CN1092825C - Method for detecting and positioning fire by using colour image three-primary colors difference - Google Patents

Method for detecting and positioning fire by using colour image three-primary colors difference Download PDF

Info

Publication number
CN1092825C
CN1092825C CN 95100519 CN95100519A CN1092825C CN 1092825 C CN1092825 C CN 1092825C CN 95100519 CN95100519 CN 95100519 CN 95100519 A CN95100519 A CN 95100519A CN 1092825 C CN1092825 C CN 1092825C
Authority
CN
China
Prior art keywords
image
fire
difference
red
blue
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
Application number
CN 95100519
Other languages
Chinese (zh)
Other versions
CN1112702A (en
Inventor
袁宏永
范维澄
王清安
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
University of Science and Technology of China USTC
Original Assignee
University of Science and Technology of China USTC
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 University of Science and Technology of China USTC filed Critical University of Science and Technology of China USTC
Priority to CN 95100519 priority Critical patent/CN1092825C/en
Publication of CN1112702A publication Critical patent/CN1112702A/en
Application granted granted Critical
Publication of CN1092825C publication Critical patent/CN1092825C/en
Anticipated expiration legal-status Critical
Expired - Fee Related legal-status Critical Current

Links

Images

Classifications

    • GPHYSICS
    • G08SIGNALLING
    • G08BSIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
    • G08B17/00Fire alarms; Alarms responsive to explosion
    • G08B17/12Actuation by presence of radiation or particles, e.g. of infrared radiation or of ions
    • G08B17/125Actuation 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)
  • Closed-Circuit Television Systems (AREA)
  • Alarm Systems (AREA)

Abstract

The present invention relates to a fire monitoring technique which is a fire difference analyzing technique, and the fire monitoring technique adopts a color CCD image pick-up system. In the fire monitoring technique, the difference calculation, the binaryzation and the second difference calculation of red and blue primary color components are carried out by the different reflections of color image three primary colors on early flame, the abnormality of fire is judged by the growth rate of a flame area difference image, and the space positioning process of the fire is carried out by a direct linear transformation technique. The fire monitoring technique enhances the reliability of the detection of the fire and can use the existing closed-circuit television system, and the automatic monitoring process, the positioning process and the linkage remedying process of the fire are carried out by a computer.

