CN102743830B - Automatic electric switch cabinet fire extinguishing system and fire recognition method - Google Patents

Automatic electric switch cabinet fire extinguishing system and fire recognition method Download PDF

Info

Publication number
CN102743830B
CN102743830B CN201210237440.XA CN201210237440A CN102743830B CN 102743830 B CN102743830 B CN 102743830B CN 201210237440 A CN201210237440 A CN 201210237440A CN 102743830 B CN102743830 B CN 102743830B
Authority
CN
China
Prior art keywords
pixel
value
fire
flame
fire extinguishing
Prior art date
Application number
CN201210237440.XA
Other languages
Chinese (zh)
Other versions
CN102743830A (en
Inventor
赵跃进
钱永强
岳青
张黎明
王冰
Original Assignee
西安交通大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 西安交通大学 filed Critical 西安交通大学
Priority to CN201210237440.XA priority Critical patent/CN102743830B/en
Publication of CN102743830A publication Critical patent/CN102743830A/en
Application granted granted Critical
Publication of CN102743830B publication Critical patent/CN102743830B/en

Links

Abstract

The invention discloses an automatic electric switch cabinet fire extinguishing system and a fire recognition method, wherein the automatic electric switch cabinet fire extinguishing system comprises a fire extinguishing agent nozzle which is arranged in an electric switch cabinet; the fire extinguishing agent nozzle is connected with a storage device in which a non-conducting fire extinguishing agent is stored through a pipeline; an electromagnetic valve which is controlled by a central controller is arranged on a pipeline; a photoelectric sensor is also arranged in the electric switch cabinet; and the photoelectric sensor outputs a detection signal to the central controller and then the central controller judges whether to open the electromagnetic valve. The automatic electric switch cabinet fire extinguishing system is capable of responding sudden electric fire and reducing the loss caused thereby. Fire extinguishing agents such as heptafluoropropane and superfine dry powder are used and non-conducting, so that secondary danger caused by conduction of fire extinguishing agents does not occur; and the fire extinguishing agents are environmental-friendly, so that the damage to the environment caused by conduction of fire extinguishing agents also does not occur. The photoelectric sensor is used, so that the fire such as electric arc and the like in the switch cabinet can be rapidly and correctly detected.

Description

A kind of cabinet automatic fire extinguishing system and method for recognizing fire disaster

Technical field

The invention belongs to electrical switchgear technical field of security and protection, relate to a kind of electrical switchgear automatic fire extinguishing system and method for recognizing fire disaster.

Background technology

Switch cubicle is a kind of electric equipment, and outside line is introduced into master switch in cabinet, then enters sub-control switch, and each shunt needs to arrange by it.As instrument, automatic control, motor magnetic switch, various A.C. contactors etc., what have also establishes hyperbaric chamber and low-pressure chamber switch cubicle, is provided with high voltage bus, as generating plant etc.

Switch cubicle divides high-tension switch cabinet and low-tension switch cabinet:

High-tension switch cabinet refers to for electric system generating, transmission of electricity, distribution, electric energy conversion and plays the effects such as break-make, control or protection in consuming; electric pressure is in the electric equipment products of 3.6kV ~ 550kV; mainly comprise primary cut-out, high voltage isolator and grounding switch, high voltage load switch, high pressure automatic reclosing and sectionaliser, several large classes such as high-voltage actuating mechanism, high-voltage explosion-proof power distribution equipment and high-tension switch cabinet.High-tension switch cabinet should meet the relevant requirements of GB3906-1991 " 3-35kV ac metal closing switch gear " standard, be made up of cabinet and isolating switch two major part, cabinet is made up of housing, electric elements (comprising insulating part), various mechanism, secondary terminals and line etc.

Low-tension switch cabinet is applicable to the industries such as generating plant, oil, chemical industry, metallurgy, weaving, skyscraper, as the use of transmission of electricity, distribution and electric energy conversion.Product meets IEC439-1, the standard regulation of GB7251.1-1997 " low-voltage complete switch equipment ".

Switch cubicle is an elementary cell in network system, the topworks of electric power system control.Along with the application of unattended operation transformer station, intelligent power grid technology, the security protection of switch cubicle becomes particularly important.

Switch cubicle is in the course of the work due to reasons such as long-play, environmental change, line load disturbances, and there will be the disasters such as localized hyperthermia, electric arc, burning, these disasters can cause cabinet badly damaged, and electric power system is destroyed and personnel's injury.Therefore be necessary the security protection studying switch cubicle, ensure that topworks normally works, loss is reduced to bottom line, to improve power supplying efficiency.

The security protection of switch cubicle mainly prevents high temperature, burning (naked light, glow), ensure that switch normally works, or avoid when equipment component breaks down other equipment damaged, loss is reduced to minimum level, as found, switch cubicle burns, and automatically puts out a fire; Discovery switch cabinet temperature is too high, reports to the police etc.

Along with the progress of society, the awareness of the importance of fire prevention of people is more and more stronger, and the demand of automatic fire extinguishing system is more and more extensive.Automatic fire extinguishing system is the fixed fire extinguishing system of commonplace use.This system has safe and reliable, economical and practical, fire-fighting efficiency advantages of higher.

The automatic fire extinguishing system type adopted at present is both at home and abroad more, and as the associating of automatic wet pipe sprinkler system, Pre-Action Automatic Sprinkling Fire System, dry type and preact automatic sprinkler system, dry automatic sprinkler system, drench with rain automatic sprinkler system, water spray fire extinguish system etc.Automatic fire extinguishing system not only generally uses in skyscraper, public building, factory and warehouse, and has developed in home dwelling and install automatic sprinkler system.

But electrical complete equipment switch cubicle is different from above-mentioned category, in switch cubicle after electric elements or electric wire, ignition of cable, can not directly fire-fight by water.Because the impurity containing conduction in water, to be sprayed on charging equipment, then the dust impurity on infiltration equipment, then more easy conductive.As fire-fighted by water, also can reduce the insulating property of electric equipment, causing ground short circuit, or jeopardize the safety of neighbouring firefighter.

Summary of the invention

The technical matters that the present invention solves is to provide a kind of cabinet automatic fire extinguishing system, overcome can not directly with water, have pollute fire extinguishing agent fire extinguishing problem basis on, by the naked light in photodetector search switch cabinet and electric arc, improve recognition capability and the automatic processing capabilities of right fire.

The present invention is achieved through the following technical solutions:

A kind of cabinet automatic fire extinguishing system, comprises the fire extinguishing agent shower nozzle be installed in cabinet, and fire extinguishing agent shower nozzle has the storage of non-conducting extinguishing agent to be connected by pipeline with storage, and pipeline is provided with the solenoid valve by central controller controls;

Also photoelectric sensor is provided with in cabinet, photoelectric sensor comprises pick-up lens and photoelectric commutator, pick-up lens is provided with infrared fileter, pick-up lens shooting forms black white image and is converted to image digital signal through photoelectric commutator, white portion in black white image is flame, black part is divided into background, and detection signal is outputted to central controller by photoelectric sensor;

Central controller identifies fire according to the flame dynamic features in the middle of detection signal and static nature, and behavioral characteristics characterizes the flash rate of flame, and static nature characterizes the irregularly shaped of flame, carries out differentiation C=P by circularity 2/ 4 π S, wherein C is circularity, and P is girth, and S is area; When detecting that flash rate is greater than the luminophor of 1.8 in 7 ~ 12HZ, circularity, determine that it is fire, release non-conducting extinguishing agent opened by central controller controls solenoid valve.

