CN107577997A - The discrimination method that mountain fire is invaded in a kind of electric transmission line channel - Google Patents
The discrimination method that mountain fire is invaded in a kind of electric transmission line channel Download PDFInfo
- Publication number
- CN107577997A CN107577997A CN201710718569.5A CN201710718569A CN107577997A CN 107577997 A CN107577997 A CN 107577997A CN 201710718569 A CN201710718569 A CN 201710718569A CN 107577997 A CN107577997 A CN 107577997A
- Authority
- CN
- China
- Prior art keywords
- image
- mountain fire
- transmission line
- electric transmission
- point
- 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.)
- Pending
Links
Landscapes
- Image Analysis (AREA)
- Fire-Detection Mechanisms (AREA)
Abstract
The present invention provides a kind of discrimination method of mountain fire invasion in electric transmission line channel, including:Step 1: extract multiple image from the monitor video of electric transmission line channel;Step 2: morphologic filtering, Denoising disposal are carried out to the image extracted;Step 3: catching the moving object in image using the method for motion analysis between image, and the region of moving object is extracted from image;Step 4: feature decision is carried out to the moving object captured, mountain fire has more obvious several features, the color of moving object that the present invention is detected by progressively extracting, kinetic characteristic, area change rate, flicker frequency, wedge angle characteristic are determined whether for mountain fire, greatly improve the accuracy of mountain fire discrimination.
Description
Technical field
The present invention relates to intelligent power grid technology field, the identification side of mountain fire invasion in specifically a kind of electric transmission line channel
Method.
Background technology
Shown according to Guo Wang companies service data statistics over the years, the circuit caused by mountain fire in overhead power transmission line passage
Trip-out rate can be in any more, accounts for the main status of power transmission and transforming equipment failure.For a long time, authorities think that electric transmission line channel is transported
Dimension is mainly problem of management, and generally use, which is manually kept watch, protects the mode such as line with the masses and tackles, and expends a large amount of manpower and materials, effect compared with
Difference.Also have and the supervision equipments such as video, radar are installed on shaft tower, manually identify defect, accuracy and reliability cannot
Ensure, be unfavorable for popularization and application.The generation of abnormal conditions in line channel is prevented, will be a long-term process, and currently
One significant technology issues of transmission line safety O&M.
At present, the inspection of domestic transmission line of electricity is more by the way of artificial line walking, but with modern machines vision technique
Development, and the exploitation of various Digital Image Processing instruments use, and also occur detecting power transmission line using image processing techniques
The method of road dangerous matter sources.Such as:A kind of " high-voltage line foreign body intrusion object detection method " (application number:2014104849.X), although
Foreign matter is detected using image processing techniques, but and not differentiate between be which kind of foreign matter threatens to caused by transmission line of electricity." power transmission line
Road foreign body intrusion intelligent video on-line monitoring assessment system " (application number:201610661214.2), although refer to distinguish foreign matter
Type, but be a lack of specific recognition methods and strategy, lack certain specific aim.
The content of the invention
For deficiencies of the prior art, the present invention provides a kind of identification of mountain fire invasion in electric transmission line channel
Method, accurately distinguished by extracting characteristic quantity possessed by mountain fire, invasion mesh can accurately be identified by this method
Whether mark is mountain fire, so as to take corresponding prevention and prediction policy, is carried for the monitoring system of electric transmission line channel foreign body intrusion
For technical support.
The discrimination method that mountain fire is invaded in a kind of electric transmission line channel, comprises the following steps:
Step 1: extract multiple image from the monitor video of electric transmission line channel;
Step 2: being pre-processed to the image extracted, the pretreatment specifically includes morphologic filtering and denoising
Processing;
Step 3: utilize the moving object in image after the method for motion analysis seizure pretreatment between image;
Step 4: feature decision is carried out to the moving object captured:The moving object that will be captured
The RGB image of body be converted to progressively extract after HSV images color, kinetic characteristic, area change rate, flicker frequency and
Wedge angle property feature, color, kinetic characteristic, area change rate, flicker frequency and the wedge angle property feature of comprehensive extraction, which differentiate, is
It is no to have mountain fire, and according to the edge contour of extraction moving object, obtain the size and tendency of mountain fire generation.
