CN104599427B - A kind of intelligent image type fire alarm system for vcehicular tunnel - Google Patents
A kind of intelligent image type fire alarm system for vcehicular tunnel Download PDFInfo
- Publication number
- CN104599427B CN104599427B CN201410834924.1A CN201410834924A CN104599427B CN 104599427 B CN104599427 B CN 104599427B CN 201410834924 A CN201410834924 A CN 201410834924A CN 104599427 B CN104599427 B CN 104599427B
- Authority
- CN
- China
- Prior art keywords
- image
- fire
- unit
- flame
- neural network
- Prior art date
- Legal status (The legal status 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 status listed.)
- Active
Links
- 238000012544 monitoring process Methods 0.000 claims abstract description 28
- 238000001514 detection method Methods 0.000 claims abstract description 20
- 239000000779 smoke Substances 0.000 claims abstract description 20
- 230000005540 biological transmission Effects 0.000 claims abstract description 19
- 238000013528 artificial neural network Methods 0.000 claims description 25
- 238000000605 extraction Methods 0.000 claims description 21
- 238000006243 chemical reaction Methods 0.000 claims description 18
- 238000000034 method Methods 0.000 claims description 16
- 230000008569 process Effects 0.000 claims description 14
- 238000012549 training Methods 0.000 claims description 11
- 238000007781 pre-processing Methods 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 10
- 230000008859 change Effects 0.000 claims description 9
- 230000005855 radiation Effects 0.000 claims description 7
- 238000003062 neural network model Methods 0.000 claims description 6
- 239000000284 extract Substances 0.000 claims description 5
- 238000011478 gradient descent method Methods 0.000 claims description 4
- 238000003384 imaging method Methods 0.000 claims description 4
- 230000002708 enhancing effect Effects 0.000 claims description 3
- 238000001914 filtration Methods 0.000 claims description 3
- 238000003909 pattern recognition Methods 0.000 claims description 3
- 230000011218 segmentation Effects 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 3
- 239000000443 aerosol Substances 0.000 claims description 2
- 239000000203 mixture Substances 0.000 claims description 2
- 230000005693 optoelectronics Effects 0.000 claims description 2
- 230000000644 propagated effect Effects 0.000 claims description 2
- 241000208340 Araliaceae Species 0.000 claims 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 claims 1
- 235000003140 Panax quinquefolius Nutrition 0.000 claims 1
- 235000008434 ginseng Nutrition 0.000 claims 1
- 238000013519 translation Methods 0.000 claims 1
- 230000006870 function Effects 0.000 abstract description 3
- 238000010586 diagram Methods 0.000 description 4
- 238000010276 construction Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000009434 installation Methods 0.000 description 3
- 230000005622 photoelectricity Effects 0.000 description 3
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- UGFAIRIUMAVXCW-UHFFFAOYSA-N Carbon monoxide Chemical compound [O+]#[C-] UGFAIRIUMAVXCW-UHFFFAOYSA-N 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 235000019504 cigarettes Nutrition 0.000 description 1
- 230000009977 dual effect Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 239000002360 explosive Substances 0.000 description 1
- 238000004079 fireproofing Methods 0.000 description 1
- 239000003546 flue gas Substances 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 239000003595 mist Substances 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000004321 preservation Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 230000007480 spreading Effects 0.000 description 1
- 231100000331 toxic Toxicity 0.000 description 1
- 230000002588 toxic effect Effects 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B17/00—Fire alarms; Alarms responsive to explosion
- G08B17/12—Actuation by presence of radiation or particles, e.g. of infrared radiation or of ions
- G08B17/125—Actuation by presence of radiation or particles, e.g. of infrared radiation or of ions by using a video camera to detect fire or smoke
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Business, Economics & Management (AREA)
- Emergency Management (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Fire-Detection Mechanisms (AREA)
- Alarm Systems (AREA)
- Fire Alarms (AREA)
Abstract
The present invention relates to a kind of intelligent image type fire alarm system for vcehicular tunnel, which includes：For the camera of real-time collection site data, the input terminal of camera connection video image acquisition and transmission module, two output terminals connect video monitoring module and image-type fire detector respectively；Image-type fire detector connects fire alarm control unit and image fire alarming and managing module respectively；Image fire alarming and managing module receives the video monitoring information from video monitoring module and the fire information from image-type fire detector respectively, image fire alarming and managing module connection database；Fire alarm control unit connects image-type fire detector；Database is set in computer.The fire alarm system of the present invention can be carried out at the same time three kinds of smoke detection, flame detecting and video monitoring functions, timely and accurately judge fire alarm and provide alarm signal, substantially reduce rate of false alarm, the interference free performance of alarm system for lifting.