Described fire extinguishing agent shower nozzle is provided with multiple, is connected with storage respectively by the pipeline being provided with solenoid valve.

Described storage is pressure storage type fire extinguishing agent storage, and its pressure is 1 ~ 4Mpa;

Described non-conducting extinguishing agent is heptafluoro-propane or ultra-fine dry powder extinguishing agent.

The hygrosensor be connected with central controller controls is also provided with in described cabinet.

Described hygrosensor is fibre optic temperature sensor or infrared temperature sensor, and fibre optic temperature sensor is arranged in cabinet in contactless mode with the mode of contact, infrared temperature sensor.

A method for recognizing fire disaster for cabinet automatic fire extinguishing system, comprises the following steps:

1) the photoelectric sensor shooting image in cabinet is arranged on and signal transmission, the pick-up lens of photoelectric sensor is provided with infrared fileter, pick-up lens shooting forms black white image and is converted to image digital signal through photoelectric commutator, white portion in black white image is flame, and black part is divided into background;

2) central controller receives detection signal, static nature and behavioral characteristics that flame is different from optical jammer source is extracted after pre-service is carried out to image digital signal, and quantized to form Fire Criterion vector, get rid of optical jammer source by analyzing criterion vector and identify fire;

Described static nature extracts and extracts by single-frame images computing the target signature comprising wedge angle, circularity and trapezoidal characteristics:

Sharp features: flame fringe there will be some wedge angles upward, the continuous change of fire angle number is the performance of flame fringe shake, adopts based target edge extracting wedge angle number;

Circularity feature: the shape of flame is irregular, is described the regularity of its shape by circularity, circularity be calculated as C=P 2/ 4 π S, wherein C is circularity, and P is girth, and S is area;

Trapezoidal characteristics: flame base, near comburant, makes the upper width of flame be less than bottom width, describes this feature by the distance of regional barycenter and center ordinate;

Behavioral characteristics be by multiple image computing extract comprise beat, edge expands the target signature of feature:

To beat feature: visually, flame flicking shows as flame and ceaselessly beats, extract the height of target in 25 two field pictures, be designated as x [n] (n=0,1,2 ... 24), draw frame number-height oscillogram according to x [n], then crest number is counted, only have when crest value 4 pixels tall all larger than the trough value of both sides, just carry out frequency counting, crest number is designated as target jumping frequency rate;

Edge expands feature: at fire early period of origination, flame main body increases gradually, is increased gradually describe this feature by the girth of target area;

3) after obtaining the characteristic vector data of flame, to these characteristic synthetic process: first sampling feature vectors is input to BP neural network; Output layer output valve is compared with ideal value and obtains output error; According to output error backpropagation, successively calculate forward the error of hidden layer each unit, and with these error correction front layer weights; So repeatedly the process of the input of sample information forward and error back propagation, is performed until output error and reaches acceptable degree, or till reaching the study number of times that presets; Input layer is the proper vector of the infrared light supply extracted, and the output valve of output layer does normalized, according to output valve, field condition is made to the judgement of fire condition, precarious position and normal condition.

Described comprises the pre-service of image digital signal:

1) binary conversion treatment, processes according to following Binarization methods:

Binary ( x , y ) = 0 if f ( x , y ) < T 255 if f ( x , y ) > = T (formula-1)

Wherein, Binary is binary map grey scale pixel value; F is gray-scale map grey scale pixel value; T is binary conversion treatment threshold value, makes T=155;

2) medium filtering: target pixel points and 2n neighborhood territory pixel point gray-scale value are carried out descending sequence, n is positive integer, get (n+1)th value and be assigned to object pixel, if had in object pixel and 2n neighborhood territory pixel gray-scale value, to be less than n+1 be 255, and object pixel gray-scale value is set to 0; Have that to be more than or equal to n+1 be 255, object pixel gray-scale value is set to 255;

3) target label: for distinguishing different bright connected regions, be different gray-scale values by different bright connected component labelings:

A, traversal bianry image, using the first aim pixel that traverses as sub pixel;

B, using the input parameter of sub pixel coordinate as labeling function, the gray-scale value of specifying with marks this point;

C, travel through 8 neighborhoods of Seed Points one by one, if object pixel, stop traversal, invocation flags function self;

D, marked this connected region after, turn back to a step, adopting use the same method continue mark other connected region;

4) holes filling: when the inner local luminance of flame is lower, flame body interior there will be some holes after binary conversion treatment, will fill these holes, to obtain a complete flame main body, the holes filling marked based on background area is:

The gray-scale value found is that the pixel of 0 is as sub pixel by a, traversing graph picture;

B, be labeled as Lab [n] (n=0,1,2 by all pixels of non-recursive algorithm to sub pixel place connected region ...);

C, continuation traversing graph picture, if traversing a gray-scale value is the pixel of 0, n=n+1, gets back to b step;

After d, view picture figure have traveled through, background pixel is marked as Lab [0], and hole is by distinguishing mark, and be that the grey scale pixel value of L [0] is set to 0 by mark value, the grey scale pixel value of non-L [0] is set to 255;

5) mark is revised: frame based on the 0th frame, judges that whether each boundary marker value in other frame is consistent with the boundary marker value in the 0th frame same position, if inconsistent, is revised as and the boundary marker value in same position in the 0th frame.

The extraction of described Sharp features is based target edge extracting wedge angle number, utilizes recurrence method and backtracking method target-marking border, and records the coordinate code of each frontier point; Then judge whether each object boundary point is summit according to coordinate chain code, and can wedge angle be formed with left and right neighborhood;

The wedge angle of flame fringe is similar to triangle, and summit is exactly the maximum point of flame fringe y coordinate, according to the coordinate chain code traveling through object boundary one by one, compares its y coordinate whether all greatly than the y coordinate figure of 5,10,15 pixels in left and right; If satisfied condition, then think that these pixels form a wedge angle;

When object edge generation small variations, also to judge whether the height of wedge angle meets the feature of fire angle, after pointed peak is determined, summit, line between the 15th, the left side, summit boundary pixel point and the 15th, the right boundary pixel point form a triangle, this leg-of-mutton height are designated as the height of wedge angle;

Triangle length of side a, b, c can be obtained by range formula:

D = ( y 1 - y 2 ) 2 + ( x 1 - x 2 ) 2 - - - ( 2 )

Leg-of-mutton area is asked again according to Heron's formula:

S = q ( q - a ) ( q - b ) ( q - c ) - - - ( 3 )

q=(a+b+c)/2?????????????(4)

Area formula S=c*h/2 is substituted into formula (3) obtain:

h = 2 S c = 2 q ( q - a ) ( q - b ) ( q - c ) c = ( a + b + c ) ( b + c - a ) ( a - b + c ) ( a + b - c ) 2 c - - - ( 5 )