Further, when extracting multiple image in the step 1 from the monitor video of electric transmission line channel, according to luring
The reason for sending out mountain fire and the sampling policy that mountain fire and season and weather relevant feature change video occurs.
Further, the step 4 is specially:
After the RGB image of the moving object captured is converted into HSV images, extracted according to the threshold value of the h components of setting
The Color Characteristic of foreign matter, the red area of foreign matter is obtained, is tentatively judged as mountain fire;
Subtract each other by the picture binaryzation of the red area of extraction, and by multiframe binaryzation picture, according to subtract each other result and
Whether scope and position in the picture, which changes, differentiates whether foreign matter has kinetic characteristic, is performed if with kinetic characteristic
In next step;
To the image of extraction after first two steps are screened color, kinetic characteristic and differentiated, (1) carries out segmentation portion to image
Reason, by moving target and background separation;(2) to two frame picture binaryzation before and after extraction, it is 1 to obtain numerical value in two frame pictures
Pixel number, tries to achieve area;(3) further according to area change rate formula, its area change rate is obtained, if area change rate reaches
Preset value is then carried out in next step;
Differentiated according to flicker frequency feature:(1) each frame selected areas is sought in the red area image of extraction
Average brightness value;(2) by FFT the brightness change frequency in time domain(i.e. flame fringe
The change frequency of pixel brightness) change frequency of brightness that is converted on frequency domain, f is to be dodged from the i-th frame to the brightness of i+1 frame
Bright frequency, t are the i-th frame to the time of i+1 frame, fAi(x,y)Represent in the i-th two field picture in Ai (x, y) if this pixel occurs
Flame brightness, its value are then 1, are otherwise 0, Fourier transform formula is:(3) in the base of preceding step
On plinth, in the multiple image of extraction per second, the gray value of same position from 0 is changed into 1 or 1 and is changed into 0 being defined as flicker once, one
In the fixed time, the change frequency of all pixels point, i.e. flicker frequency are counted, if flicker frequency meets theoretic frequency excursion
7-12Hz, then perform next step;
Differentiated according to wedge angle characteristic:(1) by the picture binaryzation of the red area of extraction;(2) pointed peak is found;
(3) judge whether pointed peak region is wedge angle;(4) summit of each flame is considered as a characteristic point, Ran Houzai
According to edge detection operator, the number for the pixel for meeting above-mentioned wedge angle feature is calculated, by wedge angle criterion, if being deposited in image
In the point with Sharp features, and wedge angle number reaches threshold value and irregular change is presented, then illustrates that the region has the several of flame
What feature, then can primitive decision be mountain fire.
Further, image is first converted into HSL spaces by the average brightness value using Matlab softwares, then to L points
Amount is asked for can obtain with mean sentences twice.
Further, it is described differentiated according to wedge angle characteristic in find pointed peak concretely comprise the following steps:It is 1 to see numerical value
The pixel that numerical value is 1 is whether there is above pixel, is not summit if having, sees left side and right side whether there is continuous numerical value if nothing
For 1 point, judge this point for summit if having.
Further, it is described to judge whether pointed peak region is that wedge angle concretely comprises the following steps:Wedge angle top is judged one by one
The following pixel number per a line of point is designated as f (n), then next line pixel number is f (n-1), the long and narrow degree f (n) of angle
Weighed with f (n-1) ratio, if the long and narrow degree of angle is less than the threshold value d set, be considered as wedge angle:, l1For
The distance of 15 pixels of a line or so, l below summit2For by the distance of 30 pixels of next line or so again in image
The present invention considers existing O&M mode servant according to the requirement of State Grid Corporation of China " 13 " program for the development of science and technology
The problem of power resource anxiety, for existing various cause calamity operating modes in electric transmission line channel, using technologies such as image procossings, lead to
The extraction to features such as color, kinetic characteristic, area change, flicker frequency, Sharp features is crossed, laddering progress is progressively sieved
Choosing, automatic identification dangerous matter sources and its movement locus, the accuracy rate that is differentiated to mountain fire will be greatly improved, finally establishes power transmission line
The real-time space protection scope in road.
