CN106096603A - A kind of dynamic flame detection method merging multiple features and device - Google Patents
A kind of dynamic flame detection method merging multiple features and device Download PDFInfo
- Publication number
- CN106096603A CN106096603A CN201610382004.XA CN201610382004A CN106096603A CN 106096603 A CN106096603 A CN 106096603A CN 201610382004 A CN201610382004 A CN 201610382004A CN 106096603 A CN106096603 A CN 106096603A
- Authority
- CN
- China
- Prior art keywords
- flame
- suspicious region
- pixel
- region
- color
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 63
- 238000000034 method Methods 0.000 claims abstract description 21
- 230000004397 blinking Effects 0.000 claims abstract description 19
- 230000000877 morphologic effect Effects 0.000 claims abstract description 13
- 238000009825 accumulation Methods 0.000 claims abstract description 7
- 238000012545 processing Methods 0.000 claims description 10
- 239000000284 extract Substances 0.000 claims description 8
- 230000011218 segmentation Effects 0.000 claims description 7
- 230000001186 cumulative effect Effects 0.000 claims description 4
- 238000002372 labelling Methods 0.000 claims description 4
- 238000007619 statistical method Methods 0.000 claims description 3
- 210000003205 muscle Anatomy 0.000 claims description 2
- 230000004069 differentiation Effects 0.000 claims 2
- 238000000205 computational method Methods 0.000 claims 1
- 238000012544 monitoring process Methods 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000004927 fusion Effects 0.000 description 3
- 238000007689 inspection Methods 0.000 description 3
- 230000003542 behavioural effect Effects 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000008676 import Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/56—Extraction of image or video features relating to colour
Abstract
The invention discloses a kind of dynamic flame detection method merging multiple features and device.Wherein method includes: set up the flame color model of multiple color spaces, apply this model that video flowing is carried out color detection, accumulation difference is used to carry out motion detection to by the region of color detection, obtain doubtful flame region, the shape and structure feature calculating doubtful flame region carries out spatial domain Morphological Identification, the blinking characteristics analyzing doubtful flame region carries out time domain dynamic discriminant, if meeting pre-provisioning request, the most doubtful flame region is real flame region.Flame detecting method of the present invention, improves the accuracy of detection of flame, reduces flame false drop rate.Successful Application for intelligent monitor system provides technical guarantee, has the strongest practical value and realistic meaning.
Description
Technical field
The present invention relates to target detection technique field, a kind of dynamic flame detection method merging multiple features and dress
Put.
Background technology
Along with the raising of China's power network schedule automation, the unattended operation of transformer station is the development of current power system
Trend, in transformer substation intelligent monitoring system, fire defector is primarily to auxiliary monitoring personnel safeguard substation safety, and it is
Carry out researching and developing based on technology such as Digital Image Processing, Digital Video Processing, computer vision and Model Identification, by means of calculating
Machine treatment technology, carries out high speed analysis to the mass data in monitor video, the flame automatically occurred in detection monitoring place, and
To monitoring, personnel report to the police, and to avoid the generation of fire, bring about great losses.
In current video flame detecting method, only consider flame individually or Partial Feature, as single color characteristic,
Behavioral characteristics or shape facility, not from overall these features of angle comprehensive consideration, in some cases, due to obtained
Characteristic set is insufficient, or different objects has identical feature, and traditional method can be other objects, such as flicker at night
Car light, be mistaken for flame.
It addition, current flame detecting method directly utilizes before and after's frame frame difference to determine flame zone when extracting behavioral characteristics
Territory, will report to the police when the object meeting flame color feature is moved, and this can cause the flase drop of flame equally.
So, how to improve fire defector precision, it is urgently to be resolved hurrily in video flame detection for reducing fire defector false drop rate
Problem.
Summary of the invention
In order to solve above-mentioned technical problem, the present invention proposes a kind of dynamic flame detection method merging multiple features and dress
Put, by the flame in Fusion of Color feature, motion feature, shape and structure feature and blinking characteristics detection video flowing.
