基于FireNet的古建筑火灾检测方法研究及改进

针对古建筑火灾检测需要快速、准确及实时的需求,建立了一个专门用于古建筑火灾检测的数据集,用于古建筑火灾检测的深度学习研究。利用CBAM注意力机制模块,结合多尺度特征融合,对FireNet网络进行改进,提出适用于古建筑火灾检测的轻量级FireNet-AMF网络,在FireNet数据集和本文构建的古建筑火灾检测数据集上验证了FireNet-AMF网络的火灾检测能力。与改进前的网络相比,FireNet-AMF网络在FireNet数据集上对火灾识别的准确率达到了95.08%,与原网络相比提高了1.17%,在本文构建的古建筑火灾检测数据集上的准确率达到了95.62%,比原网络提高了1.62%。该网络在保证轻量级的同时也保证了在古建筑火灾检测中较高的检测精度。

Performance Analysis of Support Vector Machine (SVM) on Challenging Datasets for Forest Fire Detection

This article delves into the analysis of performance and utilization of Support Vector Machines (SVMs) for the critical task of forest fire detection using image datasets. With the increasing threat of forest fires to ecosystems and human settlements, the need for rapid and accurate detection systems is of utmost importance. SVMs, renowned for their strong classification capabilities, exhibit proficiency in recognizing patterns associated with fire within images. By training on labeled data, SVMs acquire the ability to identify distinctive attributes associated with fire, such as flames, smoke, or alterations in the visual characteristics of the forest area. The document thoroughly examines the use of SVMs, covering crucial elements like data preprocessing, feature extraction, and model training. It rigorously evaluates parameters such as accuracy, efficiency, and practical applicability. The knowledge gained from this study aids in the development of efficient forest fire detection systems, enabling prompt responses and improving disaster management. Moreover, the correlation between SVM accuracy and the difficulties presented by high-dimensional datasets is carefully investigated, demonstrated through a revealing case study. The relationship between accuracy scores and the different resolutions used for resizing the training datasets has also been discussed in this article. These comprehensive studies result in a definitive overview of the difficulties faced and the potential sectors requiring further improvement and focus.

Lightning in a Forest (Wild) Fire: Mechanism at the Molecular Level

The mechanism of lightning that ignites a forest fire and the lightning that occurs above a forest fire are explained at the molecular level. It is based on two phenomena, namely, internal charge separation inside the atmospheric cloud particles and the existence of a layer of positively charged hydrogen atoms sticking out of the surface of the liquid layer of water on the surface of rimers. Strong turbulence-driven collisions of the ice particles and water droplets with the rimers give rise to breakups of the ice particles and water droplets into positively and negatively charged fragments leading to charge separation. Hot weather in a forest contributes to the updraft of hot and humid air, which follows the same physical/chemical processes of normal lightning proposed and explained recently[1]. Lightning would have a high probability of lighting up and burning the dry biological materials in the ground of the forest, leading to a forest (wild) fire. The burning of trees and other plants would release a lot of heat and moisture together with a lot of smoke particles (aerosols) becoming a strong updraft. The condition for creating lightning is again satisfied which would result in further lightning high above the forest wild fire.

基于BIMFire-YOLO的消防监控图像识别方法

在传统条件下,通常存在着火灾现场情况多变,建筑物内构造复杂等问题,使得消防监控系统难以通过现有方法精准检测识别火灾第一现场和所处室内环境。为了优化检测算法在复杂场景下的应用效果,文中提出BIMFire-YOLO检测方法,在以往YOLOv3主干网络中增加了新的卷积方式,在扩大了卷积感受野范围的同时,也保证了输出特征的不变性,对小火焰和烟雾等目标的检测能力有所提升。通过引入新的损失函数,降低了火灾发生的目标漏检率,融合BIM技术不仅提升了定位精度也实现了自动识别视频和图像中失火的场景。该算法不仅提高了检测效率,也实现了火灾的信息共享与管理的统一,在实际医院场景下的综合识别精度达到了45.6%,检测速度和精度均优于其他同类方法。

