Early Fire Detection Research Articles

Design of smoke detection system using deep learning and sensor fusion with recursive feature elimination cross-validation

The fire safety system is an important component that controls material and immaterial losses. Fire disasters are generally indicated by the appearance of excess smoke and changes in temperature, pressure, and changes in other parameters in the environment. Conventional smoke sensors are limited in reading parameter changes around their environment, making them less effective in early fire detection. This study aims to design a smoke detection system as an early fire detection system, using sensor fusion based on deep learning using the recursive feature elimination method with cross-validation (RFECV) using a random forest classifier used to select optimal parameters from public datasets as the basis for determining the sensor to be used. Based on the RFECV optimal feature, a deep learning algorithm was performed and obtained an accuracy of 0.99, a precision of 0.99, a recall of 1.00, and an F1 score of 0.99, with a latency time of 34.02 μs, which is 71.76% times faster than the original model.

Experimental study on responsiveness of multi-spectrum infrared flame detectors to diffuse flames partially hidden by obstacles installed in oil and gas processing facilities

Multi-spectrum infrared flame detectors are commonly installed in many oil and gas processing facilities for an early fire detection. Since these facilities are filled with equipment and piping, the visibility of flames from infrared flame detector is a critical factor for reliable and early detection. However, conventional methodology of fire mapping study does not consider the three-dimensional spread of flames. Therefore, this paper focuses on the responsiveness of multi-spectrum infrared flame detectors under the condition that flame is partially hidden by such obstructions.A series of full-scale experiments was conducted to examine the responsiveness of infrared flame detectors to flames partially hidden by shielding objects. These experiments simulated various conditions seen in the oil and gas processing facilities.The experiments have confirmed the relationship between the intensity of infrared energy and the obstruction ratio. Specifically, the infrared energy at a wavelength of 4.504 μm emitted from a flame hidden by shielding objects can be quantified using the obstruction ratio, the ratio of visible area of partially hidden flame. By using these indices, it has been clarified that the infrared energy emitted from a partially hidden flame at the location of an infrared flame detector can be estimated. The responsiveness of detectors depends more on the intensity of the infrared energy near 4.5 μm than on flame pulsation. Multi-spectrum flame detectors, commercially available and employed in this study, exhibited performance equal to or even more sensitive than what is stated in their data sheets. Quantitative evaluation of the responsiveness of flame detectors to partially hidden flames revealed in this study can contribute to the improvement of the fire mapping study for the realistic design evaluation in the oil and gas processing facilities.

Comparative study of teachable machine for forest fire and smoke detection by drone

Forests play a vital role in maintaining ecological equilibrium and serving as vital habitats for wildlife. They regulate global climate, safeguard soil and water resources, and provide crucial ecosystem services such as air and water purification, essential for human well-being and sustainable development. Forest fires wreak havoc on ecosystems and wildlife, emitting harmful pollutants, disrupting communities, and increasing the risk of erosion and landslides. Detecting forest fires through satellite imaging, aerial reconnaissance, and ground-based sensors is pivotal for early detection and containment, safeguarding human lives, wildlife, and preserving natural resources for future generations. Utilizing drones and deep learning (DL) algorithms can significantly enhance early fire detection and minimize their devastating impact. In this paper, we examine teachable machine, a Google tool for creating DL models. We compare the top model generated by teachable machine for fire and smoke detection to models obtained through transfer learning from established DL models in image recognition and computer vision (CV), such as VGG16, VGG19, MobileNet, MobileNetv2, and MobileNetv3. The results underscore the significance of employing the teachable machine model in specific fire and smoke detection scenarios.

Using Cognitive IOT In Vehicle System for On Actual Time Monitoring of CO2 Emission with Temperature and Alcohol Detector

One of the pressing environmental challenges we face today is the phenomenon of global warming, primarily driven by the release of greenhouse gases. Among these gases, carbon dioxide stands out as a significant contributor, with its emissions leading to a warming effect on the Earth's surface. Protecting our environment requires us to address and manage these changes effectively, presenting a considerable challenge. For long-term control of carbon dioxide emissions at their source, there is a growing recognition of the need for preventive and control technologies. This paper aims to employ IoT technology, utilizing Raspberry Pi's CO2 sensitivity, to monitor emissions from various sources such as public transportation, industries, and forest fires. Continuous monitoring of carbon dioxide levels within a city will help identify areas with the highest pollution levels. Additionally, this initiative seeks to implement an intelligent system for early detection of forest fires, also known as wildfires. These uncontrolled fires in natural areas pose significant threats to both ecosystems and human resources. It's noteworthy that wildfires emit more CO2 gas than initially assumed, impacting state climate targets. Integrating these efforts with IoT enhances security and enables various services, including the deployment of a Simple Notification Service (SNS) to alert mobile users about high CO2 levels in specific areas. Furthermore, this model aims to extend its capabilities to include the detection of temperature in heated vehicle components, such as engines, using temperature detectors. Moreover, alcohol detectors will be utilized to monitor alcohol consumption by vehicle drivers, enhancing safety measures on the road. Keywords: SNS (Simple Notification Service), IoT (Internet of Things), GPS (Global Positioning System), GVG (Green Vehicle Guide), MAQUMON (Mobile Air Quality Monitoring Network).