Description

Utilize colour image three-primary colors difference to carry out the method for detection and location
Technical field:
The present invention relates to the fire monitoring technical field.
Background technology:
Three types of the existing main thoughts cigarette of fire monitoring device, temperature-sensitive and sense flames.But temperature rising sometimes is not because the fire burning causes, but because due to weather or the normal operation release heat; Simple smog identification also lacks reliability in many occasions, because factors such as the size of the color of smog, soot, environment dust, spatial altitude, air-flow all easily cause the mistake that identification is judged, thereby causes newspaper or wrong report late; With sensor induction infrared, ultraviolet light is judged that also there is similar defective in fire, because be not to have only fire just to produce infrared, ultraviolet light, freezing etc. as electric arc welding, toluene and camera lens also usually causes the wrong report of ultraviolet light transducer or newspaper late; Infrared light transducer[sensor can only respond the flame with blinking characteristics, and is then comparatively difficult to the detection of gases at high pressure flame; The blackbody hot sources of being modulated also easily causes the wrong report of infrared sensor.Chinese patent application 93214042.0 proposes to utilize B to add special filtering apparatus and carries out fire monitoring, partly solved said method rate of false alarm problem of higher, but black-and-white image information loss is too big; The introducing of filtering apparatus has reduced the reliability and the versatility of system, because filtering apparatus can only allow infrared and near-infrared band passes through, total system can not be carried out other special monitoring simultaneously when carrying out fire monitoring; Having only filtering apparatus ceaselessly to switch the professional purpose that could carry out detection and each applying unit simultaneously before each camera surveys; The working condition of filtering apparatus directly has influence on the reliability of system; In addition, the fire location that it detects is still image coordinate, does not realize image coordinate and the conversion of coordinate on the spot, thereby the range of control of fire location and extinguishing device can't be mapped, and can't further realize the work of putting out a fire to save life and property automatically truly.
Summary of the invention:
The present invention is directed to the above-mentioned defective of existing fire monitoring technology, proposition utilizes colour TV camera shooting and monitoring environmental images, employing is carried out difference processing to colour image three-primary colors, and realize automatically the image coordinate and the corresponding conversion of coordinate on the spot realizing accurate judgement and sterically defined fire difference analysis technology to fire.
This method of utilizing the image three-primary colors difference to carry out detection and location comprises:
Absorb the signal of video signal of monitoring of environmental continuously by the CCD camera system;
Signal of video signal is deposited in computing machine or the image pick-up card by selected format;
Signal of video signal processed handle and make fire and judge unusually;
It is characterized in that: (1) described camera system adopts the colored CCD camera system; (2) describedly signal of video signal is processed to handle be that chromatic image is resolved into red (r), green (g), blue (b) three primary colours component image f r, f g, f b, carry out following calculus of differences then:
The first step is made calculus of differences with red primary image and blue primary image, gets red blue colour-difference image, and compares judgement with preset threshold T, gets the colour-difference binary image:
Figure C9510051900051
Second step is with difference moment t i, t jThe colour-difference binary image carry out calculus of differences again after, obtain the area difference image: (3) rate of growth of reference area difference image:
Figure C9510051900053
(4) judge whether to exist fire unusual:, to exist fire unusual as area rate of growth Δ S during greater than preset threshold G; (5) adopt direct linear transformation (DLT) technology to carry out fire coordinate setting.
Realize the method for image three-primary colors difference computing, it is characterized in that taking following steps: in the memory under program that (1) deposits the sequential instructions desiring to execute the task and control program in computing machine; (2) start computer system, CPU is by instruction that has deposited in and control program, sequentially the executable operations process; (3) open t iReading command constantly reads red primary component image f respectively rWith blue primary component image f b, and carry out difference operation, get i red blue colour-difference image constantly
Figure C9510051900054
, deposit in calculator memory or the hard disk; (4) if redder blue primary component is f r-f bGreater than given threshold value " T ", output is then arranged, otherwise no-output; (5) open t jReading command reads red primary component image f respectively constantly rWith blue primary component image f b, and carry out difference operation, get j red blue colour-difference image constantly
Figure C9510051900061
, deposit in calculator memory or the hard disk; (6) if redder blue primary component is f r-f bGreater than given threshold value " T ", output is then arranged, otherwise no-output; (7) compare f r-f bAfter red blue colour-difference image, and carry out the second difference computing, the area difference image: Δ D f r f b ( i , j ) ; (8) rate of growth of the rapid gained area difference of calculating previous step image ΔS = ( Σ 1 n ΔD f r f b | ΔD f r f b = 1 ) / ( t j - t i )
If Δ S then has output greater than setting threshold values G, judging has fire unusual; Otherwise circulation repeats above-mentioned (3)-(8) step.
The invention has the advantages that:
1. utilize of the difference reflection of colour image three-primary colors image first to early stage flame, the automatic differential technique realization of employing red primary image and blue primary image flame separates automatically with background images, the position of detection of fires in image, it need not to add optical filtering equipment, directly carry out the identification of fire, solved and adopted the gray scale image to add the problem that special filtering apparatus detection flame reliability is difficult to improve; The present invention adopts primary colours difference and second order difference, has embodied the intelligent characteristics of human identification large space incipient fire.It had both utilized the color character of flame, having utilized fire again is this thought of out of control burning, the employing of above-mentioned dual criterion has greatly improved the reliability of detection, has overcome other detector and has only utilized the local behavior of fire and physical features easily to produce wrong shortcoming.Because science and technology development, B is substituted by colour TV camera gradually, and this technology will more can adapt to the trend of development.Because the colourful CCD video camera price descends significantly, and need not to add filtering apparatus, adopt this method to increase under few situation in equipment cost, improved the reliability of utilizing the image detection of fires greatly; Adopt this method, can utilize existing closed-circuit television system equipment, by the automatic monitoring of computer realization fire.
