Fire Detection Research Articles

Evaluasi Trade-off Akurasi dan Kecepatan YOLOv5 dalam Deteksi Kebakaran pada Edge Devices

Real-time object detection using the YOLO (You Only Look Once) algorithm has shown promising performance in various computer vision applications. However, its application on devices with limited resources is still a challenge due to its high computational requirements. This study aims to optimize the YOLOv5 model for fire and smoke detection on Orange Pi Zero 3 devices using quantization techniques. Using a dataset of 2247 fire and smoke images, this study applies static quantization techniques to improve model efficiency. The methodology includes training of standard YOLOv5 models, conversion to ONNX format, and application of static quantization. Results show a significant improvement in computational efficiency, with a 42.2% reduction in model size and a 65.21% increase in inference speed. Despite a decrease in the mAP value by 25.6%, the optimized model was still able to perform object detection at a significantly higher speed. In conclusion, the quantization technique is effective in optimizing the YOLOv5 model for deployment on edge computing devices, despite the trade-off between speed and accuracy.

Fire image enhancement method based on generative adversarial networks for improving fire detection performance through cameras

Fire image enhancement method based on generative adversarial networks for improving fire detection performance through cameras

Micro-Incubator Protocol for Testing a CO2 Sensor for Early Warning of Spontaneous Combustion

A protocol for detecting the potential occurrence of spontaneous combustion (SC) in stored cottonseeds and peanuts using a micro-incubator is described. The protocol indicates how to quantify CO2 production rates and final CO2 levels in wet versus dry cottonseed and peanut samples, which can provide crucial data for the early detection of SC risk in storage facilities. The experimental design utilizes a micro-incubator to simulate conditions found in large bulk crop storage. Parameters monitored include CO2 concentration, temperature, and relative humidity. The protocol includes preparation methods, experimental procedures for both control (dry) and wet seed tests, and test termination criteria that allow for safe experimentation of likely pathogenic fungi. The protocol has three replicates for wet and dry conditions. The protocol is intended to facilitate future experimental studies and ultimately contribute to the development of a consistently reliable early warning fire detection system for SC in cottonseed and peanut warehouse facilities. A consistently reliable fire detection system would address a critical need in the cotton and peanut industry for improved fire risk management and insurability of storage facilities.

Modeling of Wildfire Digital Twin: Research Progress in Detection, Simulation, and Prediction Techniques

Wildfires occur frequently in various regions of the world, causing serious damage to natural and human resources. Traditional wildfire prevention and management methods are often hampered by monitoring challenges and low efficiency. Digital twin technology, as a highly integrated virtual simulation model, shows great potential in wildfire management and prevention. At the same time, the virtual–reality combination of digital twin technology can provide new solutions for wildfire management. This paper summarizes the key technologies required to establish a wildfire digital twin system, focusing on the technical requirements and research progress in fire detection, simulation, and prediction. This paper also proposes the wildfire digital twin (WFDT) model, which integrates real-time data and computational simulations to replicate and predict wildfire behavior. The synthesis of these techniques within the framework of a digital twin offers a comprehensive approach to wildfire management, providing critical insights for decision-makers to mitigate risks and improve emergency response strategies.

Fire Fighter Robot Using IOT and Mobile Application

In today's fast-paced world, fire hazards pose significant risks to life, property, and the environment. Traditional fire detection systems often rely on human intervention, leading to delays in response time and exacerbating the damage caused by fire incidents. Our project, the IoT-Based Automatic Fire Detection and Extinguisher Robot, addresses this challenge by utilizing a combination of modern sensor technology, robotics, and IoT to provide an autonomous and real-time solution. The robot is equipped with flame and temperature sensors that allow it to detect fires at an early stage. Once a fire is detected, the robot activates its movement mechanism using DC motors to navigate toward the fire while the integrated servo-controlled water pump precisely sprays water to extinguish the fire. In addition to automatic firefighting, the robot features obstacle detection using ultrasonic sensors, ensuring safe navigation in complex environments. The system includes Bluetooth and Wi-Fi communication, allowing for remote control via an Android mobile app. This app provides both manual and automatic control options and sends real-time fire alerts when the robot detects a fire. This autonomous system minimizes human intervention, speeds up response times, and reduces fire damage effectively. The robot is powered by an ESP32 microcontroller, which processes data from the sensors and controls the motors and water pump. The rescue pod feature allows the robot to perform additional operations like transporting sensitive materials during emergencies. The system's IoT functionality enables real-time alerts and remote monitoring, enhancing its capability as a scalable and cost-effective solution for residential, commercial, and industrial fire safety.

