A Multimodal ML-based Indoor Fire Detection and Alarm System Using Image Processing and IoT–based Environmental Sensing

The fire detection system combines the simultaneous measurements of smoke, carbon monoxide, and carbon dioxide. The security of campus against intruders moving in laboratories, class rooms, staffrooms or washrooms. The fire alarm system consists of Fire detectors (with can be smoke detector, heat, or Infra-Red detectors), control unit and alarm system. A fire detection system is developed based on the simultaneous measurements of temperature and smoke. The fire detection system with the alarm algorithm detected fires that were not alarmed by smoke sensors, and alarmed in shorter times than smoke sensors operating alone. Previous fire detection algorithms used data from sensors for temperature, smoke, and combustion products. The smoke sensor alarms when the analog output signal exceeds or equal the threshold value. The node includes analog sensors to measure smoke, carbon monoxide (CO) and temperature. A fire alarm system should reliably and in a timely way notify building occupants about the presence of fire indicators, such as smoke or high temperatures.

Fire can really be devastating to properties if improperly managed, it is due to this problem that the fire detection and alarm systems were sought for. However, traditional fire alarm system is based on a wiring network which have drawbacks and limitations such as inflexibility of the FDAS layout plan during building construction, and difficulties in renovation where the removal and relocation of traditional FDAS requires additional amount of work. To address these problems, a fire detection and alarm system that is based on wireless sensor network was developed. The FDAS is mainly composed of a fire detection node, a fire alarm node, and a fire alarm control panel. The wireless communication of the nodes was achieved using XBee as the wireless transceiver. The significant characteristic of XBee is the mesh routing protocol which helps the system to be more robust and flexible. The fire alarm control panel is a LabVIEW based program that utilizes the state machine architecture which follows the detect-evaluate-store algorithm. In evaluating the overall system, results illustrate the network of nodes adapts to its location and connects to other nodes whenever it is within their range. Furthermore, the average response time of each detection node with respect to the fire alarm control panel is below the standard of 10-second as stated by the National Fire Alarm and Signaling Code.

The requirement of the government thru the Bureau of Fire Protection to install Fire Detection and Alarm Systems (FDAS) to high occupancy buildings or structures has raised the need of skilled workers to perform this task. To address this matter, Technical Skills Development Authority (TESDA) included in the Electrical Installation and Maintenance (EIM) NCII qualification the competency of FDAS installation, set-ups, and troubleshooting. With the inclusion of the FDAS in the said qualification a decline of passers was observed from July 2017 to December 2017. In response, the researcher developed a portable device to be utilized in facilitating a retentive means of teaching the required knowledge, skills and attitude of students in the wiring, installation and troubleshooting of FDAS and to encourage training institutions to incorporate this training device in their training programs and institution. This study evaluated the training device to determine its acceptability in premises of design, performance, and safety. The evaluation purposively involved seven senior high school faculty who were teaching electrical installation subjects in the Northeastern part of Leyte Province. A mean of 4.937 was obtained from the evaluation of the portable device, and affirmed that the FDAS Trainer is very acceptable to the evaluators. Having a very acceptable rating in the three criterions, it can be inferred that this portable trainer device can be used as an effective training device that can develop knowledge, skills, and attitude of students and trainees in electrical installation classes, auxiliary systems installation classes, and other allied electrical installation programs. Keywords: FDAS, Fire alarm, Fire detection and alarm systems, Trainer, Technology development

In recent years the complexity and cost of Fire and Gas detection systems have increased rapidly with no measurable reduction in fatality rates or other losses. A principal reason for this disappointing outcome despite substantial investment is that systems designs have lost the focus of what they are trying to achieve. Another reason worth exploring is that investment in other areas may result in greater benefits. For instance, diverting funds to preventing lifting accidents, helicopter failures and materials handling incidents, could result in greater OVERALL benefit since these account for a larger proportion of risks. Nevertheless, Fire and Gas detections systems have a major role to play in preventing escalation, and this paper shows how they can be made more efficient in both cost and performance by combining four methods: targeting perceived hazards,integration from concept stage with other safety designs,control of design with effective performance specification, andcontrol of modifications throughout the system's life.