When h exceedes the threshold value of setting, the wedge angle number of this target is added 1, and pointed peak grey scale pixel value is set to 255; When now traveling through boundary pixel, the 16th boundary pixel point that need jump to this back, summit proceeds to judge;

During described circularity feature extraction, girth P is calculated as:

1. recurrence method and backtracking method target-marking border is adopted;

2. Freeman 8 directional chain-code coding is carried out to border;

3. girth is initialized as 0, if chain code is odd number, girth adds 1, otherwise girth adds

Area S is calculated as: utilize onrecurrent region growth method to mark each bright area, and the number of pixels that traversal view picture figure asks different mark value corresponding is as the area of each marked region;

Being calculated as of the regional barycenter of trapezoidal characteristics and the distance of center ordinate:

The ordinate of regional barycenter is gy = 1 S &Sigma; y &Element; R y - - - ( 7 )

Wherein gy is center of gravity ordinate, and S is target area area, and R is target area, and y is the ordinate of each pixel in target area;

Classification and boundary coordinate calculating center of gravity ordinate according to frontier point:

A, utilize Object Area Algorithms reference area S based on frontier point classification;

B, order initialization Y=0;

C, traversal object boundary point classification code, if frontier point (x n, y n) be lower boundary point, if frontier point (x n, y n) be coboundary point, if frontier point (x n, y n) be horizontal direction summit, Y=Y+y n; (wherein n=0,1,2 ... the height of h-1, h representative image.)

D, the final Y result of Y is updated to (7) formula, obtains center of gravity ordinate;

E, according to the minimum and maximum value of object boundary point ordinate, calculate target's center ordinate y c:

y min=Min(y);y max=Max(y);y∈R;

y c = y max + y min 2 - - - ( 8 ) .

The backpropagation of employing Sigmoid type output function is:

(1) selected weight coefficient initial value;

(2) repeat following process until convergence, characteristic quality of sample inputted network successively:

1. o is exported from each layer unit of forward calculation j

net j = &Sigma; i &omega; ij o i - - - ( 9 )

o j = 1 / ( 1 - e - net j ) - - - ( 10 )

2. each elemental error δ of output layer is calculated j

δ j=(y-o j)o j(1-o j)??????????(11)

3. reverse transfer error, calculates each elemental error δ of each hidden layer j

&delta; j = ( y - o j ) o j &Sigma; k &omega; jk &delta; k - - - ( 12 )

4. the correction of each layer weights is calculated

△ω ij(t)=α△ω ij(t-1)+ηδ jo j??(13)

5. according to each layer weights of each layer modified weight amount correction

ω ij(t+1)=ω ij(t)+△ω ij(t)?????(14)

When employing three layers of feedforward network carry out the classification of fire condition, precarious position and normal condition, input layer number is the dimension of proper vector; Hidden layer nodes is round up, wherein p is input layer number; Output layer nodes is number of categories; In forward process, weights initial value chooses usually ± random number in 0.3 interval; In reverse transfer method, step-length η directly sounds out 0.1 ~ 3; Inertia factor alpha is selected between 0.9 ~ 1.

Compared with prior art, the present invention has following useful technique effect:

Cabinet automatic fire extinguishing system provided by the invention, by the situation timing monitoring of photodetector to switch cubicle, judged by central controller, and non-conducting extinguishing agent sprays fire extinguishing by shower nozzle after solenoid valve is opened, fully achieve the self-extinguishing of cabinet.Be particularly suitable for the harsh and unforgiving environments as generating plant, power supply department transformer station, unattended operation transformer station, and volume is little, does not affect the normal operation of equipment under test, and easy installation and reliable.

Cabinet automatic fire extinguishing system provided by the invention, have employed heptafluoro-propane or ultra-fine dry powder extinguishing agent, and these fire extinguishing agents are nonconducting, therefore secondary will be caused to endanger due to fire extinguishing agent conduction; And a little fire extinguishing agent is environmental protection, therefore can not lead working the mischief to environment due to fire extinguishing agent.

Heptafluoro-propane, at normal temperatures in gaseous state, colorless and odorless, non-conductively has good gas phase electrical insulating property, corrosion-free, and limit without environmental protection, it is shorter that air retains the phase, has good spatter property (vaporize completely in an atmosphere and do not stay residue).Extinguishing mechanism mainly interrupts pyric chain, and blow-off velocity is exceedingly fast, and this is favourable to Rescued Protection sophisticated electronics and valuables.Non-toxic reaction (NOAEL) concentration of heptafluoro-propane is 9%, and toxic reaction (LOAEL) concentration is 10.5%, and the design concentration of heptafluoro-propane is generally less than 10%, to human-body safety.Its feature and the good physical property being applicable to fire extinguishing system use.

Ultra-fine dry powder extinguishing agent, nontoxic, pollution-free, do not destroy atmospheric ozone layer, corrosion-free to protection, to human body skin and respiratory tract non-stimulated, rapidly, efficiency is high in fire extinguishing.

Cabinet automatic fire extinguishing system provided by the invention, adopts photoelectric sensor, the fire in switch cubicle fast, accurately can be detected, as electric arc etc.Monitoring site situation can be reflected intuitively by detector by image, just automatically can identify fire at fire early period of origination, record fire.And monitoring range is large, this is that conventional fire detector cannot be accomplished.

Cabinet automatic fire extinguishing system provided by the invention, is applicable to various switch cubicle and installs and uses: both can be used for the switch cubicle that Novel switch cabinet also can be used for using simultaneously; Also multiple switch cubicle extinguishing device can be networked, possessed remote monitoring function further.

Accompanying drawing explanation

Fig. 1 is one of structural representation of the present invention;

Fig. 2 is structural representation two of the present invention;

Fig. 3 is structural representation two of the present invention;

Wherein: 1 is cabinet; 2 is central controller; 3 is photoelectric sensor; 4 is storage; 5 is solenoid valve; 6 is pipeline; 7 is fire extinguishing agent shower nozzle;

Fig. 4 is the image processing flow figure of detection signal;

Fig. 5 is Image semantic classification process flow diagram;

Fig. 6 is object pixel 8 neighborhood traversal order;

Fig. 7 is image characteristics extraction figure;

Fig. 8 is that Sharp features extracts schematic diagram;

Fig. 9 is Freeman 8 directional chain-code schematic diagram.

Embodiment

Be described in further detail the present invention below in conjunction with accompanying drawing, the explanation of the invention is not limited.

As shown in Figure 1, a kind of cabinet automatic fire extinguishing system, comprise the fire extinguishing agent shower nozzle 7 be installed in cabinet 1, fire extinguishing agent shower nozzle 7 has the storage 4 of non-conducting extinguishing agent to be connected by pipeline 6 with storage, and pipeline 6 is provided with the solenoid valve 5 controlled by central controller 2; Also be provided with photoelectric sensor 3 in cabinet 1, detection signal is outputted to central controller 2 by photoelectric sensor 3, is determined whether to open solenoid valve 5 by it.