2. the present invention combines direct linear transformation DLT (Direct Linear Transformation) technology first with detection, has solved to utilize the image detection of fires not have the space orientation problem that solves in the past.
3. the present invention combines the automatic detection and the space orientation of fire, except automatic warning, by control device, can also realize the early detection of fire and put out a fire to save life and property teamwork automatically, realizes the target that fire has been beaten long ago.
4. adopt the inventive method full and accurate scene of fire chromatic image of record automatically, provide the firsthand information for analyzing cause of fire.
Description of drawings:
Fig. 1 is a workflow diagram of the present invention:
Fig. 2 is the work system basic framework figure when a plurality of colourful CCD video cameras and automatic spraying device are arranged.
Embodiment:
Below in conjunction with embodiment the inventive method is done further concrete description.
Utilize mini-computer, as 386,486 serial microcomputers, variegate the multi-media image card outward, as Video Blaster SE card, colour TV camera can adopt the ccd video camera of models such as HDCL9206, HDCL9208, the HDCL9216 of Huading company.
Absorb the signal of video signal of monitoring of environmental continuously by colourful CCD video camera, utilize Color Image Acquisition card (or multimedia card) to deposit body in or hard disc of computer on to frame image collection, store with standard formats such as TIFF or PCX, utilize computer program or existing Photostyler software that every frame true color image is resolved into three kinds of basic color images as Microsoft Windows, i.e. width of cloth redness (red) component image, one width of cloth green (green) component image, one width of cloth blueness (blue) component image, depositing computing machine frame interior or image card respectively in deposits in the body, then, red component image and blue component image are subtracted each other, remove background, obtain red blue colour-difference image, and compare judgement, get the colour-difference binary image with the threshold values T that sets:
Figure C9510051900071
Suggestion threshold value T value is
Figure C9510051900072
For example for the daylight illumination space, T can be taken as 80-100 (gray level).The above results is deposited in the computing machine.
Get two different t constantly respectively 1, t 2The colour-difference binary image, they are carried out difference operation again, obtain the area difference image:
Figure C9510051900073
The rate of growth of reference area difference image:
Figure C9510051900081
When Δ S when not having fixed threshold value G, can judge to have fire.
The G value can be determined according to the size and the focus of camera experiment in monitoring space.For example, if focal length of camera f=15mm, monitoring bulk is 20m * 30m * 4m, and fire goods are based on wooden product, and then G gets 3 times of area difference image flame edge girths divided by the mistiming; Desirable 3-5 of mistiming second.
If judge to have fire, then further carry out the fire location.Closest approach image coordinate that the flame of getting fire is liftoff x ‾ = 1 n Σ x i - - - i = 1,2 , · · · n y ‾ = max { y i } - - - i = 1,2 , · · · n If there are many places to catch fire, then calculate many places image of flames coordinate x respectively 1, y 1, x 2, y 2... x k, y k, utilize DLT (direct linear transformation) formula x ‾ = L 1 X + L 2 Y + L 3 L 7 X + L 8 Y + 1 y ‾ = L 4 X + L 5 Y + L 6 L 7 X + L 8 Y + 1
Calculate fire coordinate position X, Y on the spot, coefficient L in the formula 1-L 8Find the solution by the DLT general-purpose algorithm.
X, Y are imported the system that puts out the fire automatically, can realize the interlock of device of putting out the fire, thereby realize the automatic detection of incipient fire and the work of putting out a fire to save life and property.
Its workflow can be represented by Fig. 1.
Its element can be represented with Fig. 2.
Below be a kind of expression way example of difference software flow and functional language:
Realize that with software mode the difference function is divided into five steps: one, read the ti moment R, G, B three primary colours image to internal memory, red[] [], green[] [], blue[] []
for(i=0;i<height;i++){
for(j=0;j<width;j++){
fread(&red[i][j],sizeof(unsigned?char),1,fp1):
Read the red primary image
fread(&green[i][j],sizeof(unsigned?char),1,fp2);
Read the green primary image
fread(&blue[i][j],sizeof(unsigned?char),1,fp3);
Read the blue primary image
}
Two, form t iColour-difference image D constantly Frfb1
for(i=0;i<height;i++){
for(j=0;j<width;i++){
If red[i] [j]-blue[i] [j]>T
D then Frfb1[i] [j]=1
Otherwise D Frfb1[i] [j]=0
}
Suggestion T value T=(gray scale image mean value+be not equal to zero minimum gradation value)/2 three, repeat (one) (two) step, try to achieve t jColour-difference image D constantly Frfb2Four, according to difference moment t i, t jThe colour-difference image of red blueprint elephant constantly, the visual Δ D of the difference of quadraturing Frfb
for(i=0;i<height;i++){
for(j=0;j<width;j++){
If D Frfb1[i] [j] equals D Frfb2[i] [j]
Δ D then Frfb[i] [j]=1
Otherwise Δ D Frf[i] [j]=0
}
Five, judge fire development, whether exists, the rate of rise by the calculating fire judges whether to exist out of control burning---fire.The rate of rise of fire adopts area difference image Δ D FrfbThe independent fire region area removes (t j-t i) mistiming calculates, and utilizes the method that region contour extracts in " C language image processing program storehouse " book, with Δ D FrfbEqual 1 zone and carry out the extraction and the numbering of profile, form the area difference image of each kindling point at last, their gray-scale value is become corresponding sequence number, for fire area, first place, Δ D Frfb=1; Fire area, second place, Δ D Frfb=2; Fire area, the 3rd place Δ D Frfb=3; Fire area, n place Δ D Frfb==n.
Calculate the area S (1) in n zone, S (2) ..., S (n) (is unit with the pixel)
for(i=0;i<height;i++){
for(j=0;j<width;j++){
If Δ D Frfb=1, then belong to the 1st zone
S(1)=S(1)+1;
If Δ Dfrfb=2, then belong to the 2nd zone
S(2)=S(2)+1
If Δ D Frfb=3, then belong to the 3rd zone
S(3)=S(3)+1;
…………………
If Δ D Frfb=n then belongs to the n zone
S(n)=S(n)+1
}
}
Zoning rate of rise V (i)
for(i=0;i<n;i++){
V(i)=S(i)/(t j-t i)
Judge the existence of fire
for(i=0;i<n;i++){
If V (i)>TV
Then there is fire
Otherwise there is not fire
Suggestion: t j-t i=3 seconds~5 seconds
S (i) equals three times of region contour girths
G=S(i)/(t j-t i)