Brief description of the drawings
Fig. 1 is hsv color model;
Fig. 2 is that the flow that foreign matter feature decision is carried out in the discrimination method that mountain fire is invaded in electric transmission line channel of the present invention is shown
It is intended to;
Fig. 3 is the detail flowchart of step 4 in the present invention (carrying out feature decision to the moving object captured).
Embodiment
Below in conjunction with the accompanying drawing in the present invention, the technical scheme in the present invention is clearly and completely described.
Referring to Fig. 2, the one of embodiment of discrimination method that mountain fire is invaded in electric transmission line channel of the present invention is included such as
Lower step:
Step 1: extract multiple image from the monitor video of electric transmission line channel;Multiframe figure is extracted from monitor video
As the main frequency issues for considering image sampling, for woods mountain fire disaster, mountain fire is concentrated mainly on winter and autumn,
The probability that mountain fire occurs for rainy season is smaller.So for this feature, we according to local meteorological condition, arid, rainwater compared with
The strategy of daily multiple repairing weld is carried out under few environment.Secondly, it is to be held a memorial ceremony for by people in specific red-letter day mountain fire majority to occur according to statistics
Offer sacriffices to the gods or the spirits of the dead caused by burning, such as:The Ching Ming Festival, the red-letter day such as Spring Festival belong to high-incidence season of mountain fire, so in these specific red-letter days sampling frequencies
Rate can also increase.
Step 2: being pre-processed to the image extracted, the pretreatment specifically includes morphologic filtering and denoising
Processing;Described to carry out morphologic filtering to the image that extracts, Denoising disposal is after being sampled to video image, right
Picture after sampling carries out gray proces, when the later image of gray processing is often containing many isolated points, isolated cell
Domain, small―gap suture and hole, in order to solve the problems, such as that the difference image after Threshold segmentation may have these, we used number
The methods of learning morphological images processing pre-processes to image.
Step 3: utilize the moving object in image after the method for motion analysis seizure pretreatment between image;
Step 4: feature decision is carried out to the moving object captured.The feature decision stage is that video camera detects motion
The committed step for mountain fire is determined whether after object.Because mountain fire has more obvious characteristic quantity, so using progressively extracting
Following characteristics:(1) color, (2) kinetic characteristic, (3) area change rate, (4) flicker frequency, (5) wedge angle characteristic determine whether
For mountain fire, and according to the edge contour of extraction moving object, obtain the size and tendency of mountain fire generation.The detailed stream of step 4
Journey is as shown in figure 3, be specifically described as follows:
First, the RGB image of the moving object captured is converted into HSV images, HSV is a kind of directly perceived for user
Color model, be easy to us to be used for the color segmentation specified, to extract the color characteristic of moving object.
RGB conversions HSV mathematical modeling is as follows:
V=max
R, g, b represent red, green, blueness space coordinates respectively, and between zero and one, max and min divide its span
The maxima and minima in tri- coordinate values of r, g, b is not represented.
H represents color information, and the parameter is represented with angular metric, and h value is generally 0 to 360 °, each number of degrees generation
A kind of color of table.As max=min, h=0, image is red, and when h value is 240 degree, representative is blueness.Ginseng
Number behalf saturation degree, the scope of its value is 0 to 1.Parameter v represents light levels, and scope is from 0 to 1.Hsv color model is shown in Fig. 1.