基于改进FireNet的轻型火灾实时检测方法

针对火灾检测精确不高,时间长等问题。设计了基于改进FireNet的轻型火灾实时检测方法,通过获取视频图像数据,网络模型进行火灾分析和识别;首先,在FireNet特征提取阶段使用多尺度卷积网络并引入通道注意力机制,以提高回归精度。其次,对全连接层的神经元个数进行压缩优化,减少计算耗时。实验表明,改进的FireNet算法模型检测精度达到96.43%,模型存储空间0.96MB,检测帧率37。相比标准算法精度提高2.5%,存储空间压缩85%,检测帧率提升35%。

The impact of climate change on fire risk in Daxing'anling,China

Daxing’anling is a key region for forest fire prevention in China. Assessing changes in fire risk in the future under multiple climatic scenarios will contribute to our understanding of the influences of climate change for the region and provide a reference for applying adaptive measures for fire management. This study analyzed the changes in fire weather indices and the fire season under four climate scenarios (RCP2.6, RCP4.5, RCP6.0, RCP8.5) for 2021-2050 using data from five global climate models together with observation data. The results showed that the analog data could project the average state of the climate for a given period but were not effective for simulating extreme weather conditions. Compared with the baseline period (1971-2000), the period 2021-2050 was predicted to have an increase in average temperature of 2.02-2.65 °C and in annual precipitation 25.4-40.3 mm, while the fire weather index (FWI) was predicted to increase by 6.2-11.2% and seasonal severity rating (SSR) by 5.5-17.2%. The DMC (Duff moisture code), ISI (initial spread index), BUI (build-up index), FWI and SSR were predicted to increase significantly under scenarios RCP4.5, RCP6.0, and RCP8.5. Furthermore, days with high or higher fire danger rating were predicted to be prolonged by 3-6 days, with the change in the southern region being greater under scenarios RCP4.5, RCP6.0, and RCP8.5.

Experimental Studies of Interactions Between Backfires and Coming Surface Fires\+*

In order to study the convection limits of surface fires and interactions between backfires and main fires,several experiments are conducted in a large space indoor laboratory: in which the effects of ambient wind speeds and changing temperatures can be avoided.The research shows that:(1) there is a convection field in front of coming fires in which the wind speed direction is toward the fire.In the convection area,the lower part has higher wind speed and when the height is taller than a certain value the convection wind speed is not significant;(2) the backfire and the main fire interact with each other even though they are far apart.When they come near each other to a certain distance,they begin to draw each other.This increases their rates of spread toward each other significantly.For surface fires with a fire line intensity of 160?kW\5m -1 ,their rate of spread increases by 27%.

Trends in Global and Mexico Research in Wildfires: A Bibliometric Perspective

In the last two decades, unprecedented changes have taken place in the frequency and severity of wildfires;in different regions of the world, some fires were even classified as megafires. Although there are studies about the diverse effects of fire, which have made significant theoretical contributions, a comprehensive review of the changes in fire research is required to understand worldwide patterns, particularly in those countries where fire activity is on the rise, such is the case of Mexico. The objective of this study was to analyze the trends in the research on wildfires published in Mexico and worldwide over a 40-year timescale. For this purpose, the Web of Science database, bibliometric tools, and the keywords TI = Forest fire* OR TI = Wildfire* were used to extract as many articles as possible related to fires from 1980 to 2020, without being restricted to those studies whose title included any of the variants of the keywords. There were 8458 publications about fires in the vegetation cover, with a notable increase in the frequency of studies in the previous decade;52% of the studies were concentrated in five countries and 20% of the articles focused on the study of different aspects of the soil. Mexico ranks thirteenth in volume of scientific production and studies in the country have focused mainly on the description of the quantitative relationship between the size of the affected area and the number of occurrences in the landscape, meanwhile, studies on fires and the consequences on the biotic interactions have been little explored.

Gas Temperature Distribution and Fluctuation in a Lab-Scale Fire Whirl

Fire whirls cause an increase in fire damage. This study clarified the unsteady behavior of fire whirls, considering that instantaneous changes in the temperature and flame shape of fire whirls can affect the damage to the surrounding area. Numerical simulations of a lab-scale flame that simulates a fire whirl were performed to investigate the changes in gas temperature and velocity fields under various fuel inflow velocities. The flow field was obtained by solving a continuity equation and a three-dimensional Navier-Stokes equation, and the turbulence was resolved using a large eddy simulation. A chemical equilibrium partially premixed combustion model was used, and radiation effects were considered. The time-averaged gas temperature distribution along the burner central axis revealed that the gas temperature decreased monotonically from upstream to downstream. The time-averaged velocity distribution along the burner central axis showed that the velocity decreased as one moved downstream, but the decrease was uneven. The time variation of the gas temperature demonstrated that the higher the fuel inflow velocity, especially near the burner, the greater the gas temperature flutter. Furthermore, the larger the fuel inflow velocity, the larger the flame swell and wobble. The results showed that the fuel inflow velocity affected temperature fluctuation and flame undulating movement.

