Fire Classes Research Articles

Convolutional Neural Network-based Smart Disaster Management Framework for Real-time Detection and Management of Forest Fires

Introduction: Timely detection of catastrophic natural disasters, such as forest fires, is critical to minimizing losses and ensuring rapid response. Artificial intelligence is increasingly being recognized as a valuable tool in enhancing various stages of disaster management. Method: This paper presents the development of a smart framework utilizing machine learning techniques for real-time detection and monitoring of natural disasters, specifically forest fires. The proposed approach employs a 10-layer convolutional neural network (CNN) that classifies aerial images into Fire, Non-Fire, and Smoke categories with high precision and speed. In addition to this, a CNN-based feature extraction process is performed and integrated with various ML classifiers, including support vector machine, k-nearest neighbor, decision tree, random forest, and extra trees. Results: Extensive performance analysis reveals that the proposed 10-layer CNN model outperforms other classifiers, achieving an accuracy of 97.64% in the binary classification of fire vs. nonfire and 95.61% in the three-class classification of Fire, Non-Fire and Smoke classes. Furthermore, a comparative study with existing state-of-the-art methods demonstrates the proposed model's superior performance in both accuracy and computational complexity. Conclusion: These results demonstrate the potential of the proposed CNN-based framework to serve as a reliable and effective tool for real-time disaster management across various applications, providing valuable support to emergency response teams in mitigating the impact of natural disasters.

Optimizing Fire Scene Analysis: Hybrid Convolutional Neural Network Model Leveraging Multiscale Feature and Attention Mechanisms

The rapid and accurate detection of fire scenes in various environments is crucial for effective disaster management and mitigation. Fire scene classification is a critical aspect of modern fire detection systems that directly affects public safety and property preservation. This research introduced a novel hybrid deep learning model designed to enhance the accuracy and efficiency of fire scene classification across diverse environments. The proposed model integrates advanced convolutional neural networks with multiscale feature extraction, attention mechanisms, and ensemble learning to achieve superior performance in real-time fire detection. By leveraging the strengths of pre-trained networks such as ResNet50, VGG16, and EfficientNet-B3, the model captures detailed features at multiple scales, ensuring robust detection capabilities. Including spatial and channel attention mechanisms further refines the focus on critical areas within the input images, reducing false positives and improving detection precision. Extensive experiments on a comprehensive dataset encompassing wildfires, building fires, vehicle fires, and non-fire scenes demonstrate that the proposed framework outperforms existing cutting-edge techniques. The model also exhibited reduced computational complexity and enhanced inference speed, making it suitable for deployment in real-time applications on various hardware platforms. This study sets a new benchmark for fire detection and offers a powerful tool for early warning systems and emergency response initiatives.

Distributed Fire Classification and Localization Model Based on Federated Learning with Image Clustering

Distributed Fire Classification and Localization Model Based on Federated Learning with Image Clustering

Early fire detection using wavelet based features

In recent years, millions of hectares of vegetation worldwide have been destroyed by wildfires and forest fires. Computer vision-based fire classification, which separates pixels from image or video datasets into fire and non-fire categories, has gained more attention recently due to technological innovations. Fire pixels in an image or video can be classified using either a deep learning strategy or a traditional machine learning approach. Deep learning algorithms could process enormous volumes of data, but their training model performance is constrained because they fail to consider the differences in complexity between training samples. Moreover, it is not obvious to train a deep learning model without a dedicated GPU unit.Similarly, deep learning techniques that have a scarcity of training data and insufficient features exhibit poor performance in intricate real-world fire situations. Consequently, to categorize fire and non-fire pixels from the processed photos from the publicly accessible datasets, Corsican and FLAME, as well as the aerial private dataset Firefront_Gestosa, the current study uses a lightweight technique based on SVM and a refined set of features.The present research implemented a novel framework for fire detection and classification from a variety of RGB and Infra-red images acquired during real missions, addressing the significant requirement for swift and accurate recognition of diverse types of flames, ranging from wildfires to industrial and domestic fires. The framework employs wavelet decomposition-based features, including wavelet length, standard deviation, variance, energy, and Shannon’s entropy, extracted through a sliding window sampling method within a machine learning approach. It should be noted that managing multidimensional data to train a model is difficult in machine learning applications. This issue is solved adopting a feature selection approach, which eliminates redundant or unnecessary data that affects the functionality of the model. Thus, to enhance model performance, feature selection using ranking algorithms based on theoretical mutual information is applied in combination to the Support Vector Machine (SVM) with Radial Basis Function (RBF) kernel.Extensive experiments demonstrate exceptional results, notably with Haar wavelets, achieving an impressive overall accuracy of 99.43% and remarkable performance across specificity, precision, recall, F-measure, and G-mean metrics. Thus, the present study showcases the potential of advanced image processing techniques to significantly advance fire detection and classification, thereby contributing to fire prevention, management, and research in various contexts.