With the widespread application of automatic fire alarm systems in various types of buildings, the problem of fire false alarms has gradually become prominent, which not only causes resource waste, but also may reduce users’ trust in the alarm system, thereby affecting the efficiency of emergency response in actual fires. According to data from a certain fire cloud platform, 99.85% of the suspected fires predicted by its system are false alarms. Although existing models can recognize most fire accidents, the accuracy of fire alarm recognition is only 0.15%, due to loose judgment logic, which still requires a large amount of manpower to verify alarms. This article analyzes a large amount of false alarm data and explores the main causes of false alarms, including environmental interference, equipment failure, and improper human operation. By using a fire dynamics simulator (FDS) to establish fire simulation models under different data settings, horizontal and vertical multi-scene fire simulation data are obtained. The study combines simulation and platform data to form a fire and false alarm dataset using a one-dimensional convolutional neural network (1D-CNN) and deep neural network (DNN) deep learning techniques to learn the deductive rules of the fire scene, establish a two-stage judgment model, and gradually, accurately, judge the results. By quantifying the precision, recall, and F1 score of the model, a deep learning model designed to accurately identify genuine fire alarms while filtering out false ones is proposed that can significantly reduce the false alarm rate. The results indicate that the model can identify 1705 false alarms out of 2255 highly suspected true alarms identified by existing systems in multiple practical scenarios and eliminate 75.61% of false positive alarms. On the premise of ensuring an authenticity recognition rate greater than 98%, the accuracy of fire alarm recognition increased from 0.15% to 28.85%, which will significantly reduce the workload of staff verifying alerts, and has good practical value.

The paper starts with the introduction of the basic requirements of classical intrusion detection and alarm systems and fire detection systems. The drawbacks of such conventional systems are highlighted. Techniques of computer vision are employed to remove the drawbacks and at the same time increase the reliability and response rate of the systems. For security and low level fire detection, a fuzzy logic based image comparison algorithm is deemed adequate. In order to confirm the existence of fire or smoke, techniques related to optical flow are employed as high level fire or smoke detection, which generate a velocity field for the image so that the decision can be judged by using fizzy logic. Details of implementation and some experimental results have been included in the paper for illustration.

Fire detection and alarm systems are recognized as key features of a building’s fire prevention and protection strategy. This chapter presents a systematic technique to be used by fire protection engineers in the design and analysis of detection and alarm systems. The majority of discussion is directed toward systems used in buildings. However, many of the techniques and procedures also apply to systems used to protect planes, ships, outside storage yards, and other nonbuilding environments.

An accurate and fast fire smoke detection algorithm is urgently needed to solve the emergency linkage measures to prevent early fire spread and after the fire. In this paper, the depth matrix of motion difference is constructed for background modeling, the moving target is extracted using flame smoke motion characteristics, and the suspected area of flame smoke is obtained through corrosion and expansion. The unique characteristics of flame and smoke are then extracted, and the smoke flame recognition model is built using a BP neural network optimized by a genetic algorithm. The experimental results show that the algorithm can reliably eliminate the interference of moving vehicles and lights in the tunnel, detect flames and smoke, and generate an alarm, and that it can be applied to fire smoke detection in real-world scenes. Finally, the emergency linkage measures in the corresponding area are automatically initiated based on the position of the fire point via information interaction with the fire smoke detection and alarm system.

This paper presents some fundamental factors that influence the reliability of fire and gas (F&G) alarm systems by incorporating human factor's principles. As compared, standard alarm and the safety‐related alarm will show two different values of IPF given in process hazard analysis (PHA) studies. In this study, fault tree diagram has been used to calculate the process flow diagram (PFD) value, to identify which alarm type could meet the standard of safety integrity level (SIL1) for F&G system for a single train process. Comparison of both alarms is done for three types of selected detectors; flammable gas detectors, flame detectors, and toxic gas detectors. It has been identified that standard alarm for all type of detectors does not meet the SIL1 requirement. Compared to safety‐related alarm, all detectors are able to meet the requirement of SIL1. Since the greater IPF will be given for safety‐related alarm, it is crucial to study the holistic factor that could maintain the system with good efficiency and performance. The human factor has been identified as the most critical element for safety‐related alarm. A complete coordinated approach is needed to ensure the performance and efficiency of the F&G alarm system can be maintained while process plants are able to be safely operated within as low as reasonably achievable region.