Concrete photoelectric sensor 3 comprises pick-up lens and photoelectric commutator, pick-up lens is provided with infrared fileter, pick-up lens shooting forms black white image and is converted to image digital signal through photoelectric commutator, white portion in black white image is flame, black part is divided into background, and detection signal is outputted to central controller by photoelectric sensor;

Central controller identifies fire according to the flame dynamic features in the middle of detection signal and static nature, and behavioral characteristics characterizes the flash rate of flame, and static nature characterizes the irregularly shaped of flame, carries out differentiation C=P by circularity 2/ 4 π S, wherein C is circularity, and P is girth, and S is area; When detecting that flash rate is greater than the luminophor of 1.8 in 7 ~ 12HZ, circularity, determine that it is fire, release non-conducting extinguishing agent opened by central controller controls solenoid valve.

As shown in Figure 2 and Figure 3, in order to the requirement of the different levels (twice or three layers) or different space separately that adapt to switch cubicle, space for each opposition in cabinet 1 is equipped with independently photoelectric sensor 2, and configure a fire extinguishing agent shower nozzle 7, each photoelectric sensor 3 is all connected with central controller 2, carries out enforcement monitoring, when needs time to it, central controller 2 sends instruction unpack fire extinguishing agent shower nozzle 7, and fire extinguishing agent shower nozzle 7 sprays non-conducting extinguishing agent and puts out a fire.Accordingly, central controller, when detecting that fire occurs, needs to judge fire place, and then the solenoid valve controlled in corresponding pipeline is opened, and release fire extinguishing agent is put out a fire.

Concrete, pipeline 6 adopts withstand voltage high connecting pipe, needing the position mounting spray head of fire extinguishing, nozzle connecting is received one or more fire extinguishing position, ensures that fire extinguishing agent directly sprays kindling point, improves fire-fighting efficiency.

Due to the particular/special requirement of described cabinet, described non-conducting extinguishing agent is specifically selected to be the fire extinguishing agent such as heptafluoro-propane, ultra-fine dry powder.

Heptafluoro-propane is at normal temperatures in gaseous state, colorless and odorless, non-conductive, corrosion-free, and limit without environmental protection, it is shorter that air retains the phase.Extinguishing mechanism mainly interrupts pyric chain, and blow-off velocity is exceedingly fast, and this is favourable to Rescued Protection sophisticated electronics and valuables.Non-toxic reaction (NOAEL) concentration of heptafluoro-propane is 9%, and toxic reaction (LOAEL) concentration is 10.5%, and the design concentration of heptafluoro-propane is generally less than 10%, to human-body safety.Its feature has good spatter property (vaporize completely in an atmosphere and do not stay residue), good gas phase electrical insulating property and the good physical property being applicable to fire extinguishing system use.

Ultra-fine dry powder extinguishing agent is a kind of efficient and environment-friendly type a new generation fire product.Conventional ABC ultra-fine dry powder extinguishing agent is nontoxic, pollution-free, do not destroy atmospheric ozone layer, corrosion-free to protection, to human body skin and respiratory tract non-stimulated, rapidly, efficiency is high, is the substitute of halon fire agent in fire extinguishing.

Be pressure storage type fire extinguishing agent storage in described storage, its pressure, for (1 ~ 4MPa) is after solenoid valve is opened under the control of the controller, makes the fire extinguishing agent in storage directly be sprayed onto fire extinguishing position through piping under the effect of himself pressure, realizes fire extinguishing.

Further, be also provided with the hygrosensor be connected with central controller controls in described cabinet, hygrosensor is used for the temperature of critical component in search switch cabinet; Hygrosensor is fibre optic temperature sensor or infrared temperature sensor, and fibre optic temperature sensor is arranged in cabinet in contactless mode with the mode of contact, infrared temperature sensor.

Wherein, fibre optic temperature sensor is the temperature sensor developed for the harsh and unforgiving environments of power industry (as generating plant, power supply department transformer station, unattended operation transformer station), the mode of employing ceramic package, its encapsulating material employing class of insulation is high, high temperature-resistant polymer material.Volume is little, does not affect the normal operation of equipment under test, and easy installation and reliable.

Naked light in photodetector main search switch cabinet and electric arc: this detector is by naked light and the such light signal of electric arc, is converted to image digital signal, applied for machines vision, Digital Image Processing and mode identification technology determination fire through photoelectric commutator.

Central processing unit realizes monitoring various signal, judge, controlling, and is the topworks protected switch cubicle according to preset value.When photodetector detects naked light, notice central controller, central controller is opened extinguishing device and is put out a fire.When hygrosensor detect switch in-cabinet temperature reach preset value time, notice central processing unit, central processing unit to carry out reporting to the police outwards transmission or start fire extinguishing according to setting means.

Above-mentioned cabinet automatic fire extinguishing system adopts following extinguishing method to carry out the identification of fire, comprises the following steps:

First be the collection to flare up fire:

Because the spectrum 95% of flame optical radiation concentrates on near-infrared band, therefore according to the spectral characteristic of flame, gather the image of this infrared band, thus get rid of visible ray to the interference of fire identification.Visible wavelength is generally 380-780nm, and the light of other wavelength coverage is invisible light, and flame optical wavelength mainly concentrates on 950-2000nm.

Select ccd video camera as the pick-up lens of photoelectric sensor, black-white CCD video camera not only has stronger spectral response to the visible ray of 380-780nm, and also have certain spectral response to the near infrared light of 780-1100nm, after pick-up lens adds the saturating infrared fileter of a slice, black-white CCD video camera can only be sensed, and wavelength is the near infrared light of 780 ~ 1100nm, black white image is formed in video camera, white portion is flame, black part be divided into background or other can not enter the jamming light source of camera lens, thus exclusive segment optical jammer source and obtain the black white image mainly comprising flame information.

Further because flame has the static nature and behavioral characteristics being different from optical jammer source, by mathematical model and geometric model by after these characteristic quantifications, form executable Fire Criterion vector, central processing unit is got rid of optical jammer source by analyzing criterion vector and identifies fire.Its behavioral characteristics shows that flame flashes, and fame disturbance frequency mainly concentrates on 7 ~ 12HZ, and great majority rock or fix interference source height change frequency is do not reach 8Hz.Its static nature is that flame is irregular, is angular, by circularity as distinguishing rule: C=P relative to other light sources 2/ 4 π S, C in formula---circularity; P---girth; S-area.The circularity of interference source is generally greater than 0.8 and is less than 1.5, and the circularity of flame is generally greater than 1.8.The behavioral characteristics of flame and static nature as proper vector, just can be determined whether breaking out of fire according to proper vector by central controller.Namely when finding that there is flash rate at 7 ~ 12HZ, when circularity is greater than the luminophor of 1.8, namely fire is determined that it is.

To the identification of fire mainly to the image procossing of detection signal, see Fig. 4, the image procossing of detection signal comprise gather image, Image semantic classification, characteristic vector pickup, fire judge and the judgement of position occurs fire.Monitoring site situation can be reflected intuitively by image like this, just automatically can identify fire at fire early period of origination, record fire.And monitoring range is large, this is that conventional fire detector cannot be accomplished.

Be described in detail the processing stage of each below.

1, Image semantic classification: Image semantic classification process flow diagram as shown in Figure 5, comprises binaryzation, medium filtering, cavity filling, target label, correction mark.