Improving fire severity prediction in south-eastern Australia using vegetation-specific information

Abstract. Wildfire is a critical ecological disturbance in terrestrial ecosystems. Australia, in particular, has experienced increasingly large and severe wildfires over the past 2 decades, while globally fire risk is expected to increase significantly due to projected increases in extreme weather and drought conditions. Therefore, understanding and predicting fire severity is critical for evaluating current and future impacts of wildfires on ecosystems. Here, we first introduce a vegetation-type-specific fire severity classification applied to satellite imagery, which is further used to predict fire severity during the fire season (November to March) using antecedent drought conditions, fire weather (i.e. wind speed, air temperature, and atmospheric humidity), and topography. Compared to fire severity maps from the fire extent and severity mapping (FESM) dataset, we find that fire severity prediction results using the vegetation-type-specific thresholds show good performance in extreme- and high-severity classification, with accuracies of 0.64 and 0.76, respectively. Based on a “leave-one-out” cross-validation experiment, we demonstrate high accuracy for both the fire severity classification and the regression using a suite of performance metrics: the determination coefficient (R2), mean absolute error (MAE), and root-mean-square error (RMSE), which are 0.89, 0.05, and 0.07, respectively. Our results also show that the fire severity prediction results using the vegetation-type-specific thresholds could better capture the spatial patterns of fire severity and have the potential to be applicable for seasonal fire severity forecasts due to the availability of seasonal forecasts of the predictor variables.

Using computer vision to classify, locate and segment fire behavior in UAS-captured images

The widely adaptable capabilities of artificial intelligence, in particular deep learning and computer vision have led to significant research output regarding flame and smoke detection. The composition of flame and smoke, also described as fire behavior, can be considerably different depending on factors like weather, fuels, and the specific landscape fire is being observed on. The ability to detect definable classes of fire behavior using computer vision has not been explored and could be helpful given it often dictates how firefighters respond to fire situations. To test whether types of fire behavior could be reliably classified, we collected and labeled a unique unmanned aerial system (UAS) image dataset of fire behavior classifications to be trained and validated using You Only Look Once (YOLO) detection models. Our 960 labeled images were sourced from over 21 h of UAS video collected during prescribed fire operations covering a large region of Texas and Louisiana, United States. National Wildfire Coordinating Group (NWCG) fire behavior observations and descriptions served as a reference for determining fire behavior classes during labeling. YOLOv8 models were trained on NWCG Rank 1–3 fire behavior descriptions in grassland, shrubland, forested, and combined fire regimes within our study area. Models were first trained and validated on classifying isolated image objects of fire behavior, and then separately trained to locate and segment fire behavior classifications in UAS images. Models trained to classify isolated image objects of fire behavior consistently performed at a mAP of 0.808 or higher, with combined fire regimes producing the best results (mAP = 0.897). Most segmentation models performed relatively poorly, except for the forest regime model at a box (locate) and mask (segment) mAP of 0.59 and 0.611, respectively. Our results indicate that classifying fire behavior with computer vision is possible in different fire regimes and fuel models, whereas locating and segmenting fire behavior types around background information is relatively difficult. However, it may be a manageable task with enough data, and when models are developed for a specific fire regime. With an increasing number of destructive wildfires and new challenges confronting fire managers, identifying how new technologies can quickly assess wildfire situations can assist wildfire responder awareness. Our conclusion is that levels of abstraction deeper than just detection of smoke or flame are possible using computer vision and could make even more detailed aerial fire monitoring possible using a UAS.