Design and implementation of a smoke alarm system using YOLOv8

Abstract. This paper present a smoke alarm system built using Yolov8, a state-of-the-art object detection model. The system is designed to detect smoke in real-time environments and alert users in the event of a fire or smoke detection. The use of Yolov8 in this system offers several advantages over other object detection models, making it a suitable choice for building an effective smoke alarm system. One of the critical advantages of Yolov8 is its high accuracy in detecting objects with a small size, such as smoke. This is achieved through its efficient and robust architecture, which allows it to process images quickly and accurately. In comparison to other object detection models, Yolov8 demonstrates superior performance in terms of speed and accuracy. Furthermore, Yolov8s ability to detect objects with high precision and low false positives is crucial for a smoke alarm system. This is because smoke detection needs high accuracy to prevent false alarms and promptly detect real threats. The smoke alarm system built using Yolov8 can be implemented in various environments, including homes, offices, and public buildings. It can provide early warnings for fires and smoke, enabling timely evacuation and reducing potential injuries or damages. The implementation of this system can greatly contribute to improving safety in various environments by providing early warnings of fires and smoke.

Assessment of active fire detection in Serra da Canastra National Park using MODIS and VIIRS sensors

Assessment of active fire detection in Serra da Canastra National Park using MODIS and VIIRS sensors

A Multi-Scale Approach to Early Fire Detection in Smart Homes

In recent years, advancements in smart home technologies have underscored the need for the development of early fire and smoke detection systems to enhance safety and security. Traditional fire detection methods relying on thermal or smoke sensors exhibit limitations in terms of response time and environmental adaptability. To address these issues, this paper introduces the multi-scale information transformer–DETR (MITI-DETR) model, which incorporates multi-scale feature extraction and transformer-based attention mechanisms, tailored specifically for fire detection in smart homes. MITI-DETR achieves a precision of 99.00%, a recall of 99.50%, and a mean average precision (mAP) of 99.00% on a custom dataset designed to reflect diverse lighting and spatial conditions in smart homes. Extensive experiments demonstrate that MITI-DETR outperforms state-of-the-art models in terms of these metrics, especially under challenging environmental conditions. This work provides a robust solution for early fire detection in smart homes, combining high accuracy with real-time deployment feasibility.

Fire detectors and their connection suitable for application in residential, commercial or industrial buildings

Fire detectors and their connection suitable for application in residential, commercial or industrial buildings

FireNet: A Lightweight and Efficient Multi-Scenario Fire Object Detector

Fire and smoke detection technologies face challenges in complex and dynamic environments. Traditional detectors are vulnerable to background noise, lighting changes, and similar objects (e.g., clouds, steam, dust), leading to high false alarm rates. Additionally, they struggle with detecting small objects, limiting their effectiveness in early fire warnings and rapid responses. As real-time monitoring demands grow, traditional methods often fall short in smart city and drone applications. To address these issues, we propose FireNet, integrating a simplified Vision Transformer (RepViT) to enhance global feature learning while reducing computational overhead. Dynamic snake convolution (DSConv) captures fine boundary details of flames and smoke, especially in complex curved edges. A lightweight decoupled detection head optimizes classification and localization, ideal for high inter-class similarity and small targets. FireNet outperforms YOLOv8 on the Fire Scene dataset (FSD) with a mAP@0.5 of 80.2%, recall of 78.4%, and precision of 82.6%, with an inference time of 26.7 ms. It also excels on the FSD dataset, addressing current fire detection challenges.

