. method bitemporal anomalous RXD proved increasing area terms temporal alteration performing unsupervised without relying ground truth. On other hand, main limitations non-abrupt changes, which very common low signal, especially 16-day revisit period. obtained this work able provide critical information fire accurate post-fire estimation dry ecosystems temperate forests. new comparative approach classify forests least possible interference, thus helping investigations when there available data addition favoring reduction fieldwork gross errors areas.

Language: Английский

---

Deep Learning-Based Wildfire Image Detection and Classification Systems for Controlling Biomass

International Journal of Intelligent Systems, Journal Year: 2023, Volume and Issue: 2023, P. 1 - 18

Published: Sept. 16, 2023

Forests are essential natural resources that directly impact the ecosystem. However, rising frequency of forest fires due to and artificial climate change has become a critical issue. A revolutionary municipal application proposes deploying an intelligence-based fire warning system prevent major disasters. This work aims present overview vision-based methods for detecting categorizing fires. The study employs detection dataset address classification difficulty discriminating between photos with without fire. method is based on convolutional neural network transfer learning Inception-v3. Thus, automatic identification current (including burning biomass) field research reducing negative repercussions. Early can also assist decision-makers in developing mitigation extinguishment strategies. Radial basis function Networks (RBFNs) rapid accurate image super resolution (RAISR) deep framework trained input detect active biomass. proposed RBFN-RAISR model’s performance recognizing nonfires was compared earlier CNN models using several criteria. water wave optimization technique used feature selection, noise blurring reduction, improvement restoration, enhancement restoration. When classifying no-fire photos, approach achieves 97.55% accuracy, 93.33% F-Score, 96.44% recall, 94.19% precision, error rate 24.89. Given one-of-a-kind dataset, suggested promising results categorization problem.

Language: Английский

---

Deep learning modeling of human activity affected wildfire risk by incorporating structural features: A case study in eastern China

Ecological Indicators, Journal Year: 2024, Volume and Issue: 160, P. 111946 - 111946

Published: March 1, 2024

Wildfire risk prediction is a critical component of disaster prevention and mitigation, often closely associated with local human activities in most regions. Recent studies demonstrate that employing joint modeling techniques using diverse datasets alongside Convolutional Neural Networks-Long Short-Term Memory Networks (ConvLSTM) produces favorable predictive results. However, previous research inadequately explored the different impact factors across categories spatial orientations, neglected fuels inside samples. This study focuses on six eastern provinces China, utilizing multi-source dataset comprising satellite-monitored wildfire products from 2012 to 2022, along various indicating terrestrial activities, simulated meteorological elements high-resolution vegetation imagery. By introducing channel attention mechanisms visual transformer mode, this optimizes ConvLSTM model. Results indicate noteworthy enhancement, elevating accuracy, Kappa coefficient, AUC ROC curves 91.15%, 80.87%, 97.01% 92.79%, 84.48%, 97.90%, respectively. Consequently, it reinforces accuracy by increase structural features within samples quantifying differences importance factors, which also validated application entire year 2023. Sensitivity analysis reveals current model still highly dependent factors. Notably, significantly surpasses influence terrain ecology elements, should be considered further models. Thus, has developed methodology integrating multiple sample features, could furnish high-precision daily kilometer-level products. method improve efficiency control improving narrowing high-risk areas.

Language: Английский

---

Physics Informed Machine Learning (PIML) for Design, Management and Resilience-Development of Urban Infrastructures: Concept, State-of-the-Art, Challenges and Opportunities

Archives of Computational Methods in Engineering, Journal Year: 2024, Volume and Issue: unknown

Published: June 12, 2024

Language: Английский

---

Forecasting of wind speed under wind-fire coupling scenarios by combining HS-VMD and AM-LSTM

Ecological Informatics, Journal Year: 2023, Volume and Issue: 77, P. 102270 - 102270

Published: Aug. 22, 2023

Language: Английский

---

Spain on fire: A novel wildfire risk assessment model based on image satellite processing and atmospheric information

Knowledge-Based Systems, Journal Year: 2023, Volume and Issue: 283, P. 111198 - 111198

Published: Nov. 22, 2023

Each year, wildfires destroy larger areas of Spain, threatening numerous ecosystems. Humans cause 90% them (negligence or provoked) and the behaviour individuals is unpredictable. However, atmospheric environmental variables affect spread wildfires, they can be analysed by using deep learning. In order to mitigate damage these events, we proposed novel Wildfire Assessment Model (WAM). Our aim anticipate economic ecological impact a wildfire, assisting managers in resource allocation decision-making for dangerous regions Castilla y León Andalucía. The WAM uses residual-style convolutional network architecture perform regression over greenness index, computing necessary resources, control extinction time, expected burnt surface area. It first pre-trained with self-supervision 100,000 examples unlabelled data masked patch prediction objective fine-tuned very small dataset, composed 445 samples. pretraining allows model understand situations, outclassing baselines 1,4%, 3,7% 9% improvement estimating human, heavy aerial resources; 21% 10,2% time; 18,8% Using provide an example assessment map León, visualizing resources entire region.

Language: Английский

---

Wildfire Scenarios for Assessing Risk of Cover Loss in a Megadiverse Zone within the Colombian Caribbean

Sustainability, Journal Year: 2024, Volume and Issue: 16(8), P. 3410 - 3410

Published: April 18, 2024

Rising wildfire incidents in South America, potentially exacerbated by climate change, require an exploration of sustainable approaches for fire risk reduction. This study investigates wildfire-prone meteorological conditions and assesses the susceptibility Colombia’s megadiverse northern region. Utilizing this knowledge, we apply a machine learning model Monte Carlo approach to evaluate sustainability strategies mitigating risk. The findings indicate that substantial number fires occur southern region, especially first two seasons year, northeast last seasons. Both are characterized high temperatures, minimal precipitation, strong winds, dry conditions. developed demonstrates significant predictive accuracy with HIT, FAR, POC 87.9%, 28.3%, 95.7%, respectively, providing insights into probabilistic aspects development. Various scenarios showed decrease soil temperature reduces mostly lower altitudes leaf skin reservoir content highest altitudes, as well north Sustainability strategies, such tree belts, agroforestry mosaics, forest corridors emerge crucial measures. results underscore importance proactive measures impact, offering actionable crafting effective amid escalating risks.

Language: Английский

---