LPG Gas Leak Detection System and LPG Fire Classification Based on Internet of Things and Artificial Intelligence with Telegram Bot as a Monitoring Tool

LPG gas leaks pose a serious threat in industrial kitchens as they can cause costly fires, both in terms of material and safety. To improve safety, an accurate detection system is required. This research focuses on developing an LPG gas leak detection system and LPG fire classification with Internet of Things and Artificial Intelligence technology. Supported by Telegram Bot as an emergency notification monitoring tool, this system uses MQ-2 sensors to detect LPG gas leaks and ESP32-Cam to classify LPG fires along with Pretrained-model technology such as Cascade Fire Detection on OpenCV Cloud Server. As the output of this system, the use of PWM control and automation oversees regulating the Exhaust Fan according to the detected leakage. FreeRTOS is also used for system task efficiency, and Port Forwarding with Ngrok Local Server allows public access to the ESP32-Cam. System testing was conducted by Black-Box testing, then evaluating the performance of the MQ-2 sensor against 400 ppm and 1500 ppm thresholds for LPG testing distances in open kitchens and closed kitchens, as well as analyzing system response and delay via HTTP protocol. The results demonstrated the system's success in detecting gas leaks, classifying LPG fires and facilitating emergency communication.

Real-Time Fire Classification Models Based on Deep Learning for Building an Intelligent Multi-Sensor System

Fire detection systems are critical for mitigating the damage caused by fires, which can result in significant annual property losses and fatalities. This paper presents a deep learning-based fire classification model for an intelligent multi-sensor system aimed at early and reliable fire detection. The model processes data from multiple sensors that detect various parameters, such as temperature, humidity, and gas concentrations. Several deep learning architectures were evaluated, including LSTM, GRU, Bi-LSTM, LSTM-FCN, InceptionTime, and Transformer. The models were trained on data collected from controlled fire scenarios and validated for classification accuracy, loss, and real-time performance. The results indicated that the LSTM-based models (particularly Bi-LSTM and LSTM) could achieve high classification accuracy and low false alarm rates, demonstrating their effectiveness for real-time fire detection. The findings highlight the potential of advanced deep-learning models to enhance the reliability of sensor-based fire detection systems.

Accurate classification of forest fires in aerial images using ensemble model

This paper proposes a method to identify forest fires in aerial images using three different convolutional neural networks (CNNs). Unlike general approaches that make use of a single CNN to classify the images, the proposed solution uses the outcomes of different CNNs and considers the most predicted class. This method overcomes the problems associated with using a single CNN, such as low accuracy due to the drawbacks associated with that model. The three different classifiers used here are InceptionV3, VGG-16, and ResNet50. Classification is carried out based on the presence of fire or smoke features in the images. The individual predictions are combined using max-ensembling. The performance is analyzed using metrics like precision, recall, accuracy and F1-score. From the work, it was found that the combined model resulted in an accuracy of 95.8%. The results confirm that the final model provides greater classification accuracy than the individual models. The proposed method can be used to predict forest fires from live aerial images more accurately and help reduce the damage caused.

A Comparative Study of Fire Code Classifications of Building Materials

Whether noncombustible or combustible construction is used, the presence of combustible materials is likely to be used for various reasons, such as interior finishes, flooring, and insulation. Consequently, how regulations consider the degree of combustibility in their fire classifications will influence the level of fire safety provided in these buildings and the exchanges between all actors in the construction sector. In North America, the regulation of combustibility is primarily governed by surface flame spread assessed through the Steiner tunnel test. While there is a growing prevalence of calorimetric methods globally, their incorporation into North American building codes remains notably limited. Based on ISO 5660-1 cone calorimeter test results of twenty commercially available North American building materials, a comparative study was conducted between the Canadian flame spread classification and the classifications in Japan, New Zealand and the European Union (Euroclass). The tests and their limitations are described herein, as well as the conceptual frameworks. The results suggest that as materials’ combustibility levels increase, the level of agreement between classifications decreases and remains binary. The choice between the material and system scales is crucial for determining the effective development and implementation of regulations.

FireSonic: Design and Implementation of an Ultrasound Sensing-Based Fire Type Identification System.