A lightweight multiscale smoke segmentation algorithm based on improved DeepLabV3+

AbstractFires not only cause devastating consequences for human life and property, but also lead to soil erosion in forests. Therefore, it is necessary to design a novel algorithm that can quickly monitor smoke from fires. Most existing smoke segmentation methods do not consider the segmentation accuracy of algorithms under limited computational resources. To address this research gap, this paper proposes a lightweight smoke segmentation algorithm based on DeepLabV3+ that achieves fast inference speed and high accuracy for different sizes smoke. To reduce the number of parameters, the feature extraction network of the DeeplabV3+ algorithm is replaced by MobileNetV2, which enhances the extraction ability of the algorithm in segment smoke images. Then, the Convolutional Block Attention Module (CBAM) is added to the encoder part to enhance the perception of the algorithm for small smoke and effectively alleviates smoke mis‐segmentation. Furthermore, a newly designed loss function is used in the network. The experimental results show that the proposed method has improved by 1.27% in Smoke IoU and 1.21% in mPA compared with other methods. The weight size has been reduced to 10.76% of the original DeepLabV3+, and the inference time is only 33.71ms. Therefore, it is a more suitable early fire detection algorithm for resource‐constrained environments.

SnO2/Au Microelectromechanical Systems Modified by Oxygen Vacancies for Enhanced Sensing of Dioctyl Phthalate.

Dioctyl phthalate (DOP) serves as a characteristic gas utilized in early electrical fire detection, its detection offers promising prospects for the prevention of electrical fires. In this study, we employed a modified photodeposition method to prepare Tin dioxide (SnO2) materials co-modified with Au and oxygen vacancies. Subsequently, microelectromechanical systems (MEMS) gas sensor for DOP detection were fabricated, utilizing 0.5 %Au/SnO2-I as the sensing material. Characterization results reveal the presence of abundant oxygen vacancies in 0.5 %Au/SnO2-I. The synergistic interplay of Au and oxygen vacancies resulted in a remarkable response of 9.98 to 20 ppm of DOP at operational temperature of 250 °C. This represents a significant 96 % enhancement in comparison to the response value of 4.50 exhibited by pure SnO2 at 300 °C. Notably, this gas sensor boasts low power consumption and demonstrates a quick response in the detection of overheating polyvinyl chloride (PVC) cables under simulated conditions.

Research on a capacitive particle analysis smoke detector.

Smoke detectors face the challenges of increasing accuracy, sensitivity, and high reliability in complex use environments to ensure the timeliness, accuracy, and reliability of very early fire detection. The improvement and innovation of the principle and algorithm for smoke particle concentration detection provide opportunities for improving the performance of the detector. This study represents a new refinement of the smoke concentration detection principle based on capacitive detection of cell structures, and detection signals are processed by a multiscale smoke particle concentration detection algorithm to calculate smoke concentration. Through experiments, it was found that the detector provides effective detection of smoke particle concentrations ranging from 0 to 10% obs/m; moreover, when the detection accuracy is greater than a certain number of parts per million (PPM), the sensitivity of the detector can reach the PPM level; furthermore, the detector can detect smoke particle concentrations higher than the PPM level accuracy even in an environment with a certain concentration of petroliferous and dust particles of different sizes.

Using Transfer Learning and XGBoost for Early Detection of Fires in Offshore Wind Turbine Units

To improve the power generation efficiency of offshore wind turbines and address the problem of high fire monitoring and warning costs, we propose a data-driven fire warning method based on transfer learning for wind turbines in this paper. This paper processes wind turbine operation data in a SCADA system. It uses an extreme gradient-boosting tree (XGBoost) algorithm to build an offshore wind turbine unit fire warning model with a multiparameter prediction function. This paper selects some parameters from the dataset as input variables for the model, with average cabin temperature, average outdoor temperature, average cabin humidity, and average atmospheric humidity as output variables. This paper analyzes the distribution information of input and output variables and their correlation, analyzes the predicted difference, and then provides an early warning for wind turbine fires. This paper uses this fire warning model to transfer learning to different models of offshore wind turbines in the same wind farm to achieve fire warning. The experimental results show that the prediction performance of the multiparameter is accurate, with an average MAPE of 0.016 and an average RMSE of 0.795. It is better than the average MAPE (0.051) and the average RMSE (2.020) of the prediction performance of a backpropagation (BP) neural network, as well as the average MAPE (0.030) and the average RMSE (1.301) of the prediction performance of random forest. The transfer learning model has good prediction performance, with an average MAPE of 0.022 and an average RMSE of 1.469.