Fighting against forest fire: A lightweight real-time detection approach for forest fire based on synthetic images

Forest fires are known for their high level of randomness and unpredictability, which often lead to significant ecological damage and human life loss. Existing forest fire detection technologies are not capable of detecting small-scale flames or smoke in real time, thus failing to meet the demands of real-time detection of forest fires using Unmanned Aerial Vehicles (UAVs). To overcome these limitations, we propose an efficient and lightweight forest fire detection method that utilizes synthetic images and UAVs to achieve real-time and high-precision detection of forest fires against complex backgrounds. Firstly, we propose the Dilation Repconv Cross Stage Partial Network (DRCSPNet), which enhances the detection capabilities for multiscale flames and smoke using multi-branch parallel joint dilation convolution and batch normalization, while effectively extracting features from different stages of forest fires. Secondly, to mitigate challenges associated with extreme lighting in forest scenes and large contrast variation in fire images, we propose a Global Mixed-Attention (GMA) model across feature pyramids to enhance information lost in high-dimensional feature maps and increase the robustness of the model through a multiscale fusion strategy. Finally, we present the Lite-Path Aggregation Network (Lite-PAN) with varying scales to improve effective feature flow for multilevel forest fires, addressing challenges that arise from various climatic conditions. Furthermore, we employ Unreal Engine 5 to generate forest fire datasets in four scenarios to address the issue of relatively limited aerial forest fire datasets. According to the results of the experiment, our proposed method achieves 58.39% mAP(mean Average Precision) with 5.703 GFLOPs (Giga Floating Point Operations Per Second) while yielding a frame rate of 33.5 Frames Per Second (FPS) on NVIDIA Jetson NX. Extensive experiment results demonstrate our method has the advantage of being in real time, extremely accurate, and easily implementable compared to state-of-the-art techniques.

Experimental Investigation of Thermal Runaway Behaviour and Inhibition Strategies in Large-Capacity Lithium Iron Phosphate (LiFePO4) Batteries for Electric Vehicles

Large-scale lithium-ion batteries play an important role in the electric vehicle market and the thermal runaway (TR) behaviour of batteries is one of the significant threats to the passengers. In this study, we conducted a series of thermal abuse tests concerning single battery and battery box to investigate the TR behaviour of a large-capacity (310 Ah) lithium iron phosphate (LiFePO4) battery and the TR inhibition effects of different extinguishing agents. The study shows that before the decomposition of the solid electrolyte interphase (SEI) film, temperature consistency within the battery were maintained and it was advisable to use temperature detectors. With the destruction of the internal structure of the battery and the uneven temperature distribution at the subsequent battery reaction stage, a large amount of combustible gas was released when the safety valve was opened. The combustible gas fire detector is used to improve alarm reliability and provided a stable early warning response time. Based on this TR behaviour, a LiFePO4 battery box with a standard size of 1060 × 660 × 250mm was constructed. The TR inhibition effect of two different extinguishing agents (heptafluoropropane and perfluorohexanone) were investigated using a metal tube spraying method with a spraying dose of 1.8kg (spraying rate of 0.06kg/s) and a spraying pressure of 2.5MPa. The results showed that using either of these, satisfy the requirements for TR inhibition in large-capacity LiFePO4 batteries and the flames could be completely extinguished without any reignition, and the chain reaction of the surrounding batteries was effectively prevented. Moreover, heptafluoropropane was proven to have a shorter fire-extinguishing time, whereas perfluorohexanone had the advantage of inhibiting temperature rise. This study provides valuable experimental results for early warning and inhibition of TR in large-capacity LiFePO4 batteries for electric vehicles.

Multi-Sensor Photoelectric Fire Alarm Device Implementation for Early Fire Detection in Campsites

With the growing popularity of leisure activities such as camping and glamping, the incidence of fires at camping sites has increased. This study focuses on improving the effectiveness of photoelectric fire alarm devices by incorporating temperature and humidity data for early fire detection in confined spaces, such as campsites. This study proposes a novel multi-sensor fire alarm system that dynamically adjusts fire detection threshold values based on temperature and humidity data collected by unmanned automatic weather observation systems. The prototype, which was implemented using Raspberry Pi and multiple sensors, demonstrated approximately 20% faster fire detection speed than existing photoelectric fire alarm systems, as verified through experiments in a simulated camping environment. The proposed approach is expected to advance fire alarm systems, enabling faster and more accurate fire detection in diverse environments, particularly at campsites.