Accurate and prompt determination of fire types is essential for effective firefighting and reducing damage. However, traditional methods such as smoke detection, visual analysis, and wireless signals are not able to identify fire types. This paper introduces FireSonic, an acoustic sensing system that leverages commercial speakers and microphones to actively probe the fire using acoustic signals, effectively identifying fire types. By incorporating beamforming technology, FireSonic first enhances signal clarity and reliability, thus mitigating signal attenuation and distortion. To establish a reliable correlation between fire type and sound propagation, FireSonic quantifies the heat release rate (HRR) of flames by analyzing the relationship between fire-heated areas and sound wave propagation delays. Furthermore, the system extracts spatiotemporal features related to fire from channel measurements. The experimental results demonstrate that FireSonic attains an average fire type classification accuracy of 95.5% and a detection latency of less than 400 ms, satisfying the requirements for real-time monitoring. This system significantly enhances the formulation of targeted firefighting strategies, boosting fire response effectiveness and public safety.

The evolution of reaction to fire classification of materials: A case study of Canada

AbstractCombustible and noncombustible notions have evolved with time, along with the associated fire tests by which legislation classifies building materials. New Zealand, Japan, and Europe are just some of the many legislations that have followed this evolution, except for North American regulations, which remain attached to methods dating back to 1944. To better understand this stagnation in North American practices, this document first traces the evolution of Canadian regulations on fire classification of materials. Then, a parallel is drawn with the evolution of reaction to fire tests mandated in the National Building Code of Canada. Finally, this paper will review the current fire classification of materials concerning the combustibility concept based on the Steiner tunnel test and the flame spread rating criteria. The analysis reveals that the relevance of the test and its results are questionable, and the reciprocity between test measurement and its classification does not always coincide. Despite the revisions made through time, the classification of materials based on their fire properties remains distinctly binary.

Towards early forest fire detection and prevention using AI-powered drones and the IoT

Wildfires are a significant natural hazard, resulting in financial losses, human deaths, and environmental damage. Due to the rising severity and frequency of wildfires, wildfire management and detection have recently received increased attention worldwide. Monitoring potential risk areas and early fire detection are critical factors for shortening the reaction time and reducing the potential damage. Conventional wildfire detection techniques like satellite imaging and remote camera-based sensing need more latency and low reliability. To tackle these limitations, this paper proposes a novel airborne UAV-based IoT (UIoT) system for wildfire sensing, detection, and extinguishing. It presents the design of low-cost and low-maintenance fire-detecting IoT nodes for large-scale deployment. It also proposes, investigates, and reports on several connectivity architectures using the LoRaWAN protocol for UIoT systems. We geolocate forest trees in a terrain-mapping exercise for precise fire localization. We then deploy an autonomous drone with a visual camera that utilizes our novel classification network to detect the presence of fire followed by fire detection and particle filter-based tracking to center fire at the center of the frame. We use a cloud back-end server for monitoring, analysis, and reporting. Our results show that our low-cost and long-range IoT nodes accurately detect fire within 1-5 min after fire ignition. Our fire classification network achieved an accuracy of 99.46% and a mean average precision of 99.64%. Numerical results suggest that the proposed UAV-IoT-based fire classification offers a fast and reliable wildfire recognition and extinguishing solution while minimizing power consumption and increasing the battery lifespan.

Visual fire detection using deep learning: A survey

Visual Fire Detection (VFD), through the rapid and accurate identification of smoke and flame in images and videos, is crucial for early fire warning and reducing fire hazards. In recent years, the introduction of deep learning has significantly advanced this field, especially in the automatic extraction of discriminative features necessary for VFD. This paper provides a comprehensive review of the latest technological advancements in fire detection using deep learning, offering a broad perspective. Initially, it details the publicly available benchmark datasets widely used in VFD research and the corresponding evaluation metrics, providing a basis for researchers to assess the performance of various algorithms. Subsequently, we propose a systematic categorization framework, dividing VFD tasks into three key directions: fire classification, fire localization, and fire segmentation. For these directions, we thoroughly review the innovative improvements in deep learning models tailored for image and video inputs and discusses how these advancements enhance the accuracy and efficiency of fire detection. Finally, we highlight the challenges in the field and explore future research directions, intending to inspire and guide both newcomers and seasoned researchers in this area.

