CN209433517U - It is a kind of based on more flame images and the fire identification warning device for combining criterion - Google Patents
It is a kind of based on more flame images and the fire identification warning device for combining criterion Download PDFInfo
- Publication number
- CN209433517U CN209433517U CN201821763299.6U CN201821763299U CN209433517U CN 209433517 U CN209433517 U CN 209433517U CN 201821763299 U CN201821763299 U CN 201821763299U CN 209433517 U CN209433517 U CN 209433517U
- Authority
- CN
- China
- Prior art keywords
- sensor
- fire
- image
- unit
- criterion
- 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.)
- Expired - Fee Related
Links
Landscapes
- Fire-Detection Mechanisms (AREA)
- Alarm Systems (AREA)
Abstract
The utility model discloses a kind of based on more flame images and the fire identification warning device for combining criterion, including image acquisition units, conventional fire detector cells, image-signal processing system (image layer RBF recognition unit), control unit MCU, central control system (global layer RBF recognition unit), alarm output unit.The utility model is judged using more flame images with criterion is combined, and conventional fire detector is carried out criterion with image-type fire detector and is merged.The feature of doubtful fire image is identified using RBF neural in image layer, then output and the output of conventional fire sensor are judged together, global RBF neural is carried out to be identified, avoid that conventional fire detector sensitivity is not high, be not suitable for large space environment and image-type fire detector criterion is single, the excessively high disadvantage of wrong report rate of failing to report.
Description
Technical field
The utility model belongs to fire fire-fighting security technology area, and in particular to one kind is sentenced based on more flame images with combining
According to fire identification warning device.
Background technique
When fire occurs, if it is possible to find and alarm in time, loss can be reduced to the greatest extent.Conventional fire is visited
Surveying device mainly has temperature sensitive type, sense cigarette type, photosensitive type, gaseous type and compound etc., and investigative range is restricted by space and height, when
Monitoring area occurs not making a response at once when fire, could respond when only probe value reaches a certain level.So passing
The use of the fire detector of system is limited by condition, is suitable for the identification of small space fire, in the complex environment of large space
In its accuracy be difficult to reach requirement.
Utility model content
The purpose of the utility model is to provide a kind of based on more flame images and the fire identification warning device for combining criterion,
It solves conventional fire Detection Techniques existing in the prior art and is not suitable for large space, and image-type fire detection technology criterion
It is single, it is easy to appear wrong report and fails to report the problem of misrepresenting deliberately.
The utility model is the technical scheme adopted is that a kind of based on more flame images and the fire identification report for combining criterion
Alarm device, including several video cameras, video camera are connected with video frequency collection card unit, and video frequency collection card unit connects at picture signal
Manage unit;It further include sensor a, sensor b, sensor c, sensor a, sensor b, sensor c are connected with as input terminal
Control unit MCU;Image signal processing unit connects central control system unit as input terminal with control unit MCU;Center
The output end of control system unit is connected with voice guard and fire alarm lamp;
The utility model is also characterized by:
Wherein the video camera includes several infrared C CD video cameras and two kinds of video cameras of common CCD camera;
Wherein the sensor a is temperature sensor, using NTC thermistor sensor;Sensor b is gas sensing
Device mainly includes MQ2, MQ7;Sensor c is flame sensor, by ultraviolet light sensing tube R2868 and support circuit plate C3704 group
At;
Wherein described image signal processing unit (3) and central control system unit (8) are that RBF neural identification is single
Member.
Detailed description of the invention
Fig. 1 is a kind of knot of fire identification warning device based on the more criterions of the double-deck RBF neural of the utility model
Structure schematic diagram;
In figure, 1. camera units, 2. video frequency collection card units, 3. image signal processing units, 4. sensor a, 5. are passed
Sensor b, 6. sensor c, 7. control unit MCU, 8. central control system units, 9. voice guards, 10. fire alarm lamps.
Specific embodiment
The utility model discloses a kind of based on more flame images and the fire identification warning device for combining criterion, such as Fig. 1
Shown: including several video cameras 1, video camera 1 includes several infrared C CD video cameras and two kinds of common CCD camera camera shootings
Machine, video camera 1 are connected with video frequency collection card unit 2;2 signal of video frequency collection card unit is exported to image signal processing unit 3;Figure
As signal processing unit 3 outputs signal to sensor a4 respectively, sensor a4 is temperature sensor, sensor b5, sensor b5
For gas sensor, sensor c6, sensor c6 is flame sensor: sensor a4, sensor b5, sensor c6 are as input
End is connected with control unit MCU7;Flame sensor c6 is made of ultraviolet light sensing tube R2868 and support circuit plate C3704;Gas
Sensor b5 mainly includes MQ2, MQ7;Temperature sensor a4 uses NTC thermistor sensor.3 He of image signal processing unit
Control unit MCU7 outputs signals to central control system unit 8;Central control system unit 8 is connected with 9 He of voice guard
Fire alarm lamp 10.