GNSS-R-Based wildfire detection: a novel and accurate method

ABSTRACT In recent years, global forests have faced frequent wildfires due to climate change, leading to significant ecological and economic losses. Traditional fire monitoring technologies like MODIS and VIIRS are often hindered by cloud cover and smoke. This study explores the potential of Global Navigation Satellite System Reflectometry (GNSS-R) for fire monitoring in complex environments. By analyzing cumulative burned areas in southeastern Brazil, a new fire detection method combining GNSS-R with Convolutional Neural Networks (CNN) is proposed to improve accuracy and response time. Results show high agreement between predicted and actual fire areas, with CNN outperforming Random Forest (RF) by 3.3% in prediction accuracy and achieving a 17.4% increase in model fitting. Error analysis indicates no systematic errors, with CNN exhibiting lower fire detection errors and better spatial alignment with actual conditions than RF, which tends to overestimate. This study provides a novel approach to fire detection using GNSS-R, with significant implications for fire prediction and management strategies.

Planning And Installation of Water Spray System in The Underground for The Fire Hazard Prevention in PC. GKBI Production Unit

Fire is a serious risk that can cause great losses, both moral and material. This study evaluates fire prevention and control at PC. GKBI Medari, which recently experienced a major fire due to lack of fire detection. Using the AS/NZS 4360: 2004 method, nine hazards with high, medium, and low risks were found. In the underground area of the cambric factory, there are deficiencies in the existing fire extinguishing system, especially in dealing with invisible fires. Hot and dirty air is sucked and channelled through PVC ducts to the air filter, but this system is not effective in extinguishing fires. This study recommends the installation of water spray at several underground points to improve fire protection.

INOVASI IOT UNTUK PENGELOLAAN DAN KEAMANAN RUANG ARSIP: IMPLEMENTASI ESP32 DENGAN SENSOR API DAN SUHU DHT11

This study aims to design and develop a prototype IoT-based archive room monitoring system to enhance security and efficiency in archive management at the Village Office of Desa Pingit, Pringsurat District, Temanggung Regency. The system utilizes a UV Flame Detector for fire detection and a DHT11 sensor for monitoring temperature and humidity, with an ESP32 as the main controller connected via Wi-Fi. Integrated with the Telegram app, users can receive real-time notifications and updates on archive room conditions, facilitating remote monitoring. Testing results indicate that the system effectively detects fire presence and accurately provides temperature and humidity updates. This innovation is expected to aid in risk management and archive protection, offering a practical solution for modern archive room management.

A Safe Fiber-Optic-Sensor-Assisted Industrial Microwave-Heating System.

Industrial microwave-heating systems are pivotal in various sectors, including food processing and materials manufacturing, where precise temperature control and safety are critical. Conventional systems often struggle with uneven heat distribution and high fire risks due to the intrinsic properties of microwave heating. In this work, a fiber-optic-sensor-assisted monitoring system is presented to tackle the pressing challenges associated with uneven heating and fire hazards in industrial microwave systems. The core innovation lies in the development of a sophisticated fiber-optic 2D temperature distribution sensor and a dedicated fire detector, both designed to significantly mitigate risks and optimize the heating process. Experimental results set the stage for future innovations that could transform the landscape of industrial heating technologies toward better process quality.

Real-Time Forest Fire Detection with Lightweight CNN Using Hierarchical Multi-Task Knowledge Distillation