Foreign matter differentiates concretely comprising the following steps for stage:
1st, first, the color characteristic main distinction of mountain fire is that the temperature of flame combustion is different, and the color shown also can
There is certain deviation.The light that flame combustion is sent is mainly black body radiation, and the distribution of the color (wavelength) of black body radiation is main
Depending on temperature.In general mountain fire burning color is approximately at red and Chinese red region, and the higher region of temperature is partial to white
Color.So this feature burnt for mountain fire flame, original RGB pictures are converted into HSV forms by the present invention, and h components are probably set
0 ° to 30 ° of threshold range is set to, carries out dithering, the region of general red can be thus extracted, carry out preliminary
Mountain fire differentiates.This region extracted will be used as a filter, take back the image under RGB patterns, it is possible to extract desired
Color region.Therefore, the color that we want is extracted using h threshold range is set, you can the color for extracting foreign matter is special
Sign amount.
2nd, secondly, during dithering is identified, it is similar red in addition to mountain fire that image may recognize some
Object, such as the leaf in autumn, this just needs to make a distinction the kinetic characteristic of mountain fire.Mountain fire burning is special with certain motion
Sign, and be irregular.Specially:(1) extraction multiple image is converted into hsv forms;(2) red threshold scope is set,
Such as the scope of 0 ° to 30 ° of the threshold value of h components is set;(3) by the picture binaryzation of the red area of extraction, and by multiframe two-value
Change picture to subtract each other, if the region that the bianry image Central Plains numerical value after subtracting each other is 1 is changed into 0, the region that illustrating picture has foreign matter is
It is static constant, and if multiframe picture subtracts each other the region for still having that numerical value is 1 later, and scope in the picture and position
Put and all change, then just explanation foreign matter has kinetic characteristic, it may be possible to mountain fire occurs, so as to carry out the differentiation of next step, together
When can also further screen the stationary object of the upper similar mountain fire that loses color.
3rd, judged according to the area change of moving object in image, occur the initial stage of mountain fire, the area of flame be by
It is cumulative big.That is, each two field picture that we extract, after first two steps screen color, kinetic characteristic:(1) to figure
As carrying out dividing processing, its object is to by moving target and background separation;(2) to two frame picture binaryzation before and after extraction,
The pixel number that numerical value in two frame pictures is 1 is obtained, tries to achieve area;(3) further according to area change rate formula, its area is obtained
Rate of change.When fire occurs, flame irregular movement, area also constantly changes, and its area change rate also has unstability, and
As red direction board, the car light homalographic rate of change of motion can be excluded tentatively close to 0 object, and carry out sentencing for next step
Setting analysis.The formula of area change rate is as follows:
GiRepresent area change rate, Size (bi)tNumerical value is represented in t bianry image as 1 area, Size (bi)t0Generation
Table t0Numerical value is 1 area in moment bianry image.
Furthermore 4, the similar red object in motion may also can produce interference to identification process, such as:Automobile in motion
Light.So we recycle the flicker frequency that mountain fire burns to make a distinction.The flame frequency of combustible combustion is about in 3-
Change in the range of 25Hz, be concentrated mainly on 7-12Hz.So (1) asks each frame to choose in the red area image of extraction
Image is converted into HSL spaces by the average brightness value in region, average brightness value with Matlab image processing softwares, then to L points
Amount is asked for can obtain with mean sentences twice;(2) by FFT the brightness change frequency in time domainThe change frequency for the brightness that (i.e. the change frequency of flame fringe pixel brightness) is converted on frequency domain
Rate, f are the brightness flicker frequency from the i-th frame to i+1 frame, and t is the i-th frame to the time of i+1 frame, fAi(x,y)Represent i-th
For two field picture in Ai (x, y) if flame brightness occurs in this pixel, its value is then 1, is otherwise 0.Fourier transform formula is:(3) on the basis of preceding step, in multiframe (at least 20 frames) image of extraction per second, same position
Gray value from 0 be changed into 1 or 1 be changed into 0 be defined as flicker once, within the regular hour, count all pixels point change frequency
Rate, see whether meet theoretic frequency excursion 7-12Hz, it is possible to probably mountain fire is determined whether, so as to screen out it
The influence of his foreign matter.