The utility model it is a kind of based on more flame images and main component in the fire identification warning device for combining criterion
Effect and working principle difference it is as follows:
Video camera 1 samples according to certain frequency by video monitor object region and obtains picture frame, and video camera 1 uses
Common CCD camera combination infrared CCD camera carries out scene monitoring, and the optical signal into camera lens is converted into video telecommunication
Number, by intelligent acess video server, is realized and be connected to the network using IP network and monitoring center.
Video frequency collection card unit 2 may be implemented to carry out inspection according to certain rule simultaneously to multiple target implementing monitorings.
By video frequency collection card unit, the real time video signals from different monitoring point can be received, be stored.
The result that image signal processing unit 3 (image layer RBF recognition unit) exports image acquisition units prejudges
Processing.The image obtained to sampling is compared with benchmark image, by calculating sample frame and the direct association relationship of reference frame
Variation, judge whether there is fire in advance.Then doubtful image is split and feature extraction, feature input figure will be extracted
As layer RBF neural, output is as global neural network input.
Control unit MCU7 can be realized by 8-32 bit microprocessor chip, and microprocessor chip used must have calmly
When device unit and I/O mouthfuls of user, the work master clock of microprocessor should be in 24MHz or more.Used chip such as MCS51 series,
PIC series, AVR series and ARM series have can satisfy compared with multi-product, specially MCS51 family chip, can expire completely
The demand of sufficient the utility model.
Central control system unit 8 (global layer RBF recognition unit) senses temperature sensor, gas sensor, flame
Device and the output of image layer RBF neural are combined identification judgement, if result meets criterion and is more than threshold value, drive report
Alert cell operation.
Alarm unit mainly includes that voice guard 9 and fire alarm lamp 10 form.When centralized control unit sends light
When alarm command or audible alarm instruct, fire alarm lamp 10 and voice guard 9 are responded, and fire behavior flashing light report is respectively driven
Alert and fire behavior alarm sound.
In a kind of fire identification warning device based on the more criterions of the double-deck RBF neural of the utility model, RBF
(Radial Basis Function) neural network is also referred to as radial base neural net, is a kind of with stronger input and output
The optimal network of mapping function, pace of learning, classification capacity, in terms of have obviously advantage, in mode
It has a very wide range of applications in function identification field, especially in terms of fire identification.This algorithm utilizes image procossing skill
Art extracts the characteristic information of fire image, using RBF neural as carrier, carries out the final identification judgement of fire image.
RBF neural is mainly made of input layer, hidden layer and output layer, and input node only transmits input signal to hidden
Layer.Hidden node is made of (such as Gaussian function) radial basic function, and the action function in hidden node has partial approximation energy
Power is locally generating response to input signal.Output node layer generally uses linear function, for given input vector x,
The unit of RBF network output layer exports are as follows:
It is exported inputs as global network
R in formulai(x) it is exported for hidden layer, wikFor network weight, ciFor the center vector of basic function, m is hidden layer node
Number, p are output node number, | | x-ci| | indicate x to ciThe distance between. Ri(x) in ciThere is unique maximum value at place, | | x-ci||
Increase can make Ri(x) decaying is generated.From formula (1) and formula (2) as can be seen that input layer is realized from x to Ri(x) non-linear reflects
It penetrates, output layer is realized from Ri(x) y is arrivedk(x) Nonlinear Mapping, the composition of hidden layer are one group of radial basis function, pass through radial direction
Nonlinear Mapping relationship may be implemented in basic function.Parameter vector relevant to each hidden layer node is ci(i.e. center) and σi(i.e.
Width).The learning process of RBF network is broadly divided into three phases:
(1) the center c of each RBF unit is determinedi: general ciIt is to be determined by k- mean cluster analysis technology, selection tool
Representational data are as RBF unit center, it is possible to reduce hidden layer RBF number of unit reduces network and complicates degree.
(2) radius sigma of each RBF unit is determinedi: radius sigmaiThe size for determining RBF unit acceptance region, to network essence
The influence of degree is very big.
(3) adjust weight matrix w: w refers to the weight between hidden layer and output layer here, and adjusting weight matrix can be using ladder
Degree method.