Description
Technical field
The present invention relates to technical field of fire detection, specifically, being related to a kind of for freeway tunnel
Intelligent image type fire alarm system.
Background technology
With the development that China's reform and opening-up is further goed deep into and economic construction is rapid, all trades and professions all achieve very big
Achievement.Particularly in recent years, construction industry is very prosperous, various buildings, which are such as emerged rapidly in large numbersBamboo shoots after a spring rain, rises abruptly in rapidly Chinese the earth.
Wherein the quantity of freeway tunnel is constantly increasing, but often the height of a house is high inside freeway tunnel, span is big, compares envelope
Close, escape and rescue it is relatively difficult, for such building once catch fire, fire spreading is rapid, and generation flue gas toxity is big, evacuating personnel it is difficult and
Fire attack is difficult, tends to cause very big economic loss and severe social influence, so tunnel fire proofing is one non-
Often important work the purpose is to which the fire occurred is eliminated early stage rudiment or development, is minimized loss.Fire
The selection of alarm detector must accomplish it is reliable, advanced and extremely low with highly sensitive and rate of false alarm, can it is gapless, at any time with
Ground, in the case of unmanned intervene, to tunnel space automatic fire detection-alarm.
Invention content
The object of the present invention is to provide a kind of intelligent fire detection alarm system that can be used in vcehicular tunnel, the present invention
According to the optical characteristics of video, three kinds of smoke detection, flame detecting, video monitoring functions can be realized simultaneously, and can be according to institute
The data monitored decide whether to send out fire alarm signal, integrated level height, during confirming burning things which may cause a fire disaster and realizing emergency reaction
The time of links can be greatly saved.
In order to reach foregoing invention purpose, technical solution provided by the invention is as follows：
A kind of intelligent image type fire alarm system for vcehicular tunnel, which is characterized in that the fire alarm system packet
It has included：
Field data is transmitted to and regards for acquiring the field data in vcehicular tunnel in real time by camera, the camera
Frequency Image Acquisition and transmission module；
Video image acquisition and transmission module, the input terminal connection camera, two output terminals connect regard respectively
Include sensor unit in frequency monitoring module and image-type fire detector, video image acquisition and transmission module, photoelectricity turns
Change unit and AD conversion unit；
Image-type fire detector, the image-type fire detector connect fire alarm control unit and image fire report respectively
Alert management module, the image-type fire detector are carried out at the same time smog image procossing and flame Image Processing to realize smoke detection
And flame detecting；
Video monitoring module, the output terminal connection image fire alarm management module of the video monitoring module simultaneously export real-time
Video monitoring information；
Image fire alarming and managing module receives video monitoring information from video monitoring module and from figure respectively
As the fire information of type fire detector, computer is shown in, image fire alarming and managing module connection database；
Fire alarm control unit, fire alarm control unit connection image-type fire detector；
Database, the database are set in computer, for the preservation of video and fire alarm information, in order to later right
The inquiry of historical information.
In the intelligent image type fire alarm system for being used for vcehicular tunnel in the present invention, the video image acquisition and biography
Defeated module acquires the radiation signal of scene of fire flame and smog by the imaging sensor being used as in sensor unit, and photoelectricity turns
It changes unit and radiation signal is converted into analog video signal, simultaneously boil down to is digital by analog video signal conversion for AD conversion unit
Image sequence.
In the intelligent image type fire alarm system for being used for vcehicular tunnel in the present invention, the image-type fire detector
It is equipped with smoke detection unit, flame detecting unit, image pre-processing unit, image characteristics extraction unit, intelligent recognition unit
With fire judgement unit, the data in smoke detection unit and flame detecting unit are respectively by image pre-processing unit, image
Feature extraction unit, intelligent recognition unit and fire judgement unit, the export structure linkage image fire report of fire judgement unit
Alert management module and fire alarm control unit.
In the intelligent image type fire alarm system for being used for vcehicular tunnel in the present invention, the image pre-processing unit packet
Filtering process, edge enhancing, binaryzation, segmentation and registration process are included, removal interference, noise level difference become original image
It is suitable for the form that computer carries out feature extraction.