Claims (2)

1. method of utilizing the image three-primary colors difference to carry out detection and location comprises:
Absorb the signal of video signal of monitoring of environmental continuously by the CCD camera system;
Signal of video signal is deposited in computing machine or the image pick-up card by selected format;
Signal of video signal processed handle and make fire and judge unusually;
It is characterized in that: (1) described camera system adopts the colored CCD camera system; (2) describedly signal of video signal is processed to handle be that chromatic image is resolved into red (r), green (g), blue (b) three primary colours component image f r, f g, f b, carry out following calculus of differences then:
The first step is made calculus of differences with red primary image and blue primary image, gets red blue colour-difference image, and compares judgement with preset threshold T, gets the colour-difference binary image:
Second step is with difference moment t i, t jThe colour-difference binary image carry out calculus of differences again after, obtain the area difference image:
Figure C9510051900022
(3) rate of growth of reference area difference image:
Figure C9510051900023
(4) judge whether to exist fire unusual:, to exist fire unusual as area rate of growth Δ S during greater than preset threshold G;
(5) adopt direct linear transformation (DLT) technology to carry out fire coordinate setting.
2. method that realizes the computing of image three-primary colors difference is characterized in that taking following steps:
(1) in the memory under program that deposits the sequential instructions desiring to execute the task and control program in computing machine;
(2) start computer system, CPU is by instruction that has deposited in and control program, sequentially the executable operations process;
(3) open t iReading command constantly reads red primary component image f respectively rWith blue primary component image f b, and carry out difference operation, get i red blue colour-difference image constantly
Figure C9510051900031
, deposit in calculator memory or the hard disk; (4) if redder, blue primary component is f r-f bGreater than given threshold value " T ", output is then arranged, otherwise no-output; (5) open t jReading command reads red primary component image f respectively constantly rWith blue primary component image f b, and carry out difference operation, get j red blue colour-difference image constantly
Figure C9510051900032
, deposit in calculator memory or the hard disk; (6) if redder blue primary component is f r-f bGreater than given threshold value " T ", output is then arranged, otherwise no-output; (7) compare f r-f bAfter red blue colour-difference image, and carry out the second difference computing, the area difference image:
Figure C9510051900033
(8) rate of growth of the rapid gained area difference of calculating previous step image ΔS = ( Σ 1 n Δ D f r f b | Δ D f r f b = 1 ) / ( t j - t i )
If Δ S then has output greater than setting threshold values G, judging has fire unusual; Otherwise circulation repeats above-mentioned (3)-(8) step.
CN 95100519 1995-03-08 1995-03-08 Method for detecting and positioning fire by using colour image three-primary colors difference Expired - Fee Related CN1092825C (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN 95100519 CN1092825C (en) 1995-03-08 1995-03-08 Method for detecting and positioning fire by using colour image three-primary colors difference

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN 95100519 CN1092825C (en) 1995-03-08 1995-03-08 Method for detecting and positioning fire by using colour image three-primary colors difference

Publications (2)

Publication Number Publication Date
CN1112702A CN1112702A (en) 1995-11-29
CN1092825C true CN1092825C (en) 2002-10-16

Family

ID=5073491

Family Applications (1)