1.1 binary conversion treatment, process according to following Binarization methods:

Binary ( x , y ) = 0 if f ( x , y ) < T 255 if f ( x , y ) > = T (formula-1)

In formula:

Binary---binary map grey scale pixel value;

F---gray-scale map grey scale pixel value;

T---binary conversion treatment threshold value, makes T=155.

1.2 medium filterings: be positive integer by target pixel points and flame pixels and 2n(n) neighborhood territory pixel point gray-scale value carries out descending sequence, and get (n+1)th value and be assigned to object pixel, it is larger that neighborhood is chosen, and filter effect is more obvious, but operand also corresponding increase.Due to all pixel gray-scale values non-zero namely 255 in binary map, can save the process of sequence, if had in object pixel and 2n (n>1) neighborhood territory pixel gray-scale value, to be less than n+1 be 255, and object pixel gray-scale value is set to 0; Have that to be more than or equal to n+1 be 255, object pixel gray-scale value is set to 255.

1.3 target labels: owing to may there is multiple bright connected region in piece image, in order to distinguish different bright connected regions, needing different bright connected component labelings is different gray-scale values.Object pixel 8 neighborhood traversal order as shown in Figure 6.Concrete operations are as follows:

A, traversal bianry image, using the first aim pixel that traverses as sub pixel;

B, using the input parameter of sub pixel coordinate as labeling function, the gray-scale value of specifying with marks this point;

C, by the pixel 8 neighborhood traversal order shown in Fig. 6, travel through 8 neighborhoods of Seed Points one by one, if object pixel, stop traversal, invocation flags function self;

D, marked this connected region after, turn back to a step, adopting use the same method continue mark other connected region.

1.4 holes filling methods: when the inner local luminance of flame is lower, flame body interior there will be some holes after binary conversion treatment, and these holes are classified as background.In order to extract the proper vector of target area easily, needing to fill these holes, obtaining a complete flame main body.

In binary map, black region comprises background and hole, and wherein background is a maximum black region, and hole is multiple little black regions, and background and hole are separated from each other.Utilize region growing algorithm first to background area mark, then will to mark perforated.Senior general's background segment may be crossed due to flame area and become two black regions, so first add that pixel is wide, gray-scale value is the frame of 0 to view picture figure before mark, guarantee that first Seed Points traversed is background pixel, and make background can form a maximum black connected region.

Hole filling algorithms concrete steps based on background area mark are as follows:

The gray-scale value found is that the pixel of 0 is as sub pixel by a, traversing graph picture;

B, be labeled as Lab [n] (n=0,1,2 by all pixels of non-recursive algorithm to sub pixel place connected region ...);

C, continuation traversing graph picture, if traversing a gray-scale value is the pixel of 0, n=n+1, gets back to b step;

After d, view picture figure have traveled through, background pixel is marked as Lab [0], hole is labeled as Lab [1], Lab [2], Lab [3] respectively ... be that the grey scale pixel value of L [0] is set to 0 by mark value, the grey scale pixel value of non-L [0] is set to 255, completes holes filling.

1.5 revise labeling algorithms: due to flame periphery exist flash interference or inside there is hole, bright area border in 25 two field pictures in same position can be marked as different gray-scale values, so just cannot obtain the proper vector of same bright area on 25 two field pictures, so need to revise mark value.

Modification method is frame based on the 0th frame, judges that whether each boundary marker value in other frame is consistent with the boundary marker value in the 0th frame same position, if inconsistent, is revised as and the boundary marker value in same position in the 0th frame.The method can not only revise the mark value on each border, and can delete the optical jammer source moving fast or glimmer.

2, feature extraction: due to the impact of environment, except flame, also may there is other optical jammer source in the visual field in detection, thus the identification of impact to flame.Although flame and interference source all present the region of higher brightness in black white image, flame has oneself static nature and behavioral characteristics.Static nature is the target signature extracted by single-frame images computing, mainly comprises wedge angle, circularity and trapezoidal characteristics.Behavioral characteristics is the target signature extracted by multiple image computing, mainly comprise beat, edge expand feature.Feature extraction flow process as shown in Figure 7.

2.1 Sharp features

Due to the reason that flame fringe shake and thermal current rise, flame fringe there will be some wedge angles upward, the continuous change of fire angle number is that of flame fringe shake obviously shows, and the wedge angle number as the object such as candle or electric light is almost constant.In order to improve the speed extracting wedge angle number, adopt the method for based target edge extracting wedge angle number, concrete steps are as follows:

1) utilize recurrence method and backtracking method target-marking border, and record the coordinate code of each frontier point;

2) judge whether each object boundary point is summit according to coordinate chain code, and can wedge angle be formed with left and right neighborhood.

The wedge angle of flame fringe is similar to a triangle, not only has a summit but also has certain height.Summit is exactly the maximum point of flame fringe y coordinate, according to the coordinate chain code traveling through object boundary one by one, compares its y coordinate whether all greatly than the y coordinate figure of 5,10,15 pixels in left and right; If satisfied condition, then think that these 31 pixels (15, left side pixel, 15, the right pixel) can form a wedge angle.

When object edge generation small variations, all can produce little projection at random, so also will judge whether the height of wedge angle meets the feature of fire angle.After pointed peak is determined, as shown in Figure 8, summit, line between the 15th, the left side, summit boundary pixel point and the 15th, the right boundary pixel point form a triangle, this leg-of-mutton height can be designated as the height of wedge angle.

Triangle length of side a, b, c can be obtained by range formula:

D = ( y 1 - y 2 ) 2 + ( x 1 - x 2 ) 2 - - - ( 2 )

Leg-of-mutton area is asked again according to Heron's formula:

S = q ( q - a ) ( q - b ) ( q - c ) - - - ( 3 )

q=(a+b+c)/2????????????(4)

Area formula S=c*h/2 is substituted into formula (formula-3) obtain:

h = 2 S c = 2 q ( q - a ) ( q - b ) ( q - c ) c = ( a + b + c ) ( b + c - a ) ( a - b + c ) ( a + b - c ) 2 c - - - ( 5 )

When h exceedes the threshold value of setting, the wedge angle number of this target can be added 1, and pointed peak grey scale pixel value is set to 255.Rear 15 coordinates contiguous due to this summit may also meet wedge angle condition, but they represent same wedge angle, and when now traveling through boundary pixel, the 16th boundary pixel point that need jump to this back, summit proceeds to judge.

2.2 circularity features: the shape of flame is irregular, and the shape of most of interference source (incandescent lamp, electric torch, candle flame) is all more regular, can be described the regularity of shape by circularity.The computing formula of circularity is as follows:

C = P 2 4 &pi;S - - - ( 6 )

In formula:

C---circularity;

P---girth;

S---area.

* target perimeter P calculation procedure:

1. recurrence method and backtracking method target-marking border is adopted;

2. carry out Freeman 8 directional chain-code coding to border, coding rule as shown in Figure 9;

3. girth is initialized as 0, if chain code is odd number, girth adds 1, otherwise girth adds

* target area S computing method:

Utilize onrecurrent region growth method to mark each bright area, the number of pixels that traversal view picture figure asks different mark value corresponding is as the area of each marked region.When the more or single bright area area of bright area number is larger, onrecurrent region growing algorithm is consuming time longer.