In the intelligent image type fire alarm system for being used for vcehicular tunnel in the present invention, the image characteristics extraction unit
Including smoke characteristics extraction and flame characteristic extraction, smoke characteristics include spatial resolution, frequecy characteristic and directional characteristic,
Flame characteristic includes flame color, flicker frequency and brightness change.
In the present invention in the intelligent image type fire alarm system of vcehicular tunnel, the intelligent recognition unit to use
Artificial neural network is classified to image information and is recognized, and image information classification includes fire condition, alarm condition and just
Normal state.
The present invention for vcehicular tunnel intelligent image type fire alarm system in, the fire judgement unit according to
Image information classify and preset standard compare be confirmed whether occur fire, start if fire occurs fire alarm control unit into
Row alarm, and fire information and video information are shown in image fire alarming and managing module.
Based on above-mentioned technical proposal, have in the intelligent image type fire alarm system for vcehicular tunnel of the present invention
Following technological merit：
1. the characteristics of intelligent image type fire alarm system of the present invention is maximum is that it can be carried out at the same time smoke detection, fire
Flame detects, and integrated level is high, has greatly saved system cost, the complexity for reducing system installation maintenance.
2. the intelligent image type fire alarm system of the present invention is using advanced photoelectric imaging technology, Computer Image Processing
Technology has contactless detection, high intelligence, strong sensitive feature without by spatial altitude, air velocity, explosive, toxic etc.
The limitation of environmental condition, the smog that can be detected the flame of early stage in time and generate are that vcehicular tunnel carries out detection
Effective means.
3. the intelligent image type fire alarm system of the present invention is alarmed using visualization, detection and video monitoring are realized
Dual function.Monitoring center pops up live real-time video picture, and identify fire specific location immediately when fire occurs, and supplies
Operator on duty confirms, substantially reduces determining burning things which may cause a fire disaster and realizes the time of emergency reaction process, is drawn quickly to put out burning things which may cause a fire disaster in time
It rescues lives and properties and provides " time " guarantee.
4. the intelligent image type fire alarm system investigative range of the present invention is big, apart from construction wiring side remote, simple in structure
Just, easy-to-use handy compatibility is strong.
Description of the drawings
Fig. 1 is the system structure diagram of intelligent image type fire alarm system of the present invention.
Fig. 2 is the flow chart of data processing schematic diagram of intelligent image type fire alarm system of the present invention.
Fig. 3 is the video image acquisition of the present invention and transmission module flow chart of data processing schematic diagram.
Specific embodiment
We carry out the intelligent image type fire for vcehicular tunnel to the present invention with reference to attached drawing and specific embodiment below
Calamity alarm system is described in further detail, its operation principle and concrete application are expressly understood that, but not in the hope of apparent
It can be limited the scope of the invention with this.
As shown in Figure 1, Fig. 1 is the intelligent image type fire alarm system structural representation for vcehicular tunnel of the present invention
Figure, intelligent image type fire alarm system of the invention is mainly by camera 1, video image acquisition and transmission module 2, video
7 groups of monitoring module 3, image-type fire detector 4, image fire alarming and managing module 5, fire alarm control unit 6 and database
Into.
Wherein, camera 1 is installed in vcehicular tunnel, for acquiring the field data in vcehicular tunnel in real time, and will be existing
Field data is transmitted to video image acquisition and transmission module 2.Described in video image acquisition and the connection of the input terminal of transmission module 2
Two output terminals of camera 1, video image acquisition and transmission module 2 connect video monitoring module 3 and image-type fire respectively
Detector 4 includes sensor unit, photoelectric conversion unit and analog-to-digital conversion list in video image acquisition and transmission module 2
Member.
Above-mentioned image-type fire detector 4 connects fire alarm control unit 6 and image fire alarming and managing module respectively
5, which is carried out at the same time smog image procossing and flame Image Processing to realize that smoke detection and flame are visited
It surveys.The output terminal connection image fire alarm management module 5 of video monitoring module 3 simultaneously exports real-time video monitoring information, image
Fire alarm management module 5 receives video monitoring information from video monitoring module 3 and respectively from image-type detection
The fire information of device 4, the image fire alarming and managing module 5 connection database 7, the database 7 are set in computer, should
Database 7 is connected with the image fire alarming and managing module 5, for receiving the video of its transmission and fire alarm information
It preserves, in order to later to the inquiry of historical information.Fire alarm control unit 6 connects image-type fire detector 4.