A Survey of the Machine Learning Models for Forest Fire Prediction and Detection

Forest fires are a significant threat to the environment, causing ecological damage, economic losses, and posing a threat to human life. Hence, timely detection and prevention of forest fires are critical to minimizing their impact. In this paper, we review the current state-of-the-art methods in forest fire detection and prevention using predictions based on weather conditions and predictions based on forest fire history. In particular, we discuss different Machine Learning (ML) models that have been used for forest fire detection. Further, we present the challenges faced when implementing the ML-based forest fire detection and prevention systems, such as data availability, model prediction errors and processing speed. Finally, we discuss how recent advances in Deep Learning (DL) can be utilized to improve the performance of current fire detection systems.

Selection of Fire-resistant Tree Species for Southwestern China

Selection of fire resistant tree species for the southwestern China and the planting of those species can effectively prevent large area's fire damage. In this paper the components and flammability of leaves, twigs and barks of 12 tree species in the mountain area of southwestern China have been tested and analyzed in the laboratory. The test and analysis indicate the results as follows:(1) for all the tree species, the fire resistance of leaves is much weaker than that of twigs and barks, and the broad leaves are stronger than those of conifers in fire resistance. (2) Heat value, moisture, ignition point and ash content are main indexes to affect fire resistance. Heat value relates to lignose content and benzene ethanol extractive content linearly.(3) Of all the 12 tree species, Schima superba,Castanopsis hystrix, Myrica rubra have the strongest resistance to fire; Machilus pauhoi, Michelia macclurei, Mytilaria laosensis, Camellia olifera and Manglietia tenuipes are relatively strong in fire resistance, and Lithocapus thalassica, Tsoongiodendron odorum, Cunninghamia lanceolata and Pinus massoniana are weak in fire resistance.

Rapid Detection of Accelerants in Fire Debris Using a Field Portable Mid-Infrared Quantum Cascade Laser Based Analyzer

Arson presents a challenging crime scene for fire investigators worldwide. Key to the investigation of suspected arson cases is the analysis of fire debris for the presence of accelerants or ignitable liquids. This study has investigated the application and method development of vapor phase mid-Infrared (mid-IR) spectroscopy using a field portable quantum cascade laser (QCL) based system for the detection and identification of accelerant residues such as gasoline, diesel, and ethanol in fire debris. A searchable spectral library of various ignitable fluids and fuel components measured in the vapor phase was constructed that allowed for real-time identification of accelerants present in samples using software developed in-house. Measurement of vapors collected from paper material that had been doused with an accelerant followed by controlled burning and then extinguished with water showed that positive identification could be achieved for gasoline, diesel, and ethanol. This vapor phase mid-IR QCL method is rapid, easy to use, and has the sensitivity and discrimination capability that make it well suited for non-destructive crime scene sample analysis. Sampling and measurement can be performed in minutes with this 7.5 kg instrument. This vibrational spectroscopic method required no time-consuming sample pretreatment or complicated solvent extraction procedure. The results of this initial feasibility study demonstrate that this portable fire debris analyzer would greatly benefit arson investigators performing analysis on-site.