2.3 beat feature: outwardly, fame disturbance is rambling, and the spectrum signature of flare up fire has distinctive rule in fact.Visually, flame flicking shows as flame and ceaselessly beats, and fame disturbance frequency mainly concentrates on 7 ~ 12Hz, and great majority rock or fix interference source height change frequency does not reach 8Hz, can get rid of the lower interference source of height change frequency according to this.

First extract the height of target in 25 two field pictures, be designated as x [n] (n=0,1,2 ... 24), draw frame number-height oscillogram according to x [n], then crest number is counted, crest number is designated as target jumping frequency rate.

Because there is slight vibration at fixing interference source edge, and slightly to tremble the height change caused be very little, generally can not exceed twice target perimeter, namely height change generally can not more than 4 pixels, so only have when crest value 4 pixels tall all larger than the trough value of both sides, just carry out frequency counting, the method can obtain vibration of flame frequency fast.

2.4 trapezoidal characteristics: generally, because flame base is near comburant, the upper width of flame is made to be significantly less than bottom width, its body is similar to a ladder type, generally describe this feature by the height of C.G. of target and the overall height ratio of target, height of C.G. is the difference of center of gravity ordinate and target lower boundary ordinate minimum value, and target overall height is the difference of target coboundary ordinate maximal value and lower boundary ordinate minimum value.For the jamming light source that shape is more regular, its height of C.G. and centre-height closely, and are similar to trapezoidal flame for body, and its height of C.G. is significantly less than centre-height, in order to shortcut calculation and raising degree of accuracy, by the distance of center of gravity and center ordinate, this feature is described.

The ordinate that regional barycenter is:

gy = 1 S &Sigma; y &Element; R y (formula-7)

In formula:

Gy---center of gravity ordinate;

S---target area area;

R---target area (i.e. flame region);

Y---the ordinate of each pixel in target area.

Because image all processes target area based on border, so the coordinate of target area interior pixels cannot directly obtain, so calculate center of gravity ordinate according to the classification of frontier point and boundary coordinate, computing method are as follows:

A, utilize Object Area Algorithms reference area S based on frontier point classification;

B, order initialization Y=0;

C, traversal object boundary point classification code, if frontier point (x n, y n) be lower boundary point, if frontier point (x n, y n) be coboundary point, if frontier point (x n, y n) be horizontal direction summit, Y=Y+y n; (wherein n=0,1,2 ... the height of h-1, h representative image.)

D, the final Y result of Y is updated to (7) formula, obtains center of gravity ordinate;

E, according to object boundary point ordinate maximin, target's center ordinate y can be calculated c:

y c = y max + y min 2 (formula-8)

2.5 edges expand feature

At fire early period of origination, flame main body increases gradually, and the main body of fixing interference source is substantially constant, increases gradually to describe this feature by the girth (calculating identical with above-mentioned target perimeter P) of target area.

3, fire judges: after obtaining the characteristic vector data of flame, need these characteristic synthetic process, discriminant parameter is difficult to by manually determining.First sampling feature vectors is input to BP neural network; Output layer output valve is compared with ideal value and obtains output error; According to output error backpropagation, successively calculate forward the error of hidden layer each unit, and with these error correction front layer weights.The sample information forward input so gone round and begun again and the process of error back propagation, be performed until output error and reach acceptable degree, or till reaching the study number of times that presets.

Concrete employing Sigmoid type output function, back-propagation algorithm is as follows:

(1) selected weight coefficient initial value.

(2) following process is repeated until convergence (characteristic quality of sample is inputted network successively):

1. o is exported from each layer unit of forward calculation j

net j = &Sigma; i &omega; ij o i (formula-9)

o j = 1 / ( 1 - e - net j ) (formula-10)

2. each elemental error δ of output layer is calculated j

δ j=(y-o j) o j(1-o j) (formula-11)

3. reverse transfer error, calculates each elemental error δ of each hidden layer j

&delta; j = ( y - o j ) o j &Sigma; k &omega; jk &delta; k (formula-12)

4. the correction of each layer weights is calculated

△ ω ij(t)=α △ ω ij(t-1)+η δ jo j(formula-13)

5. according to each layer weights of each layer modified weight amount correction

ω ij(t+1)=ω ij(t)+△ ω ij(t) (formula-14)

Back-propagation algorithm can revise hidden layer weights, and when adopting gradient method to ask nonlinear function extreme value, is likely absorbed in local minizing point, can not ensures to converge to global minimum point.

When employing three layers of feedforward network are classified, input layer number is the dimension of proper vector; Hidden layer nodes is round up, wherein p is input layer number; Output layer nodes is number of categories.In forward process, weights initial value chooses usually ± random number in 0.3 interval.In reverse transfer method, step-length η affects very large on convergence and optimal value, usually directly sound out 0.1 ~ 3; Inertia factor alpha affects speed of convergence, generally selects between 0.9 ~ 1.

Input layer is the proper vector of the infrared light supply extracted, and the output valve of output layer does normalized:

(1) out=1 (fire condition)

(2) out=0.5 (precarious position)

(3) out=0 (normal condition)

Or, when using BP neural network model to differentiate fire, differentiate as follows according to output interval value value to field condition:

(1) 0.75<=out<=1 fire condition

(2) 0.25<out<0.75 precarious position

(3) 0<=out<=0.25 normal condition.

Claims (5)