DATFNets-dynamic adaptive assigned transformer network for fire detection

Fires cause severe damage to the ecological environment and threaten human life and property. Although the traditional convolutional neural network method effectively detects large-area fires, it cannot capture small fires in complex areas through a limited receptive field. At the same time, fires can change at any time due to the influence of wind direction, which challenges fire prevention and control personnel. To solve these problems, a novel dynamic adaptive distribution transformer detection framework is proposed to help firefighters and researchers develop optimal fire management strategies. On the one hand, this framework embeds a context aggregation layer with a masking strategy in the feature extractor to improve the representation of low-level and salient features. The masking strategy can reduce irrelevant information and improve network generalization. On the other hand, designed a dynamic adaptive direction conversion function and sample allocation strategy to fully use adaptive point representation while achieving accurate positioning and classification of fires and screening out representative fire samples in complex backgrounds. In addition, to prevent the network from being limited to the local optimum and discrete points in the sample from causing severe interference to the overall performance, designed a weighted loss function with spatial constraints to optimize the network and penalize the discrete points in the sample. The mAP in the three baseline data sets of FireDets, WildFurgFires, and FireAndSmokes are 0.871, 0.909, and 0.955, respectively. The experimental results are significantly better than other detection methods, which proves that the proposed method has good robustness and detection performance.

Remote Sensing and Machine Learning for Accurate Fire Severity Mapping in Northern Algeria

Forest fires pose a significant threat worldwide, with Algeria being no exception. In 2020 alone, Algeria witnessed devastating forest fires, affecting over 16,000 hectares of land, a phenomenon largely attributed to the impacts of climate change. Understanding the severity of these fires is crucial for effective management and mitigation efforts. This study focuses on the Akfadou forest and its surrounding areas in Algeria, aiming to develop a robust method for mapping fire severity. We employed a comprehensive approach that integrates satellite imagery analysis, machine learning techniques, and geographic information systems (GIS) to assess fire severity. By evaluating various remote sensing attributes from the Sentinel-2 and Planetscope satellites, we compared different methodologies for fire severity classification. Specifically, we examined the effectiveness of reflectance indices-based metrics such as Relative Burn Ratio (RBR) and Difference Burned Area Index for Sentinel-2 (dBIAS2), alongside machine learning algorithms including Support Vector Machines (SVM) and Convolutional Neural Networks (CNN), implemented in ArcGIS Pro 3.1.0. Our analysis revealed promising results, particularly in identifying high-severity fire areas. By comparing the output of our methods with ground truth data, we demonstrated the robust performance of our approach, with both SVM and CNN achieving accuracy scores exceeding 0.84. An innovative aspect of our study involved semi-automating the process of training sample labeling using spectral indices rasters and masks. This approach optimizes raster selection for distinct fire severity classes, ensuring accuracy and efficiency in classification. This research contributes to the broader understanding of forest fire dynamics and provides valuable insights for fire management and environmental monitoring efforts in Algeria and similar regions. By accurately mapping fire severity, we can better assess the impacts of climate change and land use changes, facilitating proactive measures to mitigate future fire incidents.

A modified vision transformer architecture with scratch learning capabilities for effective fire detection

Fire is considered to be one of the major influencing factors that cause fatalities, property damage, and economic, and ecological disruption. To perform the early detection of fire using data from vision sensors and prevent or reduce the damage that fires cause, deep learning models have been widely adopted to overcome the limitations of conventional methods. However, mainstream convolutional neural network (CNN) models have limited generalisation abilities in unseen scenarios and struggle to obtain a good trade-off among the accuracy, inference speed, and model size. Currently, vision transformers (ViT) outperform conventional CNN models; however, they are computationally expensive and required more data for training. They provide a limited performance for small and medium-sized datasets, which are very common in the fire scene classification domain. In this work, we employ a novel ViT architecture by combining shifted patch tokenisation and local self-attention modules for efficient fire scene classification and enable the model to learn from scratch even on small and medium-sized datasets. Furthermore, to make the model suitable for real-time inferencing, we modify the transformer encoder and eventually achieve a reduced number of floating-point operations and a reduced model size. Additionally, in this work, a medium-scale fire dataset is developed that contains complex real-world scenarios. Our model is assessed on three benchmark and a self-created datasets using several evaluation metrics, including a novel cross-corpse evaluation metric, as well as a robustness evaluation metric. The experimental results indicate that our model achieved an overwhelmingly better performance compared to existing methods in terms of the accuracy and model complexity.