Since the randomness that fire occurs is strong, the place that fire occurs can not be estimated, camera 1 is carried out in vcehicular tunnel
During installation, according to the size in space, structure, density, the position of the installation of camera 1 are considered, the visual angle of monitoring will avoid monitoring existing
There is dead angle and fail to report, when selection of camera considers its clarity, signal-to-noise ratio, outer synchronization and spectral response characteristic.
Fig. 2 is the flow chart of data processing schematic diagram of intelligent image type fire alarm system of the present invention.Pass through camera 1 first
The video monitoring image at scene is obtained, then obtains digital image sequence after video image acquisition and transmission module 2 are handled,
Image-type fire detector 4 receives video image acquisition and the image information of transmission module 2 pre-processes it, proposes
Smog, the visual signature information progress intellectual analysis of flame image and identification.
It can be with fire behavior point, by linkage image fire alarming and managing module 5, fire alarming and controlling if found after image identification
Device 6, by image fire alarming and managing module 5 by the fire parameters such as spatial positional information of fire occurrence point information and video monitoring
Information is shown on computer, and carries out burning things which may cause a fire disaster positioning, is automatically controlled nozzle and is put out a fire in real time accurately and in time, while by fire
Alarm controller 6 starts sound-light alarm and warning message is notified related personnel in the form of short message.
Wherein, video image acquisition and transmission module 2 include 3 big units as shown in Figure 3：Image sensor cell, photoelectricity
Converting unit, A/D converting units.It acquires the radiation of scene of fire flame and smog in real time by image sensor cell first
Signal, into analog signal vision signal, is finally forming digital picture sequence through opto-electronic conversion after A/D conversions and compressed encoding
Row.
The video image acquisition and transmission module acquires fire by the imaging sensor being used as in sensor unit
Radiation signal is converted to analog video signal, analog-to-digital conversion list by the radiation signal of live flame and smog, photoelectric conversion unit
Analog video signal is converted simultaneously boil down to digital image sequence by member.
In the intelligent image type fire alarm system for vcehicular tunnel of the present invention, described image type fire detector 4
Practical is image processing process, can be carried out at the same time smog image procossing, flame Image Processing, realize smoke detection and flame
Detection.Image-type fire detector is equipped with smoke detection unit, flame detecting unit, image pre-processing unit, characteristics of image
Extraction unit, intelligent recognition unit and fire judgement unit, the data in smoke detection unit and flame detecting unit pass through respectively
Cross image pre-processing unit, image characteristics extraction unit, intelligent recognition unit and fire judgement unit, fire judgement unit it is defeated
Go out structure linkage image fire alarming and managing module and fire alarm control unit.It needs to pre-process before image is handled
Including filtering process, edge enhancing, binaryzation, segmentation and registration etc., it is therefore an objective to eliminate certain interference, be conducive to improve image
The efficiency of analysis is handled, characteristics of image proposition is then carried out using the image processing techniques based on Wavelet transformation, then using people
The method of artificial neural networks carries out pattern recognition classifier according to characteristics of image, is then determined whether fire occurs by fire judgement unit
Calamity.
Above-mentioned image-type fire detector 4 is as follows：
Data pre-processing unit：Smog and Digital Flame Image of the image-type fire detector first to receiving
Ordered series of numbers smoothly and Edge contrast, eliminate or reduce influence of noise as possible, then carries out binary conversion treatment to digital picture, carries
The whole target image with local feature of reflection image is taken out, then removes the background of image using image subtraction operation.
Image characteristics extraction unit：Target image is passed through flexible peace by image-type fire detector 4 using wavelet transformation
After shifting, the smog that can represent and distinguish the target image, the various change characteristic information of flame are extracted respectively, specifically includes cigarette
Mist feature extraction and flame characteristic extraction, smoke characteristics include spatial resolution, frequecy characteristic and directional characteristic, flame characteristic
Include flame color, flicker frequency and brightness change.
Intelligent recognition unit：Intelligent recognition unit is classified and is recognized to image information using artificial neural network, figure
Classify as information and include fire condition, alarm condition and normal condition, specifically using BP networks mould in artificial neural network
Type carries out pattern recognition classifier, and the course of work can be divided into study and two stages of classification, the former is the sample number according to image
According to by network in itself according to certain rule progress self study；The latter is then to be classified using the structure of study to entire image.