Innovative pre-fire alert smart detection system-based embedded system

Early fire identification is critical to saving lives and property. One viable strategy for improving fire detection systems’ efficacy and precision is the use of Internet of Things (IoT) technology. This article details research into the conception and development of an IoT, ESP32-based, Arduino MEGA-connected smart pre-alert detection system for fires. Because of its flexible nature, the Arduino microcontroller can easily exchange data with various sensors for real-time monitoring and analysis. This study surveys the existing literature to provide insight into many approaches to early fire detection (forecasting), such as sensor fusion methods and the use of the intelligence concept as a pre-predictor. The research highlights the need to set up sophisticated fire monitoring and alarm systems and to develop predictive models based on sensor data. This study illustrates how embedded systems via IoT technology and the ESP32 architecture may improve the efficiency and dependability of very early fire detection systems. This research adds to the expanding corpus of information that may be used to improve fire safety in contemporary settings. The proposal’s sensors include an Arduino MEGA, an ESP32, a MAX6675 module, an MQ-Smoke, and an MQ-Natural Gas for collecting data, which are then processed using Cloud Firebase and a sleek mobile interface that can be reached overseas at any time.

Early detection of deep-seated smouldering fires in wood waste storage using ERT

Incidents of waste and biofuel fires are common at all stages of the waste recycling chain and have grave implications for business, employees, firefighters, society, and environment. An early detection of waste and biofuel fires in the smouldering stage could save precious lives, resources, and our environment. Existing fire detection methodologies e.g. handheld temperature sensors, IR cameras, gas sensors, and video and satellite-based monitoring techniques have inherent limitations to efficiently detect smouldering fires. An attempt was made to explore the potential of electrical resistivity tomography (ERT) as an alternate tool to address the problem. In the experiments an externally powered resistive wire was employed to initiate the smouldering fire inside the test material (wood pellets, wood shavings, wood fines). Time series of ERT that followed the initiation and development of smouldering were recorded using an automated monitoring instrument setup. The actual geometry of the experimental sample container and electrode setup was integrated in the 3D finite element method (FEM) model grid to perform inverse numerical modelling (inversion) and to develop resistivity tomographic images. The study shows a sharp increase in ratio of resistivity (R/Ro ≥ 50 %) in the test material in the region of smouldering hotspot and demonstrates the potential use of ERT technique for the detection of smouldering hotspots in silos and pile storage of organic material such as wood-based fuels, wood waste, coal, municipal solid waste (MSW), recyclables etc. More research is however required for enabling the use of this technique at the practical scale for different storage conditions.

Fire detection and suppression in rooms with different geometries

The paper presents experimental research findings on the integral characteristics of physicochemical processes in the area around a fire at different stages of its development and suppression. The purpose of this study was to determine the characteristics of the operation of fire detectors in rooms of various geometries. The temperatures in different cross sections, as well as the composition of gaseo us products of thermal decomposition and flame combustion of materials were investigated. The study dealt with class A fires involving wood. Mock-ups of rooms with distinctly different geometries were used: a parallelepiped, a rectangular truncated pyramid and a bow-roof (arched) structure. Efficient arrangement of technical equipment (light and gas analysis sensors, heat and smoke detectors, nozzles) was determined to provide early detection of a fire, prompt initiation of suppression and completion of combustion and smoldering. Minimum volumes of a firefighting liquid (water in this case) and spraying times were identified that are necessary and sufficient to suppress combustion. Guidelines were proposed on the arrangement of technical equipment for early fire detection and efficient containment of fire in the shortest time. A mathematical apparatus has been developed that allows predicting the performance of fire detectors and sensors. The research results can be used to evaluate the most effective locations for fire sensors in the design of buildings and structures.

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.

Rational Design of Amorphous Carbon-Coated Laminar-Structured Wood for Integrating Repeatable Early Fire Detection and High-Temperature Affordable Flexible Pressure Sensing in One System.