Forest fires pose a significant threat to ecosystems, property, and human life, making their early and accurate detection crucial for effective intervention. This study presents a novel, lightweight approach to real-time forest fire detection that is optimized for resource-constrained devices like drones. The method integrates multi-task knowledge distillation, transferring knowledge from a high-performance DenseNet201 teacher model that was trained on a hierarchically structured wildfire dataset. The dataset comprised primary classes (fire vs. non-fire) and detailed subclasses that account for confounding elements such as smoke, fog, and reflections. The novelty of this approach lies in leveraging knowledge distillation to transfer the deeper insights learned by the DenseNet201 teacher model—specifically, the auxiliary task of recognizing the confounding elements responsible for false positives—into a lightweight student model, enabling it to achieve a similar robustness without the need for complex architectures. Using this distilled knowledge, we trained a MobileNetV3-based student model, which was designed to operate efficiently in real-time while maintaining a low computational overhead. To address the challenge of false positives caused by visually similar non-fire elements, we introduced the Confounding Element Specificity (CES) metric. This novel metric, made possible by the hierarchical structure of the wildfire dataset, is unique in its focus on evaluating how well the model distinguishes actual fires from the confounding elements that typically result in false positives within the negative class. The proposed approach outperformed the baseline methods—including single-task learning and direct multi-task learning—achieving a primary accuracy of 93.36%, an F1-score of 91.57%, and a higher MES score, demonstrating its enhanced robustness and reliability in diverse environmental conditions. This work bridges the gap between advanced deep learning techniques and practical, scalable solutions for environmental monitoring. Future research will focus on integrating multi-modal data and developing adaptive distillation techniques to further enhance the model’s performance in real-time applications.

Transforming fire protection: AI's evolving role in proactive and reactive applications

Abstract. This paper comprehensively reviews the applications of artificial intelligence (AI) in the field of fire protection. It traces the historical development of AI, from its origins in the 1930s and 1940s to its current state, focusing on major milestones and algorithms that have shaped the field. The paper discusses the evolution of AI, particularly in terms of machine learning, deep learning, and the emergence of various models such as deep neural networks (DNNs), transformer models, and graph neural networks (GNNs). The application of AI in fire protection is explored in two primary categories: proactive and reactive applications. Proactive applications encompass AI's role in decision-making, construction, and operation, including the use of machine learning for environmental analysis, building information modelling (BIM) for design optimization, 3D printing and intelligent robots in construction, and intelligent fire facility management. Reactive applications focus on early fire detection and monitoring, especially in forests, utilizing terrestrial, aerial, and satellite systems equipped with advanced sensors and deep learning techniques. This paper highlights the significant contributions of AI in mitigating fire risks, improving fire detection and response, and enhancing the safety and efficiency of fire protection systems.

Deep Learning Approaches for Wildfire Severity Prediction: A Comparative Study of Image Segmentation Networks and Visual Transformers on the EO4WildFires Dataset

This paper investigates the applicability of deep learning models for predicting the severity of forest wildfires, utilizing an innovative benchmark dataset called EO4WildFires. EO4WildFires integrates multispectral imagery from Sentinel-2, SAR data from Sentinel-1, and meteorological data from NASA Power annotated with EFFIS data for forest fire detection and size estimation. These data cover 45 countries with a total of 31,730 wildfire events from 2018 to 2022. All of these various sources of data are archived into data cubes, with the intention of assessing wildfire severity by considering both current and historical forest conditions, utilizing a broad range of data including temperature, precipitation, and soil moisture. The experimental setup has been arranged to test the effectiveness of different deep learning architectures in predicting the size and shape of wildfire-burned areas. This study incorporates both image segmentation networks and visual transformers, employing a consistent experimental design across various models to ensure the comparability of the results. Adjustments were made to the training data, such as the exclusion of empty labels and very small events, to refine the focus on more significant wildfire events and potentially improve prediction accuracy. The models’ performance was evaluated using metrics like F1 score, IoU score, and Average Percentage Difference (aPD). These metrics offer a multi-faceted view of model performance, assessing aspects such as precision, sensitivity, and the accuracy of the burned area estimation. Through extensive testing the final model utilizing LinkNet and ResNet-34 as backbones, we obtained the following metric results on the test set: 0.86 F1 score, 0.75 IoU, and 70% aPD. These results were obtained when all of the available samples were used. When the empty labels were absent during the training and testing, the model increased its performance significantly: 0.87 F1 score, 0.77 IoU, and 44.8% aPD. This indicates that the number of samples, as well as their respectively size (area), tend to have an impact on the model’s robustness. This restriction is well known in the remote sensing domain, as accessible, accurately labeled data may be limited. Visual transformers like TeleViT showed potential but underperformed compared to segmentation networks in terms of F1 and IoU scores.