Investigation on Fire Protection System of Highrise Buildings in the Context of Bir Uttam Aminul Haque Avenue in Dhaka, Bangladesh

Rapid high-rise building construction of different occupancies is becoming popular in Dhaka due to its economic gain and functional flexibility. Fire prevention poses a significant difficulty to this type of construction due to its complexity and economic worth as well. Therefore, construction of high-rise building without following the proper fire safety measures, is a common practice at present in the city and it poses a greater threat to urban life considering its associated loss. Even though, most of the owners or authorities do not respect the construction code and the majority of them lack sufficient awareness and basic fire control knowledge and practice. More often, they are unable to comprehend the magnitude and severity of severe fire hazards, and recognize the causes and implementing effective mitigation measures are rare. As a result, the number of fire hazard in high-rise construction is increasing day by day. Hence, investigating the present condition of high-rise buildings (already built) in terms of Fire Protection System is imperative to prevent the upcoming fire hazard. An urban chunk of 33 plots along with Bir Uttam Aminul Haque Avenue, located in Banani, has been chosen for the study area considering its vulnerability to fire hazard. Among these buildings, STAR Tower and HBR Tower, these two high-rise buildings are assessed thoroughly using FRI (Fire Risk Index) Method. The method is semi-quantitative in nature with seventeen parameters associated with fire protection. This assessment represents a scenario in what extent fire codes are being followed in the selected area. It concludes that most of the buildings lack in providing protection against fire hazards for not following the code strictly. This research provides some recommendations which can be followed to improve the fire safety measures in this existing context. It is expected that this research could be a unique addition to firefighting knowledge by contributing to mitigating the consequences of fire related hazards in any densely populated city like Dhaka.

Investigation on the Fire Resistance of Cellular Steel Beam with Sinusoidal Openings

This paper investigated the fire resistance of CSBs with various parameters under high temperature rise due to fire using finite element software ABAQUS. The mechanical parameters of CSBs are analyzed, including load-bearing capacity and the temperature distribution during the heating process. Through structural analysis simulation of the entire heating process, the structural response of the CSBs is divided into five stages: elastic stage, elastic-plastic stage, self-balancing stage, catenary stage and ultimate destruction stage. The results indicate that the opening diameter-to-height ratio, opening spacing-to-height ratio and load ratio significantly affect the structural responses of CSBs in fire, followed by opening shape as secondary effects. In all the numerical analyzes, CSBs are analyzed with a uniformly distributed load and having simply supported boundary conditions.

Usage of Cool Burning as a Contributor to Bushfire Mitigation

This paper considers and describes the cool burning techniques long utilized by Australian Indigenous people as a contributor to bushfire mitigation. Indigenous fire management involves lighting the “cool” fires in selected areas between March and July, in Australia, during the early dry season. The fires burn gradually, reducing fuel loads and creating fire breaks and not all of the area is burnt. Late in the dry season, when the weather is very hot, the method removes fuel for larger fires while maintaining and protecting habitat for mammals, reptiles, insects and birds. The management of Indigenous cultural fire offers an Indigenous viewpoint for wider control of fire and cultural fire management is an opportunity for collaborations to encourage Aboriginal empowerment with public and private sector organisations. Effective cool burning in contemporary prescribed burning activities can be achieved through implementation of good training, strong partnerships, carefully considered on ground practices and appropriate and effective techniques.

Determination of fire-access zones along road networks in firesensitive forests

Wild forest fires are one of the greatest environmental disasters affecting forest resources. Along the coastal zone of the Mediterranean region in Turkey, forested areas are classified as first-degree, fire-sensitive areas. Every year, thousands of hectares of forests have been destroyed in Turkey. In this study, fire-access zones were determined in the Mediterranean forests of Turkey, by utilizing geographic information systems （GIS） technology. The effective reach distance of fire hoses from both sides of roads was considered in order to delineate fireaccess zones. The effective reach distance can vary based on the technical capabilities and hydraulic capacity of fire trucks （minimum and maximum pressures on water pump）; terrain structures （uphill, downhill and flat）; and ground slope. These factors and their influences were studied in fire sensitive forest areas located in the eastern Mediterranean city of Kahramanmaras in Turkey. First, terrain structures on both sides of the road network and groundslope classes were determined based on GIS data layers. Then, fire access zones were delineated according to water pressure data, terrain structures, and ground-slope classes. The results indicated that 69.30 % of the forested areas were within the fire-access zones, while the rest of the forest was out of reach the fire hoses. The accessible areas were also calculated for forested areas with different fire- sensitivity degrees. The accessible areas were 69.59, 69.96, and 67.16 % for the forested areas that are sensitive to fires at the first, second, and third degrees, respectively. This finding has implications for the monitoring and management of fire threats in areas outside of the reach distance. The outside areas should receive extra attention and monitoring during the fire season so that fires are detected ahead of time and management has sufficient time to react. Besides, new roads should be considered for these areas in order to access more lands in a shorter amount of time.