High-temperature affordable flexible polymer-based pressure sensors integrated with repeatable early fire warning service are strongly desired for harsh environmental applications, yet their creation remains challenging. This work proposed an approach for preparing such advanced integrated sensors based on silver nanoparticles and an ammonium polyphosphate (APP)-modified laminar-structured bulk wood sponge (APP/Ag@WS). Such integrated sensors demonstrated excellent fire warning performance, including a short response time (minimum of 0.44 s), a long-lasting alarm time (>750 s), and reliable repeatability. Moreover, it achieved high-temperature affordable flexible pressure sensing that exhibited an almost unimpaired working range of 0-7.5 kPa and a higher sensitivity (in the low-pressure range, maximum to 226.03 kPa-1) after fire. The high stability was attributed to reliable structural elasticity, and the wood-derived amorphous carbon is capable of repeatable fire warnings. Finally, a Ag@APP/WS-based wireless fire alarm system that realized reliable house fire accident detection was demonstrated, showing great promise for smart firefighting application.

The Smart Agriculture System Using IOT and ML

Numerous countries have abundant resources, including land, rivers, groundwater, the environment, with agriculture serving as the primary source of income for many people in country. Nonetheless, resource shortage has grown in recent decades, particularly for groundwater and river water, so remote control smart irrigation system is developed to avoid access use and loss of water in the agriculture. A lack of information about the best use of available resources leads to increased resource consumption and worse crop yields. Choosing crops that are unsuited for the soil or climate reduces both quality and quantity of crop. As a result, system that recommends appropriate crops and fertilizers based on soil NPK values, soil colour, Season, PH, Rainfall, temperature etc has been developed. Fires are one of the most destructive global disasters, thus early detection is necessary to avoid the damages in agriculture. To reduce losses, an automated system capable of early fire detection through alarm systems and prompt extinguishing procedures through the roof structure pipeline in the farm is developed. Another use of that same pipeline is for spraying the pesticides and fertilizers. A potential resolution to address all these challenges involves the development and implementation of an intelligent agricultural system integrating Internet of Things (IOT) and Machine Learning technologies.

YOLO-Based Models for Smoke and Wildfire Detection in Ground and Aerial Images

Wildland fires negatively impact forest biodiversity and human lives. They also spread very rapidly. Early detection of smoke and fires plays a crucial role in improving the efficiency of firefighting operations. Deep learning techniques are used to detect fires and smoke. However, the different shapes, sizes, and colors of smoke and fires make their detection a challenging task. In this paper, recent YOLO-based algorithms are adopted and implemented for detecting and localizing smoke and wildfires within ground and aerial images. Notably, the YOLOv7x model achieved the best performance with an mAP (mean Average Precision) score of 80.40% and fast detection speed, outperforming the baseline models in detecting both smoke and wildfires. YOLOv8s obtained a high mAP of 98.10% in identifying and localizing only wildfire smoke. These models demonstrated their significant potential in handling challenging scenarios, including detecting small fire and smoke areas; varying fire and smoke features such as shape, size, and colors; the complexity of background, which can include diverse terrain, weather conditions, and vegetation; and addressing visual similarities among smoke, fog, and clouds and the the visual resemblances among fire, lighting, and sun glare.

F2M: Ensemble-based uncertainty estimation model for fire detection in indoor environments

Early fire detection and timely notification are paramount for preventing human and material casualties caused by fire. As a result, scientists have developed various fire monitoring systems based on sensors and images. Image-based systems have proven more advantageous than sensor-based ones as they provide additional details about the fire, such as its location, intensity, and progression. However, accurately determining the shape and boundaries of flames in image-based systems is difficult due to various factors, including backgrounds, different fire sizes, and interference from objects that resemble flames. In this study, we first evaluate convolutional neural networks used in related research and benchmark them on our challenging indoor fire dataset. Secondly, we propose a novel Feature Merging Model (F2M), which combines the output fire segmentation masks of the top five segmentation models obtained through the performed evaluation. The F2M relies on a novel mechanism that introduces bias into the mask estimation and achieves better results than all the tested methods, along with Monte-Carlo dropout for uncertainty estimation. The F2M ensemble-based uncertainty estimation model achieved improvements in comparison to the best performance convolutional neural network U-Net++, on three different image resolutions: 256 × 256, 640 × 640 and 800 × 800, according to the following metrics: Total error, Dice coefficient, and IoU score.