1. a method for recognizing fire disaster for cabinet automatic fire extinguishing system, is characterized in that, comprises the following steps:
1) the photoelectric sensor shooting image in cabinet is arranged on and signal transmission, the pick-up lens of photoelectric sensor is provided with infrared fileter, pick-up lens shooting forms black white image and is converted to image digital signal through photoelectric commutator, white portion in black white image is flame, and black part is divided into background;
2) central controller receives detection signal, static nature and behavioral characteristics that flame is different from optical jammer source is extracted after pre-service is carried out to image digital signal, and quantized to form Fire Criterion vector, get rid of optical jammer source by analyzing criterion vector and identify fire;
Described static nature extracts and extracts by single-frame images computing the target signature comprising wedge angle, circularity and trapezoidal characteristics:
Sharp features: flame fringe there will be some wedge angles upward, the continuous change of fire angle number is the performance of flame fringe shake, adopts based target edge extracting wedge angle number;
Circularity feature: the shape of flame is irregular, is described the regularity of its shape by circularity, circularity be calculated as C=P 2/ 4 π S, wherein C is circularity, and P is girth, and S is area;
Trapezoidal characteristics: flame base, near comburant, makes the upper width of flame be less than bottom width, describes this feature by the distance of regional barycenter and center ordinate; Being calculated as of the regional barycenter of trapezoidal characteristics and the distance of center ordinate:
The ordinate of regional barycenter is gy = 1 S &Sigma; y &Element; R y - - - ( 7 )
Wherein gy is center of gravity ordinate, and S is target area area, and R is target area, and y is the ordinate of each pixel in target area;
Classification and boundary coordinate calculating center of gravity ordinate according to frontier point:
A, utilize Object Area Algorithms reference area S based on frontier point classification;
B, order initialization Y=0;
C, traversal object boundary point classification code, if frontier point (x n, y n) be lower boundary point, if frontier point (x n, y n) be coboundary point, if frontier point (x n, y n) be horizontal direction summit, Y=Y+y n; Wherein n=0,1,2 ... the height of h-1, h representative image;
D, the final Y result of Y is updated to (7) formula, obtains center of gravity ordinate;
E, according to the minimum and maximum value of object boundary point ordinate, calculate target's center ordinate y c:
y min=Min(y);y max=Max(y);y∈R;
y c = y max + y min 2 - - - ( 8 ) ;
Behavioral characteristics be by multiple image computing extract comprise beat, edge expands the target signature of feature:
To beat feature: visually, flame flicking shows as flame and ceaselessly beats, and extracts the height of target in 25 two field pictures, is designated as x [n], n=0,1,2 ... 24, draw frame number-height oscillogram according to x [n], then crest number is counted, only have when crest value 4 pixels tall all larger than the trough value of both sides, just carry out frequency counting, crest number is designated as target jumping frequency rate;
Edge expands feature: at fire early period of origination, flame main body increases gradually, is increased gradually describe this feature by the girth of target area;
3) after obtaining the characteristic vector data of flame, to these characteristic synthetic process: first static nature and behavioral characteristics vector are input to BP neural network; Output layer output valve is compared with ideal value and obtains output error; According to output error backpropagation, successively calculate forward the error of hidden layer each unit, and with these error correction front layer weights; So repeatedly the process of the input of sample information forward and error back propagation, is performed until output error and reaches acceptable degree, or till reaching the study number of times that presets; Input layer is the proper vector of the infrared light supply extracted, and the output valve of output layer does normalized, according to output valve, field condition is made to the judgement of fire condition, precarious position and normal condition.
2. the method for recognizing fire disaster of cabinet automatic fire extinguishing system as claimed in claim 1, it is characterized in that, described comprises the pre-service of image digital signal:
1) binary conversion treatment, processes according to following Binarization methods:
Binary ( x , y ) = 0 if f ( x , y ) < T 255 if f ( x , y ) > = T (formula-1)
Wherein, Binary is binary map grey scale pixel value; F is gray-scale map grey scale pixel value; T is binary conversion treatment threshold value, makes T=155;
2) medium filtering: target pixel points and 2n neighborhood territory pixel point gray-scale value are carried out descending sequence, n is positive integer, get (n+1)th value and be assigned to object pixel, if had in object pixel and 2n neighborhood territory pixel gray-scale value, to be less than n+1 be 255, and object pixel gray-scale value is set to 0; Have that to be more than or equal to n+1 be 255, object pixel gray-scale value is set to 255;
3) target label: for distinguishing different bright connected regions, be different gray-scale values by different bright connected component labelings:
A, traversal bianry image, using the first aim pixel that traverses as sub pixel;
B, using the input parameter of sub pixel coordinate as labeling function, the gray-scale value of specifying with marks this point;
C, travel through 8 neighborhoods of Seed Points one by one, if object pixel, stop traversal, invocation flags function self;
D, marked this connected region after, turn back to a step, adopting use the same method continue mark other connected region;
4) holes filling: when the inner local luminance of flame is lower, flame body interior there will be some holes after binary conversion treatment, will fill these holes, to obtain a complete flame main body, the holes filling marked based on background area is:
The gray-scale value found is that the pixel of 0 is as sub pixel by a, traversing graph picture;
B, be labeled as Lab [n] (n=0,1,2 by all pixels of non-recursive algorithm to sub pixel place connected region ...);
C, continuation traversing graph picture, if traversing a gray-scale value is the pixel of 0, n=n+1, gets back to b step;
After d, view picture figure have traveled through, background pixel is marked as Lab [0], and hole is by distinguishing mark, and be that the grey scale pixel value of L [0] is set to 0 by mark value, the grey scale pixel value of non-L [0] is set to 255;
5) mark is revised: frame based on the 0th frame, judges that whether each boundary marker value in other frame is consistent with the boundary marker value in the 0th frame same position, if inconsistent, is revised as and the boundary marker value in same position in the 0th frame.
3. the method for recognizing fire disaster of cabinet automatic fire extinguishing system as claimed in claim 1, it is characterized in that, the extraction of described Sharp features is based target edge extracting wedge angle number, utilizes recurrence method and backtracking method target-marking border, and records the coordinate code of each frontier point; Then judge whether each object boundary point is summit according to coordinate chain code, and can wedge angle be formed with left and right neighborhood;
The wedge angle of flame fringe is similar to triangle, and summit is exactly the maximum point of flame fringe y coordinate, according to the coordinate chain code traveling through object boundary one by one, compares its y coordinate whether all greatly than the y coordinate figure of 5,10,15 pixels in left and right; If satisfied condition, then think that these pixels form a wedge angle;
When object edge generation small variations, also to judge whether the height of wedge angle meets the feature of fire angle, after pointed peak is determined, summit, line between the 15th, the left side, summit boundary pixel point and the 15th, the right boundary pixel point form a triangle, this leg-of-mutton height are designated as the height of wedge angle;
Triangle length of side a, b, c can be obtained by range formula:
D = ( y 1 - y 2 ) 2 + ( x 1 - x 2 ) 2 - - - ( 2 )
Leg-of-mutton area is asked again according to Heron's formula:
S = q ( q - a ) ( q - b ) ( q - c ) - - - ( 3 )
q=(a+b+c)/2??(4)
Area formula S=c*h/2 is substituted into formula (3) obtain:
h = 2 S c = 2 q ( q - a ) ( q - b ) ( q - c ) c = ( a + b + c ) ( b + c - a ) ( a - b + c ) ( a + b - c ) 2 c - - - ( 5 )
When h exceedes the threshold value of setting, the wedge angle number of this target is added 1, and pointed peak grey scale pixel value is set to 255; When now traveling through boundary pixel, the 16th boundary pixel point that need jump to this back, summit proceeds to judge;
During described circularity feature extraction, girth P is calculated as:
1. recurrence method and backtracking method target-marking border is adopted;
2. Freeman 8 directional chain-code coding is carried out to border;
3. girth is initialized as 0, if chain code is odd number, girth adds 1, otherwise girth adds
Area S is calculated as: utilize onrecurrent region growth method to mark each bright area, and the number of pixels that traversal view picture figure asks different mark value corresponding is as the area of each marked region.
4. the method for recognizing fire disaster of cabinet automatic fire extinguishing system as claimed in claim 1, is characterized in that, the backpropagation of employing Sigmoid type output function is:
(1) selected weight coefficient initial value;
(2) repeat following process until convergence, characteristic quality of sample inputted network successively:
1. ο is exported from each layer unit of forward calculation j
net j = &Sigma; i &omega; ij o i - - - ( 9 )
o j = 1 / ( 1 - e - net j ) - - - ( 10 )
2. each elemental error δ of output layer is calculated j
δ j=(y-ο j)ο j(1-ο j)??(11)
3. reverse transfer error, calculates each elemental error δ of each hidden layer j
&delta; j = ( y - o j ) o j &Sigma; k &omega; jk &delta; k - - - ( 12 )
4. the correction of each layer weights is calculated
Δω ij(t)=αΔω ij(t-1)+ηδ jο j??(13)
5. according to each layer weights of each layer modified weight amount correction
ω ij(t+1)=ω ij(t)+Δω ij(t)??(14)。
5. the method for recognizing fire disaster of cabinet automatic fire extinguishing system as claimed in claim 1, it is characterized in that, when employing three layers of feedforward network carry out the classification of fire condition, precarious position and normal condition, input layer number is the dimension of proper vector; Hidden layer nodes is round up, wherein p is input layer number; Output layer nodes is number of categories; In forward process, weights initial value chooses ± random number in 0.3 interval; In reverse transfer method, step-length η sounds out between 0.1 ~ 3; Inertia factor alpha is selected between 0.9 ~ 1.
CN201210237440.XA 2012-07-10 2012-07-10 Automatic electric switch cabinet fire extinguishing system and fire recognition method CN102743830B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201210237440.XA CN102743830B (en) 2012-07-10 2012-07-10 Automatic electric switch cabinet fire extinguishing system and fire recognition method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201210237440.XA CN102743830B (en) 2012-07-10 2012-07-10 Automatic electric switch cabinet fire extinguishing system and fire recognition method