5th, finally differentiated according to another obvious feature of mountain fire, have during flame combustion one it is more obvious several
What feature is exactly Sharp features, and wedge angle number is more, and irregular change is presented.Wedge angle is by pixel one by one in image
Point composition, so, (1) handles gradation of image, extracts the color gamut of 0 ° to 30 ° of threshold value, then image is switched into binary map
Picture;(2) pointed peak is found:See the pixel top that numerical value is 1 whether there is the pixel that numerical value is 1, be not summit if having, if
Without then seeing left side and right side whether there is the point that continuous numerical value is 1, judge this point for summit if having.(3) judge one by one under summit
Pixel number of the face per a line is designated as f (n), then next line pixel number is f (n-1), angle long and narrow degree f (n) and f
(n-1) ratio is weighed.If the long and narrow degree of angle is less than the threshold value d set, it is considered as wedge angle.
l1For the distance of 15 pixels of a line below summit or so, l2For by next line or so 30 again in image
The distance of individual pixel.(4) summit of each flame is considered as a characteristic point, then further according to edge detection operator, meter
Calculate the number for the pixel for meeting above-mentioned wedge angle feature.If by above-mentioned criterion, exist in image with Sharp features
Point, and the irregular change of the more presentation of wedge angle number, then illustrate that the region has the geometric properties of flame, then can primitive decision be
Mountain fire occurs.Wedge angle quantity can be typically taken more than 3 because its threshold value acquirement of different application scenarios is variant.
Using above method, by being carried to features such as color, kinetic characteristic, area change, flicker frequency, Sharp features
Take, laddering carry out Stepwise Screening, the accuracy rate that is differentiated to mountain fire will be greatly improved.
Claims (6)
1. the discrimination method that mountain fire is invaded in a kind of electric transmission line channel, it is characterised in that comprise the following steps:
Step 1: extract multiple image from the monitor video of electric transmission line channel;
Step 2: being pre-processed to the image extracted, the pretreatment specifically includes morphologic filtering and Denoising disposal;
Step 3: utilize the moving object in image after the method for motion analysis seizure pretreatment between image;
Step 4: feature decision is carried out to the moving object captured:The RGB image of the moving object captured is converted to
Color, kinetic characteristic, area change rate, flicker frequency and wedge angle property feature, the face of comprehensive extraction are progressively extracted after HSV images
Color, kinetic characteristic, area change rate, flicker frequency and wedge angle property feature discriminate whether mountain fire, and according to extraction
The edge contour of moving object, obtain the size and tendency of mountain fire generation.
2. the discrimination method that mountain fire is invaded in electric transmission line channel as claimed in claim 1, it is characterised in that:The step 1
In when multiple image is extracted from the monitor video of electric transmission line channel, according to the reason for inducing mountain fire and occurring mountain fire and season
Section and weather relevant feature change the sampling policy of video.
3. the discrimination method that mountain fire is invaded in electric transmission line channel as claimed in claim 1, it is characterised in that:The step 4
Specially:
After the RGB image of the moving object captured is converted into HSV images, foreign matter is extracted according to the threshold value of the h components of setting
Color Characteristic, obtain the red area of foreign matter, be tentatively judged as mountain fire;
Subtract each other by the picture binaryzation of the red area of extraction, and by multiframe binaryzation picture, according to subtracting each other result and scheming
Whether scope and position as in, which change, differentiates whether foreign matter has kinetic characteristic, is performed if with kinetic characteristic next
Step;
To the image of extraction after first two steps are screened color, kinetic characteristic and differentiated, (1) carries out dividing processing to image, will
Moving target and background separation;(2) to two frame picture binaryzation before and after extraction, the pixel that numerical value in two frame pictures is 1 is obtained
Point number, tries to achieve area;(3) further according to area change rate formula, its area change rate is obtained, if area change rate reaches default
Value is then carried out in next step;
Differentiated according to flicker frequency feature:(1) being averaged for each frame selected areas is asked in the red area image of extraction
Brightness value;(2) by FFT the brightness change frequency in time domain(i.e. flame fringe pixel
The change frequency of point brightness) change frequency of brightness that is converted on frequency domain, f is from the i-th frame to the brightness flicker of i+1 frame frequency
Rate, t are the i-th frame to the time of i+1 frame, fAi(x,y)Represent in the i-th two field picture in Ai (x, y) if flame occurs in this pixel
Brightness, its value are then 1, are otherwise 0, Fourier transform formula is:(3) on the basis of preceding step,
In the multiple image of extraction per second, the gray value of same position from 0 is changed into 1 or 1 and is changed into 0 being defined as flicker once, certain
In time, the change frequency of all pixels point, i.e. flicker frequency are counted, if flicker frequency meets theoretic frequency excursion 7-
12Hz, then perform next step;
Differentiated according to wedge angle characteristic:(1) by the picture binaryzation of the red area of extraction;(2) pointed peak is found;(3)
Judge whether pointed peak region is wedge angle;(4) summit of each flame is considered as a characteristic point, then further according to
Edge detection operator, the number for the pixel for meeting above-mentioned wedge angle feature is calculated, by wedge angle criterion, if tool in image be present
There is the point of Sharp features, and wedge angle number reaches threshold value and irregular change is presented, then illustrates that the region has the geometry of flame special
Sign, then can primitive decision be mountain fire.