Fire judgement unit：The fire judgement unit is compared according to image information classification and preset standard to confirm
No generation fire, if fire occurs starting fire alarm control unit alarms, and fire information and video information are shown
In image fire alarming and managing module.Its operation result is transferred to fire and sentenced by BP neural network model in intelligent recognition unit
Other unit judges fire condition by it according to 0 (k).
In intelligent recognition unit, BP neural network learning process is as follows：
A. training sample obtains first, and sample classification, Ran Houcong are carried out to target image according to M characteristic parameter of extraction
One is respectively randomly selected in M group target images, forms a M dimensional feature vector, extracts repeatedly composition training sample set；
B. BP neural network model is then built so that the input node of BP neural network corresponds to M spy of target image
Parameter is levied, for node i=M of BP neural network input layer, the output node of BP neural network is made to correspond to fire identification result 0
(k), the value of 0 (k) is 0 (k)[0,1], wherein, 0 (k)[0.75,1] it is " fire condition ", 0 (k)[0.25,0.75]
For " alarm condition ", 0 (k)[0,0.25] it is " normal condition ", number of nodes k=1 of corresponding BP neural network output layer;Then
Calculate the number of nodes of BP neural network hidden layer；
C. the training of neural network, neural network using three etale topology structures, i.e. input layer, hidden layer, output layer,
Input layer information is determined by the fire hazard aerosol fog that extracts, the both image change characteristics of flame.The training flow of neural network is as follows：
Parameter initialization：The parameters such as input layer setting maximum iteration, minimal error, learning rate are set, will be inputted
LayerWith hidden layerConnection weight, hidden layerWith output layerConnection weightIt is initialized as（-1,1）Between
Random number.
Information propagated forward process calculates, by training sampleThe input layer of BP neural network model is input to,
Pass through inputObtain the input value of hidden layer node, whereinFor input layerThe output valve of node, that is,The input value of node.So hidden layerOutput be；Then pass throughObtain output layerThe input of node
For；Finally obtain output layerThe output of node is。
Error back-propagating process calculates.Assuming that desired output is, then the mean error of output terminal be, the weights between hidden layer and output layer are then modulated using gradient descent method.Similarly, according to hidden
Containing the error on node layer, the weights between input layer and hidden layer are adjusted using gradient descent method, circulation step 2 and 3,
Until reaching maximum frequency of training or meeting minimal error.
The weights of each layer are preserved, and new samples are identified using the parameter succeeded in school.
Intelligent image type fire alarm system of the present invention is mainly used to monitor the flame and smog in vcehicular tunnel, can be to tunnel
Fire conditions carry out the real-time monitoring of non-intervention type and can decide whether to send out fire automatically according to the data monitored in road
Calamity pre-warning signal.By the present invention, accurately and reliably judge fire alarm and provide alarm signal, reduce rate of false alarm, alarm system for lifting
Antijamming capability.
Claims (5)
1. a kind of intelligent image type fire alarm system for vcehicular tunnel, which is characterized in that the fire alarm system includes
Have：
Field data is transmitted to video figure by camera, the camera for acquiring the field data in vcehicular tunnel in real time
As acquisition and transmission module；
Video image acquisition and transmission module, the input terminal connection camera, two output terminals connect video prison respectively
Module and image-type fire detector are controlled, includes sensor unit, opto-electronic conversion list in video image acquisition and transmission module
Member and AD conversion unit；
Image-type fire detector, the image-type fire detector connect fire alarm control unit and image fire alarm pipe respectively
Module is managed, which is carried out at the same time smog image procossing and flame Image Processing to realize smoke detection and fire
Flame detects；
Video monitoring module, the output terminal connection image fire alarm management module of the video monitoring module simultaneously export real-time video
Monitoring information；
Image fire alarming and managing module receives video monitoring information from video monitoring module and from image-type respectively
The fire information of fire detector, image fire alarming and managing module connection database；
Fire alarm control unit, fire alarm control unit connection image-type fire detector；
Database, the database are set in computer；
The video image acquisition and transmission module acquires scene of fire by the imaging sensor being used as in sensor unit
Radiation signal is converted to analog video signal by the radiation signal of flame and smog, photoelectric conversion unit, and AD conversion unit will
Analog video signal conversion and boil down to digital image sequence；
The image-type fire detector is equipped with smoke detection unit, flame detecting unit, image pre-processing unit, image
Feature extraction unit, intelligent recognition unit and fire judgement unit, the data point in smoke detection unit and flame detecting unit
Not Jing Guo image pre-processing unit, image characteristics extraction unit, intelligent recognition unit and fire judgement unit, fire judgement unit
Export structure linkage image fire alarming and managing module and fire alarm control unit；
