基于多光谱图像的烟雾检测

烟雾检测对于火灾早期防范非常重要,传统的智能视频和图像处理技术易受背景运动信息影响,抗干扰性差,且不容易区分森林水雾和燃烧产生的烟雾,森林防火误报率高。为此提出一种新的多光谱图像检测方法检测烟雾。采用多光谱成像系统,获取400至720 nm波段范围的烟雾、水雾光谱图像序列,对图像进行分层像素整合处理;利用欧氏距离度量不同分块光谱特征差异,获取动态区域光谱特征向量,根据目标与背景间光谱特征向量差异,提取烟雾、水雾区域。室内外试验结果表明：多光谱图像检测方法可用于烟雾检测,能够有效地检测并区分烟雾和水雾,与视频图像方法结合,可有效地用于森林火灾监测,降低森林火灾检测误报率。     

A study on forest fire automatic detection systems I.—Smoke plume model

Summary A mathematical model of the smoke plume produced by a forest fire is described. The model serves to obtain an estimate of the infra-red and optical signatures of the plume, taking into account several factors, such as fuel composition, combustion process characteristics, smoke diffusion, and the optical properties of the smoke and surrounding environment. The results of this model are used in a further paper to analyse the features and performances of automatic detection systems.     

基于光散射的火灾烟雾粒子激光图像探测技术研究

本文结合火灾探测领域的特点,分析了火灾烟雾粒子的光散射特性。通过与其他类型的光电感烟火灾烟雾探测技术的比较,阐述了基于光散射的火灾烟雾粒子激光图像探测技术的原理、特点和技术优势。     

Low-cost active optical system for fire surveillance

Detection of smoke plumes using active optical sensors provides many advantages with respect to passive methods of fire surveillance. However, the price of these sensors is often too high as compared to passive fire detection instruments, such as infrared and video cameras. This article describes robust and cost effective diode-laser optical sensor for automatic fire surveillance in industrial environment. Physical aspects of the sensing process allowing to simplify the hardware and software design, eventually leading to significant reduction of manufacturing and maintenance costs, are discussed. The article is published in the original.     

光散射在烟尘探测中的应用

利用烟尘微粒对光的散射作用，通过测量散射光的强度，可以作为烟尘的传感信号，这种信号可用于火警初期的报警信号，进行火警探测，从而扩充了光散射现象的应用范围。     

基于类间方差的MODIS森林火灾火点检测方法

提出一种基于火点与背景的类间方差和烟雾掩膜的红外光谱火点检测算法。该算法的潜在火点判定阈值设为305K,再利用MODIS热红外光谱通道亮温数据的类间方差值将森林火灾火点从背景中区分出来,并对低温闷烧火点采用烟雾掩膜进行识别。应用该算法对福建省和黑龙江地区的林火数据进行探测处理,与传统MODIS火点检测方法比较分析表明,该算法使用简单高效,在不同环境适应性较强,可以更准确地检测到高温明火点和一些低温闷烧火点,降低了火点的误报率和小火点的漏报率,检测效果更理想。     

Metrological characteristics of thermal video fire detection systems

We consider the possibilities for designing a new hardware and software system for fire detection, based on combination of video analytics technology and analysis of IR signals from sensors in the mid-IR and thermal IR ranges. We consider the parameters (in particular, the maximum distance for detection of the seat of a fire) for IR sensors based on various photodetectors. As a detector in the visible range, designed for early smoke or open flame detection based on analysis of sequential images of a scene, we suggest using a standard IP network camera.     