Numerical Studies of Wind Speed Fields in Front of Forest Fire Fronts

In the study a fire and fire environment model is set up and by using PHEONICS software 3 cases of surface fires are studied. The results fit the experimental studies well generally. The simulation reveals that (1) The wind speed fields in front of fire front generally can be divided into 3 zones and there is always an eddy immediately at the corner between just in front of the fire and the ground. (2) The shape and dimension of the division of the 3 zones is mainly decided by slope angle and ambient wind speed given fire line intensity. (3) There exits an upwind zone in front of fire front. Ambient wind speeds have little effect on the magnitude of the upwind speed when slope angle is 0. But when the slope angle is negative, the upwind is apparently stronger.

THE ROLE OF FIRE IN FOREST DYNAMICS IN DAXINGANLING REGION

Recent fire statistics and preliminary fire history data suggest that fire has been historically responsible for maintaining the vegetative communities up to present in Daxinganling region. Forest types, and even tree species, arc dependent on the degree of fire intensity, fire size, depth of burn and fire frequency. Selected samples of larch, pine, birch and spruce forest were studied in terms of species composition as determined by fire frequency which mainly depends on topography and site conditions. Intervals between fires range between 6 and 170 years.

METHODS OF EVALUATING FOREST FIRE HISTORY

Forest fire history can be reconstructed over past centuries across a wide variety of forest types. Fire scars on living tress, and age classes of forest stands, are the two sources of information for these reconstructions. Point and area frequencies are used to reconstruct fire history. Point frequencies are useful in forest types that burn with frequent, low intensity fire so that many fire-scarred residual trees exist. A true point is a single tree, but more often point estimates are made by combining fire scar records from several adjacent trees. Area frequences are applied where fires are infrequent but of moderate to high intensity, so that stand ages are used across wide areas to estimate fire return interals. Proper selection and application of fire history methods are essential to deriving useful ecological implications from fire history studies.This review evaluates the common methods of determining fire history :what the techniques are, where they are best applied, and how to interpret them in an ecological context. Emphasis is placed on fire freqency and predictability, but choosing a proper technique may also be a function of fire intensity.

Impacts of fire severity and cattle grazing on early plant dynamics in old-growth Araucaria-Nothofagus forests

Background: Fire has historically shaped the dynamics of Araucaria araucana-Nothofagus pumilio forests in the Andean region of South America. Nevertheless, human disturbances and stronger drought events have resulted in increased occurrence and severity of wildfires. Regardless of their conservation relevance, the initial response to fire of Araucaria-Nothofagus forests has not been well documented. Through this research we tested the hypothesis that plant composition of Araucaria-Nothofagus forest after fire will initially differ in its recovery depending on fire severity, but over time, plant similarity will increase among areas with different levels of fire severity. The study was carried out in old-growth Araucaria-Nothofagus forests in south-central Chile(38°S) that were burned in 2015. We studied vegetation response to three levels of fire severity(low(LS), moderate(MS) and high(HS) severity), also including an adjacent unburned forest(UN), which were assessed one, two and three years after fire. We also evaluated the impact of cattle on plant recovery after fire. We measured species richness, abundance, plant diversity and origin of all vascular plant species.Results: Species richness and plant abundance responded differently to fire severity. Time-since-fire had a significant effect on plant richness and abundance. Plant composition within LS areas was similar to UN areas, but MS and HS areas were significantly less diverse compared to UN and LS. In absence of other major disturbances, similarity in plant composition tended to increase over time across the different levels of fire severity, becoming more similar to UN. We also found a synergetic effect of cattle and fire on species richness. This interaction promotes the establishment of nonnative species and limits the recovery of native species following fire.Conclusions: Fire severity had a significant impact on the plant community of old-growth Araucaria-Nothofagus forests,especially in areas of high fire severity, where the native tree N. pumilio was completely absent. Nevertheless,plant communities can recover gradually, depending on fire severity, time-since-fire, and the absence of cattle grazing. Management actions should include the exclusion of cattle within fire-affected areas and afforestation with N. pumilio in areas high fire severity due to its unsuccessful recruitment after fire.