The image pre-processing unit smog and Digital Flame Image ordered series of numbers that receive are carried out first smoothly at sharpening
Reason is eliminated or reduces influence of noise as possible, then carries out binary conversion treatment to digital picture, extracts reflection image entirety drawn game
The target image of portion's feature then removes the background of image using image subtraction operation；
Image characteristics extraction unit using wavelet transformation by target image by flexible and translation after, extracting can represent simultaneously respectively
The smog of the target image, the various change characteristic information of flame are distinguished, smoke characteristics extraction is specifically included and flame characteristic carries
It takes, smoke characteristics include spatial resolution, frequecy characteristic and directional characteristic, and flame characteristic includes flame color, flicker frequency
Rate and brightness change；
Intelligent recognition unit is classified and is recognized to image information using artificial neural network, and image information classification includes fire
Calamity state, alarm condition and normal condition specifically carry out pattern recognition classifier using BP network models in artificial neural network,
Its course of work can be divided into study and two stages of classification, the former is the sample data according to image, by network in itself according to certain
Kind rule carries out self study；The latter is then to be classified using the structure of study to entire image；
Fire judgement unit is compared according to image information classification and preset standard to be confirmed whether that fire occurs, if fire occurs
Start fire alarm control unit to alarm, and fire information and video information are shown in image fire alarming and managing module
In；Its operation result is transferred to fire judgement unit by BP neural network model in intelligent recognition unit, is sentenced by it according to 0 (k)
Disconnected fire condition；
In intelligent recognition unit, BP neural network learning process is as follows：
A. training sample obtains first, sample classification is carried out to target image according to M characteristic parameter of extraction, then from M groups
One is respectively randomly selected in target image, forms a M dimensional feature vector, extracts repeatedly composition training sample set；
B. BP neural network model is then built so that the input node of BP neural network corresponds to the M feature ginseng of target image
Number for node i=M of BP neural network input layer, makes the output node of BP neural network correspond to fire identification result 0 (k), and 0
(k) value is 0 (k)[0,1], wherein, 0 (k)[0.75,1] it is " fire condition ", 0 (k)[0.25,0.75] is " alert
Lodge a complaint with state ", 0 (k)[0,0.25] it is " normal condition ", number of nodes k=1 of corresponding BP neural network output layer;Then it calculates
The number of nodes of BP neural network hidden layer；
C. the training of neural network, neural network is using three etale topology structures, i.e. input layer, hidden layer, output layer, input
Layer information determines that the training flow of neural network is as follows by the fire hazard aerosol fog that extracts, the both image change characteristics of flame：
Parameter initialization：Input layer setting maximum iteration, minimal error, learning rate parameter are set, by input layerWith it is hidden
Containing layerConnection weight, hidden layerWith output layerConnection weightIt is initialized as（-1,1）Between random number；
Information propagated forward process calculates, by training sampleThe input layer of BP neural network model is input to, by defeated
EnterObtain the input value of hidden layer node, whereinFor input layerThe output valve of node, that is,Node
Input value, then hidden layerOutput be；Then pass throughObtain output layerThe input of node
For；Finally obtain output layerThe output of node is；
Error back-propagating process calculates；Assuming that desired output is, then the mean error of output terminal be
, the weights between hidden layer and output layer are then modulated using gradient descent method, similarly, according to the mistake on hidden layer node
Difference adjusts the weights between input layer and hidden layer using gradient descent method, circulation step 2 and 3, until reaching maximum instruction
Practice number or meet minimal error；The weights of each layer are preserved, and new samples are identified using the parameter succeeded in school.
2. a kind of intelligent image type fire alarm system for vcehicular tunnel according to claim 1, which is characterized in that
The image pre-processing unit includes filtering process, edge enhancing, binaryzation, segmentation and registration process, removal interference, noise
Original image is become suitable for the form of computer progress feature extraction by grade difference.
3. a kind of intelligent image type fire alarm system for vcehicular tunnel according to claim 1, which is characterized in that
The image characteristics extraction unit includes smoke characteristics extraction and flame characteristic extraction, and smoke characteristics include spatial discrimination
Rate, frequecy characteristic and directional characteristic, flame characteristic include flame color, flicker frequency and brightness change.
4. a kind of intelligent image type fire alarm system for vcehicular tunnel according to claim 1, which is characterized in that
The intelligent recognition unit uses artificial neural network, is classified to image information and is recognized, and image information classification includes
There are the non-fire of the first kind and the second class fire.