4. the discrimination method that mountain fire is invaded in electric transmission line channel as claimed in claim 3, it is characterised in that:It is described average bright
Image is first converted into HSL spaces by angle value using Matlab softwares, then L * component is asked for obtaining with mean sentences twice
Arrive.
5. the discrimination method that mountain fire is invaded in electric transmission line channel as claimed in claim 3, it is characterised in that:It is described according to point
Angle characteristic is found pointed peak in being differentiated and concretely comprised the following steps:See the pixel top that numerical value is 1 whether there is the pixel that numerical value is 1
Point, it is not summit if having, if seeing left side and right side whether there is the point that continuous numerical value is 1 without if, judges this point for top if having
Point.
6. the discrimination method that mountain fire is invaded in electric transmission line channel as claimed in claim 3, it is characterised in that:It is described to judge point
Whether angular vertex region is that wedge angle concretely comprises the following steps:Judge that the pixel number below pointed peak per a line is designated as one by one
F (n), then next line pixel number are f (n-1), and the long and narrow degree of angle is weighed with f (n) and f (n-1) ratio, if angle
Long and narrow degree is less than the threshold value d set, then is considered as wedge angle:l1For 15 pixels of a line below summit or so away from
From l2For by the distance of 30 pixels of next line or so again in image.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710718569.5A CN107577997A (en) | 2017-08-21 | 2017-08-21 | The discrimination method that mountain fire is invaded in a kind of electric transmission line channel |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710718569.5A CN107577997A (en) | 2017-08-21 | 2017-08-21 | The discrimination method that mountain fire is invaded in a kind of electric transmission line channel |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107577997A true CN107577997A (en) | 2018-01-12 |
Family
ID=61034241
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710718569.5A Pending CN107577997A (en) | 2017-08-21 | 2017-08-21 | The discrimination method that mountain fire is invaded in a kind of electric transmission line channel |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107577997A (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110363192A (en) * | 2018-04-11 | 2019-10-22 | 大众电脑股份有限公司 | Object image identification system and object image discrimination method |
CN112133052A (en) * | 2020-09-22 | 2020-12-25 | 岭澳核电有限公司 | Image fire detection method for nuclear power plant |
CN112216055A (en) * | 2020-09-24 | 2021-01-12 | 佛山市天然气高压管网有限公司 | Fire early warning detection method and system for gas field station |
CN113077424A (en) * | 2021-03-23 | 2021-07-06 | 广东电网有限责任公司广州供电局 | Power transmission line channel environment change detection method and system based on deep learning |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
GB2428473A (en) * | 2005-07-18 | 2007-01-31 | Sony Uk Ltd | Fire detection by processing video images |
CN101515326A (en) * | 2009-03-19 | 2009-08-26 | 浙江大学 | Method for identifying and detecting fire flame in big space |
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 |
CN107025753A (en) * | 2017-06-05 | 2017-08-08 | 天津汉光祥云信息科技有限公司 | A kind of wide area fire alarm installation analyzed based on multispectral image |
-
2017
- 2017-08-21 CN CN201710718569.5A patent/CN107577997A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
GB2428473A (en) * | 2005-07-18 | 2007-01-31 | Sony Uk Ltd | Fire detection by processing video images |
CN101515326A (en) * | 2009-03-19 | 2009-08-26 | 浙江大学 | Method for identifying and detecting fire flame in big space |