Central control system unit 8 (global layer RBF recognition unit) in the utility model passes temperature sensor, gas
Sensor, flame sensor and the output of image layer RBF neural carry out identification judgement, and design output neuron number is 1, value
0 or 1.0 indicates do not have fire, and 1 indicates fire.Its result drives alarm unit work.
RBF neural main feature value has fire angle number, area change rate, consistency, stretches in the utility model
Length, center of mass point offset distance will extract characteristic value input picture layer RBF neural;Due to 5 feature calculations of input layer
Value difference is not too big, in order to be in input data in [0,1] section, needs to be normalized.The advantages of processing, is in this way
Making RBF neural to treated, data are easier study and training, and input node only transmits input signal to hidden layer.Hidden layer
Node is made of (such as Gaussian function) radial basic function, and the action function in hidden node has partial approximation ability, to defeated
Enter signal and locally generates response.Initial weight assignment is given by random function, calculates the error of target output and reality output;It is logical
Least square method is crossed with the linear weight value of learning rate appropriate adjustment output layer, network then terminates to instruct when reaching defined precision
Practice.In prediction, if the gap with actual result has been more than scheduled boundary, network meeting adjust automatically, this prediction model
It can adapt to the needs of fire variability.
The utility model is a kind of based on more flame images and the fire identification warning device for combining criterion, and working principle is such as
Under: the sampled images of monitoring area are obtained by 1 infrared CCD camera of video camera and video frequency collection card unit 2, calculate sample graph
As the similarity measure with benchmark image or background image, done when the similarity of consecutive numbers frame sampling image is higher than given threshold
Further judgement.Doubtful fire sampled images are pre-processed first, are divided, the processing such as feature extraction, are then joined feature
Number inputs in trained image-signal processing system unit 3 (image layer RBF neural) in advance, but at this time at picture signal
The output of reason system unit 3 can not entirely accurate judge whether to be fire.By sensor a4, sensor b5, pass
Analog-to-digital conversion parallel connection input control unit MCU7 is passed through in the output of sensor c6, then passes through the signal output of control unit MCU7
Universal serial bus feeds back to central control system unit 8, and central control system unit 8 is single by image signal processing unit 3 and control
Normalized is done in the input of first MCU7, inputs trained RBF neural, is judged whether by the output of global neural network
For fire generation, fire alarm lamp and voice guard is then driven to be responded in case of fire.
Bright the utility model advantage from the principle:
The utility model is based at digital picture with the fire identification warning device for combining criterion based on more flame images
The fire detection technology of reason intercepts doubtful fire image from monitor video and is analyzed, then extracts the feature ginseng of suspicious region
Number is simultaneously compared with presetting characteristic threshold value, so that judging whether is that fire occurs.Multi-layer image type fire detection technology
Can be with effective solution large space fire safety problem, but since large space fire image background is complicated, flame region is difficult to essence
Quasi- segmentation, characteristic criterion is single, so there is the wrong report higher problem of rate of failing to report.By conventional fire Detection Techniques and image-type fire
Calamity Detection Techniques carry out more criterion combinations, then establish fire identification model using RBF neural, the multilayer graph that will be extracted
As criterion feature carries out Classification and Identification to fire image as input quantity.By a series of fire tests the result shows that, to difference
The fire identification of scene accuracy rate with higher, can be effectively reduced fire false alarm rate, improve the accuracy of fire alarm;
On the other hand it is defeated RBF neural to be combined as with conventional fire Detection Techniques using the fire detection of multi-layer image type
Enter, overcomes conventional fire detector and be not suitable for that large space and image-type fire detection technology criterion are single, it is high to fail to report wrong report
The disadvantages of, whole system does not need manually to carry out relevant operation intervention, fire alarm fast response time, with higher accurate
Property.
Claims (4)
1. a kind of based on more flame images and the fire identification warning device for combining criterion, which is characterized in that including several camera shootings
Machine (1), video camera (1) are connected with video frequency collection card unit (2), and video frequency collection card unit (2) connects image signal processing unit
(3);It further include sensor a (4), sensor b (5), sensor c (6), sensor a (4), sensor b (5), sensor c (6) make
Control unit MCU (7) are connected with for input terminal;Image signal processing unit (3) and control unit MCU (7) connect as input terminal
Connect central control system unit (8);The output end of central control system unit (8) is connected with voice guard (9) and fire report
Warning lamp (10).
2. it is according to claim 1 a kind of based on more flame images and the fire identification warning device for combining criterion, it is special
Sign is that video camera (1) includes several infrared C CD video cameras and two kinds of video cameras of common CCD camera.