Publications (2)

Publication Number Publication Date
CN102743830A CN102743830A (en) 2012-10-24
CN102743830B true CN102743830B (en) 2015-01-21

Family

ID=47024544

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201210237440.XA CN102743830B (en) 2012-07-10 2012-07-10 Automatic electric switch cabinet fire extinguishing system and fire recognition method

Country Status (1)

Country Link
CN (1) CN102743830B (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3321908A1 (en) * 2016-11-11 2018-05-16 Kidde Technologies, Inc. Fiber optic based monitoring of temperature and/or smoke conditions at electronic components
WO2018089477A1 (en) * 2016-11-11 2018-05-17 Carrier Corporation High sensitivity fiber optic based detection

Families Citing this family (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103845832B (en) * 2012-11-30 2016-04-27 珠海格力电器股份有限公司 Extinguishing device and control method, electrical box, switch cubicle and electric equipment
CN103423763B (en) * 2013-07-18 2015-12-02 武汉九州三维燃烧科技有限公司 A kind of method revising radiation energy signal static deviation
CN103861222A (en) * 2014-02-24 2014-06-18 钟海平 Mixed early-warning cabinet fire extinguishing device and early-warning method thereof
CN104623831B (en) * 2014-12-31 2018-08-31 曙光节能技术(北京)股份有限公司 Intelligent fire extinguishing system and device for server cabinet
CN105046868B (en) * 2015-06-16 2019-09-20 苏州华启智能科技股份有限公司 A kind of fire alarm method based on thermal infrared imager in long and narrow environment
CN105337180A (en) * 2015-10-16 2016-02-17 常熟开关制造有限公司(原常熟开关厂) Power distribution cabinet and power distribution system
CN105261030B (en) * 2015-11-26 2019-01-15 四川汇源光通信有限公司 The method and device of flame is detected from infrared video
CN105788139A (en) * 2016-03-31 2016-07-20 嘉兴恒创电力设计研究院有限公司嘉善分公司 Intelligent fire-extinguishing distribution box
CN106231872B (en) * 2016-08-23 2019-04-16 江苏明强电气有限公司 A kind of arc protection is from control cabinet of putting out a fire

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101315667A (en) * 2008-07-04 2008-12-03 南京航空航天大学 Multi-characteristic synthetic recognition method for outdoor early fire disaster
CN101577033A (en) * 2009-05-26 2009-11-11 官洪运 Multiband infrared image-type fire detecting system and fire alarm system thereof
CN201603324U (en) * 2009-12-09 2010-10-13 上海中发电气(集团)股份有限公司;中发电气股份有限公司 An electrical switchgear sensing automatic spray system

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7728715B2 (en) * 1996-01-23 2010-06-01 En-Gauge, Inc. Remote monitoring

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101315667A (en) * 2008-07-04 2008-12-03 南京航空航天大学 Multi-characteristic synthetic recognition method for outdoor early fire disaster
CN101577033A (en) * 2009-05-26 2009-11-11 官洪运 Multiband infrared image-type fire detecting system and fire alarm system thereof
CN201603324U (en) * 2009-12-09 2010-10-13 上海中发电气(集团)股份有限公司;中发电气股份有限公司 An electrical switchgear sensing automatic spray system

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
野外早期火灾图像识别方法研究;徐仕玲等;《计算机技术与发展》;20080630;第18卷(第6期);摘要,第215页左侧栏第1.2节,右侧栏第2.1节,第216页左侧栏第2.4节和右侧栏第4节 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3321908A1 (en) * 2016-11-11 2018-05-16 Kidde Technologies, Inc. Fiber optic based monitoring of temperature and/or smoke conditions at electronic components
WO2018089477A1 (en) * 2016-11-11 2018-05-17 Carrier Corporation High sensitivity fiber optic based detection

Also Published As

Publication number Publication date
CN102743830A (en) 2012-10-24

Similar Documents

Publication Publication Date Title
Marbach et al. An image processing technique for fire detection in video images
DE10011411C2 (en) Imaging fire detector
US20150215685A1 (en) Methods, apparatus, and systems for monitoring transmission systems
US20060115154A1 (en) Fire detection and smoke detection method and system based on image processing
Honorton Error some place.
CN101388130B (en) Multi-grade stereo anti-intruding system, apparatus and implementing method based on wireless sensing net
US7456749B2 (en) Apparatus for detecting a fire by IR image processing
CN1331101C (en) Device and method for detecting fire sources or gas impurites
EP1275094A2 (en) Early fire detection method and apparatus
US6720880B2 (en) Vision-based method and apparatus for automatically activating a child safety feature
CN104299351B (en) Intelligent early-warning firefighting robot
CN101751755B (en) Forest Fire Warning System
Toreyin et al. Online detection of fire in video
US2942247A (en) Alarm warning system for swimming pools
US20030178568A1 (en) Fire detector with electronic frequency analysis
CN103400465A (en) Multifunctional and highly integrated fire safety precaution emergency system device
CN1639751A (en) System and method for territory thermal monitoring
Xu et al. Automatic fire smoke detection based on image visual features
Yu et al. A real-time video fire flame and smoke detection algorithm
AU2002220440B2 (en) Video smoke detection system
CN101334924B (en) Fire hazard probe system and its fire hazard detection method
PT1628260E (en) Method and system for automatic forest fire recognition
CN101577033A (en) Multiband infrared image-type fire detecting system and fire alarm system thereof
CA2275893A1 (en) Low false alarm rate video security system using object classification
EP1286727A2 (en) Method and device for the early detection of fire and for fighting fire indoors and outdoors, especially in living areas, of homes and buildings

Legal Events

Date Code Title Description
PB01 Publication
C06 Publication
SE01 Entry into force of request for substantive examination
C10 Entry into substantive examination
GR01 Patent grant
C14 Grant of patent or utility model