The functioning fire protection system in the State Forests, which covers 76.9% of the forest area in the country, is constantly improved thanks to close cooperation with the Forest Research Institute. The results of the research conducted by the Institute are thus implemented in forestry practice, examples of which are presented in this article. The basis for planning and organizing rescue operations is the assessment of potential forest fire risk, which is carried out using the method of classifying forest areas to the forest fire risk category. It is determined for the forest district (average area 17500 ha) for 10 years on the basis of the frequency of fires, forest stand, climatic and anthropogenic conditions according to the formulas developed for each of the factors. The category determines the distribution of funds for fire protection and the method of forest preparation in the event of a fire, concerning fire detection, communication and alarm systems, the density of fire access roads and water supply points, and fire extinguishing equipment. The macroscopic method of forest fire risk categorization is complemented by the method of forest stand flammability classes, which determines their susceptibility to fire due to the presence of flammable material. It is based on the forest habitat type and soil cover type and is determined at the level of separation (average area 3 ha), with the possibility of generalization to the division (20 ha) and forestry (1300 ha). Both methods are used to map the potential fire risk of forests in Poland. The method of determining the degree of forest fire risk on the basis of air temperature and relative humidity, precipitation and direct measurements of moisture content of pine litter (Pinus sylvestris) is used to assess the dynamic forest fire risk, shaped by weather conditions. Determining the degree of risk is performed from the 1st of March to the 30th of September, daily at 9.00 and 13.00 in 60 forecast zones with the use of an automated network of 145 meteorological measuring points. Information on the degree of risk determines the operational readiness of ground and air forces. In the event of a fire and determination of its forest address, we can download from the Information System of the State Forests, data on the flammability class of the forest stand and the characteristics of the flammable material (type, fuel load), using the developed fuel models for 7 types of soil cover, the fires of which constitute nearly 90 % of all events. The possibility of a soil cover fire turning into a crown fire is determined based on the developed algorithms, taking into account the age of the forest stand, habitat and stand conditions, meteorological conditions and the duration of the fire. Organizing a firefighting action, including in particular having the right amount of forces and resources to extinguish a fire with the smallest possible area, can be supported by the application Model of a forest fire, which the forest service has at its disposal.

Sensors for fire alarms require a high level of predictive variables to ensure accurate detection, injury prevention, and loss prevention. Bayesian networks can aid in enhancing early fire detection capabilities and reducing the frequency of erroneous fire alerts, thereby enhancing the effectiveness of numerous safety monitoring systems. This research explores the development of optimized probabilistic graphic models for the discretization thresholds of alarm system predictor variables. The study presents a statistical model framework that increases the efficacy of fire detection by predicting the discretization thresholds of alarm system predictor variable fluctuations used to detect the onset of fire. The work applies the Bayesian networks and probabilistic visual models to reveal the specific characteristics required to cope with fire detection strategies and patterns. The adopted methodology utilizes a combination of prior knowledge and statistical data to draw conclusions from observations. Utilizing domain knowledge to compute conditional dependencies between network variables enabled predictions to be made through the application of specialized analytical and simulation techniques.

The combination of ultrasonic and microwave sensors offers versatile sensor applications and new solutions in terms of reliability of decision and robustness against distortions. For motion detection in an intruder alarm system, the redundant sensor information from the Doppler multisensor is used to distinguish a moving person from distortions. The superior performance is based on the fact that a moving person causes coherent frequencies of the ultrasound and the microwave Doppler signal, whereas noise and distortions influence both waves differently. In this paper, a new complementary approach is described to recognise a fire as the source of a corrupted coherence between the ultrasonic and the microwave Doppler signal. As a result, a comprehensive protection system for both motion tracking and fire detection is presented which is very sensitive, fast in response, and robust against false alarms. Results from measurements with 40 kHz ultrasound and 24 GHz microwaves show the potential of the new approach.