Unraveling the Dynamics of User Acceptance on the Internet of Things: A Systematic Literature Review on the Theories and Elements of Acceptance and Adoption

The Internet of Things (IoT) connects various devices to the internet, enabling seamless communication and coordination. Its sensor and action capabilities aid in early fire detection and response. This article investigates the theory and elements of Internet of Things (IoT) acceptance and acceptance to see how IoT can assist in monitoring and controlling fire safety management. This study uses a systematic literature review (SLR) methodology to search, identify, appraise, synthesize, analyze, and summarize studies on the Internet of Things (IoT). The study highlights the importance of theoretical frameworks like TPB, TAM, UTAUT, and TOE in understanding the Internet of Things (IoT) acceptance and adoption. TPB focuses on attitudes, subjective standards, and perceived behavioral control, TAM emphasizes ease of use, UTAUT focuses on performance expectancy, effort expectancy, social influence, and facilitating conditions, and TOE examines organizational, technological, and environmental aspects affecting commercial IoT adoption. These frameworks provide actionable insights for successful IoT implementation and integration in individual and organizational settings.

Early fire detection technology based on improved transformers in aircraft cargo compartments

The implementation of early and accurate detection of aircraft cargo compartment fire is of great significance to ensure flight safety. The current airborne fire detection technology mostly relies on single-parameter smoke detection using infrared light. This often results in a high false alarm rate in complex air transportation environments. The traditional deep learning model struggles to effectively address the issue of long-term dependency in multivariate fire information. This paper proposes a multi-technology collaborative fire detection method based on an improved transformers model. Dual-wavelength optical sensors, flue gas analyzers, and other equipment are used to carry out multi-technology collaborative detection methods and characterize various feature dimensions of fire to improve detection accuracy. The improved Transformer model which integrates the self-attention mechanism and position encoding mechanism is applied to the problem of long-time series modeling of fire information from a global perspective, which effectively solves the problem of gradient disappearance and gradient explosion in traditional RNN (recurrent neural network) and CNN (convolutional neural network). Two different multi-head self-attention mechanisms are used to classify and model multivariate fire information, respectively, which solves the problem of confusing time series modeling and classification modeling in dealing with multivariate classification tasks by a single attention mechanism. Finally, the output results of the two models are fused through the gate mechanism. The research results show that, compared with the traditional single-feature detection technology, the multi-technology collaborative fire detection method can better capture fire information. Compared with the traditional deep learning model, the multivariate fire prediction model constructed by the improved Transformer can better detect fires, and the accuracy rate is 0.995.

ISSUES OF CREATING A SYSTEM FOR EARLY DETECTION OF FOREST FIRES BASED ON UNMANNED AERIAL VEHICLES

The urgent importance of the task of monitoring the state of forests lies in the danger of forest fires. Forest fires are a frequent occurrence in almost all continents. Despite the high information content of the spectral signs of a forest fire, the corresponding functionality is not fully used in the known implementations of the forest fire early detection system. The spectral region in which it is possible to distinguish forest areas in a state of fire from their usual state is quite wide and covers the middle and thermal zone of the infrared part of the spectrum. The installation of spectral equipment (multi- and hyperspectrometers) on board the UAV (unmanned aerial vehicle) can certainly expand the functionality of unmanned equipment in the early detection of forest fires. The optimal procedure for using UAVs equipped with hyperspectral equipment for early detection of forest fires is proposed. An optimization problem has been formulated and solved to determine the most appropriate mode of changing the UAV flight altitude with an increase in the wavelength of the measurements being carried out. According to the proposed method, on the right branch of the Planck curve, with increasing wavelength, the flight altitude should be increased, and on the left branch it should be reduced. This order corresponds to the case of increasing the wavelength on both branches and follows from the integral restriction set on the function of the dependence of the flight altitude on the wavelength, which comes from the limited real energy resources on board the UAV. Based on the developed method, a new technique for early detection of forest fires based on the use of hyperspectral method is proposed.

A comprehensive survey on several fire management approaches in wireless sensor networks

The majority of the fires are activated through environmental reasons although a minority of them are self-activated. To detect fires several safety systems were introduced. There are wired systems, cameras, satellite systems, and bluetooth feasible to provide a complete image of the world but after a long search period. These systems are not perfect since it prevents fire from finding just at the time, the fire initiates. But, recent technological development in wireless sensor networks (WSN) has spread out its fire detection application. A comprehensive survey on several fire management approaches in WSN propose to discuss various fire detection approaches like early fire detection, energy efficient fire detection, mobile agent-based fire detection, unmanned aerial vehicle (UAV)-based fire detection, threshold-based fire detection, machine learning based fire detection and secure fire detection approaches. Moreover, the comprehensive tabular study of the fire management technique is given that will assist in the suitable selection of approaches to be applied for the detection of fire. Furthermore, WSN uses the clustering method to minimize redundant dataandsecure fire detection approaches collect authenticated data related to fire detection. Early fire detection approaches detects the fire early. Machine learning algorithm detects the fire efficiently.