5. a kind of intelligent image type fire alarm system for vcehicular tunnel according to claim 1, which is characterized in that
The fire judgement unit is compared according to image information classification and preset standard to be confirmed whether to occur fire, and fire such as occurs
Then start fire alarm control unit to alarm, and fire information and video information are shown in image fire alarming and managing module
In.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410834924.1A CN104599427B (en) | 2014-12-30 | 2014-12-30 | A kind of intelligent image type fire alarm system for vcehicular tunnel |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410834924.1A CN104599427B (en) | 2014-12-30 | 2014-12-30 | A kind of intelligent image type fire alarm system for vcehicular tunnel |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104599427A CN104599427A (en) | 2015-05-06 |
| CN104599427B true CN104599427B (en) | 2018-06-26 |
Family
ID=53125176
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410834924.1A Active CN104599427B (en) | 2014-12-30 | 2014-12-30 | A kind of intelligent image type fire alarm system for vcehicular tunnel |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104599427B (en) |
Families Citing this family (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106875611A (en) * | 2015-12-12 | 2017-06-20 | 上海腾盛智能安全科技股份有限公司 | Traffic tunnel fire detecting system |
| CN106846699A (en) * | 2017-03-10 | 2017-06-13 | 深圳实现创新科技有限公司 | The accident localization method and system of urban fire control system |
| CN108310700A (en) * | 2018-01-30 | 2018-07-24 | 国网上海市电力公司 | A kind of substation cable layer fire extinguishing system based on image fire detection device |
| CN108682105B (en) * | 2018-05-29 | 2019-11-05 | 贵州电网有限责任公司 | One kind is based on multispectral transmission line forest fire exploration prior-warning device and method for early warning |
| CN108564760A (en) * | 2018-06-06 | 2018-09-21 | 广西防城港核电有限公司 | Fire detection device under nuclear power station extreme environmental conditions and detection method |
| CN109920200B (en) * | 2018-08-01 | 2021-06-18 | 赫威环境科技(安徽)有限公司 | Automatic humidifying heating pipe |
| CN110314314A (en) * | 2019-05-27 | 2019-10-11 | 安徽中科中涣防务装备技术有限公司 | A kind of gas station's Intelligent preventive control apparatus and system |
| CN111242053B (en) * | 2020-01-16 | 2023-07-18 | 国网山西省电力公司电力科学研究院 | Power transmission line flame detection method and system |
| CN112043991B (en) * | 2020-09-15 | 2023-06-20 | 河北工业大学 | Tunnel guide rail traveling fire-fighting robot system and use method |
| CN112767644B (en) * | 2020-12-31 | 2022-05-24 | 千方捷通科技股份有限公司 | Method and device for early warning fire in highway tunnel based on video identification |
| CN112927411A (en) * | 2021-01-21 | 2021-06-08 | 深圳摩萨迪智能科技有限公司 | Access control management method based on image processing technology |
| CN113593171B (en) * | 2021-02-27 | 2024-01-12 | 国家电网有限公司 | Smoke sensing early warning method and system for cable tunnel space |
| CN113409535A (en) * | 2021-04-29 | 2021-09-17 | 辛米尔视觉科技(上海)有限公司 | Tunnel flame detection system based on computer vision and artificial intelligence algorithm |
| CN113379998A (en) * | 2021-06-09 | 2021-09-10 | 南京品傲光电科技有限公司 | Automatic fire alarm system in petrochemical tank district |
| CN115641688A (en) * | 2022-11-03 | 2023-01-24 | 中铁第一勘察设计院集团有限公司 | Fire extremely early detection early warning system and method |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2001067566A (en) * | 1999-08-30 | 2001-03-16 | Fujitsu Ltd | Fire detecting device |
| CN101334924A (en) * | 2007-06-29 | 2008-12-31 | 丁国锋 | Fire hazard probe system and its fire hazard detection method |
| CN101527073A (en) * | 2009-02-17 | 2009-09-09 | 丁国锋 | Fire detection system and method thereof |
-
2014
- 2014-12-30 CN CN201410834924.1A patent/CN104599427B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2001067566A (en) * | 1999-08-30 | 2001-03-16 | Fujitsu Ltd | Fire detecting device |
| CN101334924A (en) * | 2007-06-29 | 2008-12-31 | 丁国锋 | Fire hazard probe system and its fire hazard detection method |
| CN101527073A (en) * | 2009-02-17 | 2009-09-09 | 丁国锋 | Fire detection system and method thereof |