Fire hazards pose a significant risk to human lives, infrastructure, and the environment, necessitating the development of efficient fire detection and localization systems. Traditional methods, including smoke and heat sensors, suffer from high false alarm rates and delayed response times. In recent years, deep learning has emerged as a transformative approach, leveraging computer vision to enhance accuracy in fire detection and localization. This paper provides a comprehensive survey of deep learning techniques employed for fire detection, including Convolutional Neural Networks (CNNs), object detection models such as YOLO and Faster R-CNN, and hybrid approaches integrating multimodal data. The study further explores publicly available fire datasets, preprocessing techniques, and performance evaluation metrics. Additionally, key challenges such as real-time processing constraints, environmental variability, and model generalization are discussed. The paper concludes with an outlook on future advancements, including lightweight AI models, multi-sensor fusion, and synthetic dataset generation for robust fire detection systems.

In this paper, the optimal threshold for fire detection is derived based on the probability density functions of the fire and background pixels. The threshold values and the commission and omission errors are computed according to several detection criteria: maximum likelihood, null-hypothesis testing, and minimum cost. This statistical approach enables the determination of optimal threshold value according to the specific requirements of the detection tasks I. INTRODUCTION Active fire detection using satellite thermal sensors usually involves thresholding the detected brightness temperature in several bands. The frequently used features for fire detection are the brightness temperature in the 4-µm wavelength band (T4) and the brightness temperature difference between the 4- µm and 11- µm bands (∆T=T4-T11) (1), (2), (3). Due to the statistical nature of the fire and background temperature distribution, no matter how the thresholds are chosen, there are bound to be commission errors (i.e. false alarms) and omission errors (false negatives) in the detection outcomes. In the commonly used fire detection algorithms, the thresholds are often arbitrarily defined, and hence may not be optimal for a given fire detection task. The optimal threshold values are expected to be site dependent. They are different in different regions and for different fire conditions (such as smoldering and flaming fires). Hence, a global set of threshold values may not work optimally for a specific site. The task of active fire hot spot detection can be modeled as a stochastic target detection procedure (4). In this model, each pixel of the image occupies a point (T4, ∆T) in the two- dimensional feature space, and each pixel belongs to either the fire class or the non-fire background class. This approach considers the probability density functions of the fire and background pixels and optimal thresholds are derived depending on the specific objectives of the detection tasks. In a previous work (4), we implemented a method for deriving the optimal detection threshold by minimizing a cost function which is a weighted sum of the omission and commission errors. Alternatively, the threshold can also be derived based on other criteria. In this paper, we derive the optimal thresholds for detecting active fires in MODIS scenes using various detection criteria: maximum likelihood, constant false alarm rate, and null-hypothesis testing. The test data set consists of several MODIS scenes over Sumatra and Borneo islands and coincidental high resolution SPOT scenes (5). The SPOT data serves as ground truth for validation of MODIS hotspots detected by the new algorithms based on stochastic models. The probability density functions (pdf's) of the fire and background pixels are derived from this dataset and these pdf's are used to derive the optimal thresholds for active fire detection.

Forest fires causes a severe risk to human being and the environment. By 2030, woodland fires are expected to have destroyed half of the world’s forests. Assuming early fire recognition approaches is the most actual strategy to limit the damage caused by forest fires. Various industrial fireplace detection sensor structures exists on the market today, but they are all problematic to use in wide open areas like as woods due to their slow response time. Image processing has been used in this study for a variety of reasons, including the rapid advancement of virtual cameras era. Accurate forest fire detection algorithms remain a challenge since certain items have similar characteristics to fire, potentially resulting in a high false alarm rate. To reduce the error rate at the fire detection, initially detect the background from foreground image regions. Second, the RGB shade space is used to select prospective fireplace zones. Finally, features extraction is employed to distinguish between real picture and fireplace-like items, as candidate areas can also include moving patch-like things. Finally, the region of interest is classified as either true fire or non-fire using machine learning and deep learning techniques.KeywordsDeep learningForest fire recognitionForeground separationImage miningMachine learning