基于图像探测的早期火灾烟气模式研究

烟气是火灾过程中的重要物理现象。为改进早期火灾探测和预警,需要对基于图像探测的火灾烟气模式加以研究。以不同应用背景为依据,对烟气模式提出分类研究的思想。并以提高图像探测的可靠性和灵敏度为目的,对烟气的图像信息特征进行分析,提出了建立早期火灾烟气模式的设想。     

最佳端部耦合技术在光波导损耗测量中的实践

探讨了最佳端部耦合技术应用中面临的若干问题,如系统损耗的标定,光路调整等等。采用标样标定法、可见光辅助、平面镜调节光电监测等技术,得到了满意的结果。     

Fully automatic system for the coupling alignment of optical fibres

To realize the low-loss connection of two optical fibres intended as a transmission medium for optical signals, very tight tolerances have to be observed on account of the small geometrical dimensions involved. The present work describes a fully automatic system for the alignment of optical fibres in thex-y planes relative to a light source, which has only a single intensity peak, so that maximum optical power is launched into the fibre. The light source can be, for instance, the radiating end of a fibre, a semi-conductor laser or a light-emitting diode. The optical power coupled into the fibre serves as the control signal. Three precision displacement stages driven by stepping motors serve to align the fibres. Various sequential control algorithms for optimal coupling are investigated with reference to theory. Two algorithms are tested with the aid of a computer with reference to the alignment accuracy and speed requirements which have to be met and the one shown by theoretical simulation to be the more favourable is selected for technical realization. The resulting system was used for numerous alignment operations for the coupling of two graded-index fibres. The alignment time is below 1 s and the coupling efficiency compared with manual coupling alignment efficiency (= 100%) is above 99%, corresponding to an additional loss of <0.05 dB.     

Temperature and humidity behaviour of a corona based smoke sensor

Ionization smoke sensors are amongst the best smoke sensors; however, the little radioactive source that they include is no longer desirable since recycling gets more complicated. In this paper we discuss an electrostatic system in which corona phenomena is used to generate the ions needed to smoke detection. We show how the velocity of ions is reduced in our system for a better interaction between smoke and drifting ions. Influence of smoke, temperature and moisture is studied. It is shown that the proposed sensor has a good sensitivity compared with conventional ionic and optical smoke sensors.     

Discrimination of Smoke Particles Using Infrared Photoelectrical Detection

A method of discrimination of smoke particles using infrared photoelectrical detection is proposed in this paper. Principle of the method is that smoke particles of different materials have different scattering light intensities at same scattering angle according to Mie Theory. Facility used to realize the method includes an infrared laser as light source and an array of avalanche photodiodes distributed at assigned angles as sensing devices. Scattered light signal at these angles are collected and processed. Categories of smoke particle are discriminated by signal processing results. Experiments show that by using this method cigarette smoke, kerosene smoke and water vapor can be discriminated with rate of discrimination of greater than 95%. Application of this technology in the field of fire detection will eliminate some effects of human activities such as cigarette smoking and steam on smoke detection, and improve sensitivity as well as reliability of smoke detection.     

火灾烟颗粒的分形结构形状模拟与光散射计算

针对火灾烟颗粒的形状特点，提出并建立火灾烟颗粒分形结构凝团的形状模型，并对烟颗粒扫描电镜(SEM)图像进行分析，获取分形结构模型中的单个凝团中基本颗粒个数、凝团分形维数、基本颗粒半径等参数.利用该模型对火灾烟颗粒的形状进行模拟的结果表明，该模型能够较好反映出烟颗粒的形貌特征.利用形状模型对火灾烟颗粒散射进行初步计算表明，在其他参数相同的情况下，相对于同体积的球形颗粒，分形凝团具有前向散射较弱，后向散射较强的特征.     

基于神经网络的火灾烟雾识别方法

提出了一种基于神经网络的火灾烟雾识别方法，以波长为670nm、1060nm、1550nm的三束激光的三对消光系数比作为网络的输入，网络的输出为“火灾烟雾”和“非火灾因素”，从典型火灾烟雾和非火灾因素对多波长激光的衰减实验中选取数据，组成26种网络样本模式定义表，经391次仿真训练后，输出误差小于0．0001，并经验证实验表明，本方法对火灾烟雾和非火灾因素能进行有效的识别，是处理烟雾识别等非结构问题的一种行之有效的方法。     

Absorption losses in glasses and glass fibre waveguides

Glasses prepared from high purity starting materials were used for pulling optical waveguide fibre by a double crucible technique. The observed optical loss of the glasses and fibres can be accounted for by absorption bands arising from the presence of traces of transition metal ion impurities and the low energy tails of u.v. absorption bands. The relatively high loss of the fibres is attributed to leaching of impurities from the platinum crucibles.     