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 |
CN107025753A (en) * | 2017-06-05 | 2017-08-08 | 天津汉光祥云信息科技有限公司 | A kind of wide area fire alarm installation analyzed based on multispectral image |
Non-Patent Citations (2)
Title |
---|
熊国良 等: "火焰特性识别的Matlab实现方法", 《计算机工程与科学》 * |
陈磊 等: "基于视频的火焰检测方法", 《计算机工程与设计》 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110363192A (en) * | 2018-04-11 | 2019-10-22 | 大众电脑股份有限公司 | Object image identification system and object image discrimination method |
CN110363192B (en) * | 2018-04-11 | 2023-04-14 | 大众电脑股份有限公司 | Object image identification system and object image identification method |
CN112133052A (en) * | 2020-09-22 | 2020-12-25 | 岭澳核电有限公司 | Image fire detection method for nuclear power plant |
CN112216055A (en) * | 2020-09-24 | 2021-01-12 | 佛山市天然气高压管网有限公司 | Fire early warning detection method and system for gas field station |
CN112216055B (en) * | 2020-09-24 | 2021-05-11 | 佛山市天然气高压管网有限公司 | Fire early warning detection method and system for gas field station |
CN113077424A (en) * | 2021-03-23 | 2021-07-06 | 广东电网有限责任公司广州供电局 | Power transmission line channel environment change detection method and system based on deep learning |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN115601367B (en) | LED lamp wick defect detection method | |
CN107577997A (en) | The discrimination method that mountain fire is invaded in a kind of electric transmission line channel | |
CN108307146B (en) | System and method for detecting potential safety hazard of high-voltage transmission line | |
CN111967393B (en) | Safety helmet wearing detection method based on improved YOLOv4 | |
CN106680285B (en) | Method for recognizing insulator contamination state based on infrared image assisted visible light image | |
CN103442209A (en) | Video monitoring method of electric transmission line | |
CN100520362C (en) | Method for detecting forest fire fog based on colorful CCD image analysis | |
CN111814678B (en) | Method and system for identifying coal flow in conveyor belt based on video monitoring | |
CN103778418A (en) | Mountain fire image identification method of image monitoring system of electric transmission line tower | |
CN105868722A (en) | Identification method and system of abnormal power equipment images | |
CN111814686A (en) | Vision-based power transmission line identification and foreign matter invasion online detection method | |
CN103235938A (en) | Method and system for detecting and identifying license plate | |
CN113887412B (en) | Detection method, detection terminal, monitoring system and storage medium for pollution emission | |
CN105812618B (en) | A kind of method for testing motion and motion detection apparatus | |
CN106096603A (en) | A kind of dynamic flame detection method merging multiple features and device | |
CN103679704A (en) | Video motion shadow detecting method based on lighting compensation | |
CN112287823A (en) | Facial mask identification method based on video monitoring | |
CN109241847A (en) | The Oilfield Operation District safety monitoring system of view-based access control model image | |
CN102737221A (en) | Method and apparatus for vehicle color identification | |
Abbas et al. | Automated pavement distress detection using image processing techniques | |
CN107680089B (en) | A kind of automatic judging method of ultra-high-tension power transmission line camera image exception | |
CN110533626B (en) | All-weather water quality identification method | |
CN104168462B (en) | Camera scene change detection method based on image angle point set feature | |
CN115797775A (en) | Intelligent illegal building identification method and system based on near-earth video image | |
CN108154490A (en) | Based on the high-voltage transmission line insulator image enchancing method for improving fuzzy set theory |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180112 |