3. it is according to claim 1 a kind of based on more flame images and the fire identification warning device for combining criterion, it is special
Sign is that sensor a (4) is temperature sensor, using NTC thermistor sensor;Sensor b (5) is gas sensor, main
It to include MQ2 and MQ7;Sensor c (6) is flame sensor, is made of ultraviolet light sensing tube R2868 and support circuit plate C3704.
4. it is according to claim 1 a kind of based on more flame images and the fire identification warning device for combining criterion, it is special
Sign is that described image signal processing unit (3) and central control system unit (8) are RBF neural recognition unit.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201821763299.6U CN209433517U (en) | 2018-10-29 | 2018-10-29 | It is a kind of based on more flame images and the fire identification warning device for combining criterion |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201821763299.6U CN209433517U (en) | 2018-10-29 | 2018-10-29 | It is a kind of based on more flame images and the fire identification warning device for combining criterion |
Publications (1)
Publication Number | Publication Date |
---|---|
CN209433517U true CN209433517U (en) | 2019-09-24 |
Family
ID=67967999
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201821763299.6U Expired - Fee Related CN209433517U (en) | 2018-10-29 | 2018-10-29 | It is a kind of based on more flame images and the fire identification warning device for combining criterion |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN209433517U (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112327652A (en) * | 2020-11-13 | 2021-02-05 | 杭州慧光健康科技有限公司 | Household-old-age-care-oriented intelligent kitchen monitoring system and method |
CN113378804A (en) * | 2021-08-12 | 2021-09-10 | 中国科学院深圳先进技术研究院 | Self-service sampling detection method and device, terminal equipment and storage medium |
CN117011993A (en) * | 2023-09-28 | 2023-11-07 | 电子科技大学 | Comprehensive pipe rack fire safety early warning method based on image processing |
-
2018
- 2018-10-29 CN CN201821763299.6U patent/CN209433517U/en not_active Expired - Fee Related
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112327652A (en) * | 2020-11-13 | 2021-02-05 | 杭州慧光健康科技有限公司 | Household-old-age-care-oriented intelligent kitchen monitoring system and method |
CN113378804A (en) * | 2021-08-12 | 2021-09-10 | 中国科学院深圳先进技术研究院 | Self-service sampling detection method and device, terminal equipment and storage medium |
CN117011993A (en) * | 2023-09-28 | 2023-11-07 | 电子科技大学 | Comprehensive pipe rack fire safety early warning method based on image processing |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111739250B (en) | Fire detection method and system combining image processing technology and infrared sensor | |
CN209433517U (en) | It is a kind of based on more flame images and the fire identification warning device for combining criterion | |
CN101334924B (en) | Fire hazard probe system and its fire hazard detection method | |
CN110516522B (en) | Inspection method and system | |
CN106251568A (en) | A kind of fire alarm safety-protection system based on ultraviolet and image detecting technique and method | |
KR102195706B1 (en) | Method and Apparatus for Detecting Intruder | |
CN108319926A (en) | A kind of the safety cap wearing detecting system and detection method of building-site | |
US20090195382A1 (en) | Video sensor and alarm system and method with object and event classification | |
JP2012518846A (en) | System and method for predicting abnormal behavior | |
CN103839373A (en) | Sudden abnormal event intelligent identification alarm device and system | |
CN103324919B (en) | Video monitoring system and data processing method thereof based on recognition of face | |
CN111986436B (en) | Comprehensive flame detection method based on ultraviolet and deep neural networks | |
KR102149832B1 (en) | Automated Violence Detecting System based on Deep Learning | |
KR102043922B1 (en) | A cctv divided management system and method it | |
WO2018232846A1 (en) | Large-scale peripheral security monitoring method and system | |
CN111325954B (en) | Personnel loss early warning method, device, system and server | |
CN114724330A (en) | Implementation method of self-adaptive mode switching multi-channel video fire real-time alarm system | |
CN113160513A (en) | Flame detection device based on multisensor | |
CN109544870A (en) | Alarm decision method and intelligent monitor system for intelligent monitor system | |
CN201091014Y (en) | Fire detecting device | |
KR102233679B1 (en) | Apparatus and method for detecting invader and fire for energy storage system | |
CN112263803A (en) | Unmanned vehicle intelligent security system based on real-time scene inspection and automatic detection fire extinguishing and control method | |
KR101989376B1 (en) | Integrated track circuit total monitoring system | |
CN212569977U (en) | Fire identification alarm device based on grey and RBF double-layer neural networks | |
KR20190136515A (en) | Vehicle recognition device |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20190924 Termination date: 20201029 |
|
CF01 | Termination of patent right due to non-payment of annual fee |