火灾烟气毒害物HCl和HCN的分析及危害性评价方法

火灾是一种失去控制的燃烧,具有双重性规律。火灾产生的有害烟气是造成人员伤亡的主要原因。本文选定火灾中主要的有害燃烧产物氯化氢(HCl)和氰化氢(HCN)作为目标气体,考虑了影响实验数据的各种因素,通过对实验数据的分析,证明了“等生成率假设”在考虑火灾中其它影响因素如温度时是不正确的;分析实验浓度变化曲线达到致死浓度的关键时间,提出了一种新的火灾烟气对人员伤亡影响的危险性评价体系,为火灾燃烧有害产物释放机理研究在工程技术上的应用奠定了基础。     

用消光系数比表征火灾烟雾的分类特征

提出了一种表征火灾烟雾分类特征的方法,采用两个已知波长的消光系数的比值表征火灾烟雾的本质特征.研究表明:两个已知波长的消光系数比仅与烟雾粒子群的粒径分布、平均粒径和折射率有关,排除了烟雾粒子浓度的影响,通过对火灾烟雾进行简单的光学测量和计算,即可得到火灾烟雾的分类特征 关键词： 消光系数比 火灾烟雾 分类特征     

Laser rangefinder architecture as a cost-effective platform for lidar fire surveillance

The possibility of early fire detection via lidar (light detection and ranging) technology implemented through a low-cost rangefinder is investigated. The evaluation is based on the variation of signal-to-noise ratio (SNR) with distance calculated on the basis of a theoretical model and determined experimentally. The theoretical SNR is obtained by combining a hydrodynamic model of the smoke plume taking into consideration the effect of wind (which enables calculation of smoke–particle distribution) and a lidar model that enables backscattered radiation intensity, detected power and, eventually, SNR to be assessed using Mie theory. The calculated values of SNR agree reasonably well with the experimental results obtained using small-scale experimental fires and show that in favourable conditions detection ranges up to about 4 km are achievable.     

枪口焰烟光学特性测试研究

针对狙击作战中对狙击手的探测、定位及告警的需求,开展了枪口焰烟光学特性测试研究。根据枪口焰烟文献资料与光学测量设备的性能参数,设计了枪口焰烟光学特性外场测试详细方案。对某5.8 mm与7.62 mm口径弹道步枪的枪口焰烟光学特性进行了外场测试。测试结果表明,枪口焰烟辐射主要集中在2μm～5μm,持续时间约为6 ms,辐射能量最大值出现在2 ms附近,5.8 mm弹道步枪枪口焰烟在中波和长波的辐射强度均高于7.62 mm。枪口焰烟光学特性的测试结果为枪口焰烟的探测及武器发射药的改进提供了重要依据。     

火灾烟雾颗粒532 nm光散射矩阵实验研究

基于自主研制的结合偏振调制和锁相检测技术的光散射实验装置,测量了两种典型火灾烟雾颗粒(棉绳阴燃烟雾和正庚烷池火烟雾)及超声雾化水滴颗粒的532 nm光散射矩阵元素随散射角的分布. 通过水滴颗粒测量结果与数值计算结果的比较验证了实验装置的可靠性. 对比分析了棉绳阴燃烟雾和正庚烷池火烟雾光散射矩阵元素随散射角的分布特征,讨论了该特征在颗粒区分上的应用. 研究了烟雾颗粒微观形貌特征对光散射矩阵的影响,发现可以利用Lorenz-Mie理论描述棉绳阴燃烟雾的光散射,表明其形貌为球形,并利用模拟退火拟合的方法得到了棉 关键词： 光散射 散射矩阵 烟雾