Non-Patent Citations (2)
| Title |
|---|
| 基于图像处理的火灾智能监视识别技术的研究;鲁维;《中国优秀硕士学位论文全文数据库信息科技辑》;20091215;摘要，第一至第四章，图4.1 * |
| 基于视频的高速公路隧道火灾检测技术研究;杨伟松;《中国优秀硕士学位论文全文数据库信息科技辑》;20131215;摘要，第四章 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104599427A (en) | 2015-05-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104599427B (en) | A kind of intelligent image type fire alarm system for vcehicular tunnel | |
| CN104966372B (en) | The forest fire intelligent identifying system and method for multi-data fusion | |
| CN109637068A (en) | Intelligent pyrotechnics identifying system | |
| CN104143248B (en) | Forest fire detection based on unmanned plane and preventing control method | |
| CN100504942C (en) | Module set of intelligent video monitoring device, system and monitoring method | |
| CN106600888B (en) | Automatic forest fire detection method and system | |
| CN108389359B (en) | Deep learning-based urban fire alarm method | |
| CN108052900A (en) | A kind of method by monitor video automatic decision dressing specification | |
| CN107437318B (en) | Visible light intelligent recognition algorithm | |
| US20090195382A1 (en) | Video sensor and alarm system and method with object and event classification | |
| CN109484935A (en) | A kind of lift car monitoring method, apparatus and system | |
| KR101853903B1 (en) | Invasion detecting method and system including CCTV, drone and fiber to the fence | |
| CN103839373A (en) | Sudden abnormal event intelligent identification alarm device and system | |
| CN107688793A (en) | A kind of outside transformer substation fire automatic monitoring method for early warning | |
| CN111899452A (en) | Forest fire prevention early warning system based on edge calculation | |
| CN111985365A (en) | Straw burning monitoring method and system based on target detection technology | |
| KR101204259B1 (en) | A method for detecting fire or smoke | |
| CN109218667B (en) | Public place safety early warning system and method | |
| KR102162735B1 (en) | Fire mornitoring method and system using cctv image recognition based on deep learning | |
| CN111223263A (en) | Full-automatic comprehensive fire early warning response system | |
| CN103324910A (en) | Fire alarming method based on video detection | |
| Yang et al. | Early detection of forest fire based on unmaned aerial vehicle platform | |
| CN109544870A (en) | Alarm decision method and intelligent monitor system for intelligent monitor system | |
| KR20200052418A (en) | Automated Violence Detecting System based on Deep Learning | |
| CN116434533A (en) | AI wisdom highway tunnel synthesizes monitoring platform based on 5G |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: 214028 Jiangsu province Wuxi city Wuxi District sensor network university science and Technology Park Qingyuan Road career building area C layer Applicant after: WUXI BOHUI PHOTOELECTRONIC TECHNOLOGY CO.,LTD. Applicant after: SHANGHAI BANDWEAVER TECHNOLOGY Co.,Ltd. Address before: 214028 Jiangsu province Wuxi city Wuxi District sensor network university science and Technology Park Qingyuan Road career building area C layer Applicant before: WUXI BOHUI PHOTOELECTRONIC TECHNOLOGY CO.,LTD. Applicant before: Shanghai Bandweaver Technologies Co.,Ltd. |
|
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20171023 Address after: 200120 Shanghai city Pudong New Area Branch Road No. 88 west five building room 503 Applicant after: SHANGHAI BANDWEAVER TECHNOLOGY Co.,Ltd. Address before: 214028 Jiangsu province Wuxi city Wuxi District sensor network university science and Technology Park Qingyuan Road career building area C layer Applicant before: WUXI BOHUI PHOTOELECTRONIC TECHNOLOGY CO.,LTD. Applicant before: SHANGHAI BANDWEAVER TECHNOLOGY Co.,Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP03 | Change of name, title or address | ||
| CP03 | Change of name, title or address |
Address after: Room 103, Building 299, Zhongchen Road, Songjiang District, Shanghai, 20113 Patentee after: Shanghai Bohui Technology Co.,Ltd. Address before: Room 503, 5th floor, West Tower, 88 Shangke Road, Pudong New Area, Shanghai, 200120 Patentee before: SHANGHAI BANDWEAVER TECHNOLOGY Co.,Ltd. |