Application Number Title Priority Date Filing Date
CN 95100519 Expired - Fee Related CN1092825C (en) 1995-03-08 1995-03-08 Method for detecting and positioning fire by using colour image three-primary colors difference

Country Status (1)

Country Link
CN (1) CN1092825C (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101020104A (en) * 2007-03-27 2007-08-22 南京大学 Embedded forest fire early-warning system based on video detection
CN1943824B (en) * 2006-09-08 2010-06-16 浙江工业大学 An automatic fire fighting unit based on omnibearing visual sensor

Families Citing this family (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN100559887C (en) * 2007-08-09 2009-11-11 山东省科学院自动化研究所 Utilize flame color template to carry out the method for detection
CN100511305C (en) * 2007-08-10 2009-07-08 山东省科学院自动化研究所 Double wave band image identification fire hazard detection alarming system and monitoring method
DE102008039132A1 (en) 2008-08-21 2010-02-25 Billy Hou Intelligent image smoke/flame sensor i.e. personal computer/CPU based intelligent image smoke/flame sensor, for intelligent image smoke/flame detection system in e.g. gym, has digital signal processor for turning on infrared lamp
CN101764999B (en) * 2009-07-28 2011-09-14 北京智安邦科技有限公司 Sub-camera video capture device
CN102903106B (en) * 2012-09-21 2015-09-02 东方网力科技股份有限公司 A kind of flame method for rapidly positioning of image-type fire detector and device
CN102999994A (en) * 2012-11-15 2013-03-27 镇江石鼓文智能化系统开发有限公司 Flame detection device based on video image analysis
CN103020587B (en) * 2012-11-15 2016-12-21 镇江石鼓文智能化系统开发有限公司 Based on video image analysis flame regarding figure analysis method
CN103324910A (en) * 2013-05-06 2013-09-25 南京新奕天智能视频技术有限公司 Fire alarming method based on video detection
CN105405244B (en) * 2015-12-22 2018-05-22 山东神戎电子股份有限公司 A kind of interference source screen method for forest fire protection

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1943824B (en) * 2006-09-08 2010-06-16 浙江工业大学 An automatic fire fighting unit based on omnibearing visual sensor
CN101020104A (en) * 2007-03-27 2007-08-22 南京大学 Embedded forest fire early-warning system based on video detection
CN101020104B (en) * 2007-03-27 2011-01-19 南京大学 Embedded forest fire early-warning system based on video detection

Also Published As

Publication number Publication date
CN1112702A (en) 1995-11-29

Similar Documents

Publication Publication Date Title
CN1092825C (en) Method for detecting and positioning fire by using colour image three-primary colors difference
CN110197589B (en) Deep learning-based red light violation detection method
CN101882034B (en) Device and method for discriminating color of touch pen of touch device
CN103632158B (en) Forest fire prevention monitor method and forest fire prevention monitor system
CN102298693B (en) Expressway bend detection method based on computer vision
TWI308725B (en)
CN1207924C (en) Method for testing face by image
CN107478590A (en) A kind of method of combination motor vehicle intelligent vision identification and remote exhaust emission detection
CA2091081A1 (en) System and method for image recovery
CN103020623A (en) Traffic sign detection method and equipment
CN103761529A (en) Open fire detection method and system based on multicolor models and rectangular features
CN114113129A (en) Lens tiny defect identification and grabbing system and method
CN113470050A (en) Method for analyzing, identifying and quantifying VOC (volatile organic compound) leakage by utilizing double-spectrum image
CN111652883A (en) Glass surface defect detection method based on deep learning
Chakraborty et al. Bangladeshi road sign detection based on YCbCr color model and DtBs vector
CN105195438A (en) Embedded type automatic pearl sorting device and method based on image recognition
CN103034997A (en) Foreground detection method for separation of foreground and background of surveillance video
JP2004030225A (en) White smoke detection method and apparatus
CN106128112B (en) Night bayonet vehicle identifies grasp shoot method
CN105872464B (en) Factory smoke discharge automatic monitoring system based on vision signal
CN106944362A (en) A kind of intelligent classification reminding method based on video image identification, apparatus and system
CN110991255A (en) Method for detecting fake-licensed vehicle based on deep learning algorithm
CN113569880B (en) Quick primary separation system and primary separation method for naturally curled tobacco leaves after baking
CN211904213U (en) Vehicle bottom checking system
CN109028234B (en) Range hood capable of identifying smoke grade

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant
C19 Lapse of patent right due to non-payment of the annual fee
CF01 Termination of patent right due to non-payment of annual fee