Fire image detection and classification analysis based on VGG16 image processing model

Fire image classification technology refers to the classification and recognition of fire images through computer vision technology in order to take timely countermeasures. With the development of computer vision technology, fire image classification technology has been widely studied and applied. Deep learning techniques have achieved great success in the field of image classification, with researchers classifying and identifying fire images by using deep learning models such as convolutional neural networks. This paper introduces the research background and application of fire image classification technology, and the experimental results of detection and classification analysis of fire image based on vgg16 image processing model. The model can classify and identify fire images well and achieve good prediction effect. The accuracy of training set and test set is stable at 99%, the accuracy and accuracy of the model reach 98%, and the recall rate and F1 score reach 99%. The application of fire image classification technology is of great significance. By using computer vision technology to classify and identify fire images, it can improve the accuracy and efficiency of fire monitoring and early warning system, timely detect and control fire, and reduce the loss caused by fire. At the same time, the continuous development of deep learning technology also provides a broader space for the research and application of fire image classification technology.

DSRO based data annotation with improved EfficientNet for forest fire detection using image processing in IoT environment

<p>The increasing risk of forest fires demands sophisticated detection systems in order to mitigate the environment effectively. The technology under consideration enhances real-time monitoring and reaction by functioning inside an Internet of Things (IoT) architecture. Even though Artificial Intelligence (AI) algorithms have improved fire detection systems, they are quite expensive and energy-intensive due to their high computing needs. With the use of creative methods for data augmentation and optimization as well as a shared feature extraction module, this research study offers a thorough fire detection model using an improved EfficientNet that tackles these issues. Three technical components are creatively combined in the realm of forest fire detection by this study. The first stage is the use of diagonal swap of random (DSRO) data annotation, which makes use of spatial connections in the data to improve the model’s understanding of complex aspects that are essential for precisely identifying possible fire breakouts. By adding a shared feature extraction module across three functions, the second stage solves difficulties in feature extraction and target identification. This greatly increases the model’s performance in complicated forest scenes while reducing false positives and false negatives. The third and final stage focuses on improving the EfficientNet model’s capacity for accurate forest fire categorization. When taken as a whole, these technical components upon creative combination improve the existing technology in forest fire detection and provide a thorough and practical strategy for reducing environmental hazards. For the purpose of hyperparameter tuning in the EfficientNet for the classification of forest fires, an improved Harris Hawks optimization (HHO) is used. By using the Cauchy mutation approach with adaptive weight, HHO expands the search space, boosts population diversity, and improves overall exploration. By including the sine-cosine algorithm (SCA) into the optimization process, the likelihood of local extremum occurrences is decreased. The proposed strategy is successful compared to other existing models, as shown by the experimental findings that show an improvement of 5% in accuracy compared to the standard existing model, and an improvement of 2% compared to EfficientNet model in detecting forest fire.</p>

FireMatch: A semi-supervised video fire detection network based on consistency and distribution alignment

Deep learning techniques have greatly enhanced the performance of fire detection in videos. However, video-based fire detection models heavily rely on labeled data, and the process of data labeling is particularly costly and time-consuming, especially when dealing with videos. Considering the limited quantity of labeled video data, we propose a semi-supervised fire detection model called FireMatch, which is based on consistency regularization and adversarial distribution alignment. Specifically, we first combine consistency regularization with pseudo-label. For unlabeled data, we design video data augmentation to obtain corresponding weakly augmented and strongly augmented samples. The proposed model predicts weakly augmented samples and retains pseudo-label above a threshold, while training on strongly augmented samples to predict these pseudo-labels for learning more robust feature representations. Secondly, we generate video cross-set augmented samples by adversarial distribution alignment to expand the training data and alleviate the decline in classification performance caused by insufficient labeled data. Finally, we introduce a fairness loss to help the model produce diverse predictions for input samples, thereby addressing the issue of high confidence with the non-fire class in fire classification scenarios. The FireMatch achieved an accuracy of 76.92% and 91.80% on two real-world fire datasets, respectively. The experimental results demonstrate that the proposed method outperforms the current state-of-the-art semi-supervised classification methods.