According to an aspect of the present invention, it is provided that a kind of video flame detecting method based on multi-feature fusion, described method
Comprise the following steps:
Step 1, sets up flame color model based on RGB, HSI and YUV color space；
Step 2, the color model that applying step 1 is set up carries out color detection to current frame to be detected, obtains having flame
The color detection target area of color characteristic；
Step 3, extracts adjacent 2 frames of frame to be detected, uses accumulation difference to carry out motion detection, obtains motion detection target
Region, will be that the color detection target region with motion detection target area is as flame suspicious region simultaneously；
Step 4, calculates the shape and structure feature of flame suspicious region, flame suspicious region is carried out spatial domain Morphological Identification,
Reject the flame suspicious region not meeting spatial domain Morphological Identification condition；
Step 5, on the basis of step 4, carries out statistical to the situation of change of flame suspicious region pixel gray value
Analysis, obtains the flame suspicious region dynamic blinking characteristics of time domain, differentiates described flame suspicious region according to the dynamic blinking characteristics of time domain
Whether it is real flame region.
According to a further aspect of the invention, it is provided that a kind of video flame based on multi-feature fusion detection device, described dress
Put and include:
Flame color model building module, for setting up flame color model based on RGB, HSI and YUV color space；
Color target detection module, for current frame to be detected being carried out color detection according to the color model set up,
Obtain the color detection target area with flame color feature；
Moving object detection module, for extracting adjacent 2 frames of frame to be detected, uses accumulation difference to carry out motion detection,
Obtain motion detection target area, using doubtful as flame for the region being color detection target and motion detection target area simultaneously
Region；
Spatial domain Morphological Identification module, for calculating the shape and structure feature of flame suspicious region, enters flame suspicious region
Line space territory Morphological Identification, rejects the flame suspicious region not meeting spatial domain Morphological Identification condition；
Time domain dynamic discriminant module, for carrying out statistical to the situation of change of flame suspicious region pixel gray value
Analysis, obtains the flame suspicious region dynamic blinking characteristics of time domain, differentiates described flame suspicious region according to the dynamic blinking characteristics of time domain
Whether it is real flame region.
Technical scheme, by setting up the flame color model of multiple color spaces, applies this model to flow to video
Row color detection, uses accumulation difference to carry out motion detection to by the region of color detection, obtains doubtful flame region, calculate
The shape and structure feature of doubtful flame region carries out spatial domain Morphological Identification, and the blinking characteristics analyzing doubtful flame region carries out time domain
Dynamic discriminant, if meeting pre-provisioning request, the most doubtful flame region is real flame region.This video flame detecting method,
Improve the accuracy of detection of video flame, reduce video flame false drop rate, the successful Application for intelligent monitor system provides
Technical guarantee, has the strongest practical value and realistic meaning.
Accompanying drawing explanation
Fig. 1 is total steps flow chart schematic diagram of a kind of dynamic flame detection method merging multiple features of the present invention.
Fig. 2 is that the steps flow chart of the motion detection of a kind of dynamic flame detection method merging multiple features of the present invention shows
It is intended to.
Fig. 3 is the step of the flame sudden strain of a muscle frequency feature extraction of a kind of dynamic flame detection method merging multiple features of the present invention
Rapid schematic flow sheet.
Detailed description of the invention
In order to make those skilled in the art be more fully understood that the present invention.Attached below in conjunction with in the embodiment of the present invention
Figure, is purged the method step in the embodiment of the present invention.It is fully described by.Based on the embodiment in the present invention, this area
That those of ordinary skill is obtained under not making creative work premise so other embodiments, all should belong to the application
The scope of protection.
The overall flow figure of a kind of dynamic flame detection method merging multiple features that Fig. 1 provides for the present invention, Fig. 2 is this
The flow chart of steps of the motion detection in inventive method, Fig. 3 is the flame flicking feature extraction flow chart in the inventive method.
A kind of dynamic flame detection method merging multiple features that the present invention provides specifically includes following steps:
Step 1, foundation flame color model based on RGB, HSI and YUV color space, in one embodiment, flame face
Color model is expressed as follows:
Flame color model:
Wherein, R, G, B are the red, green, blue color components of flame；RtRed component threshold value for pixel；Y is flame
Luminance component, Y=0.299*R+0.587*G+0.114*B, YtLuminance threshold for pixel；Trg=| R-G |, Tgb=| G-B |,
Thrg、ThgbIt is T respectivelyrg、TgbThreshold value, in a preferred embodiment, each threshold value value is Yt=180, Rt=135, Thrg=
35, Thgb=35；
Certainly, in other embodiments, it is also possible to set up other flame color models that can detect flame color.
The flame color model that step 2, applying step 1 are set up is to the i-th frame video F in video to be detectediCarry out color inspection
Survey, the pixel meeting flame color model is composed original value, otherwise, composes 0 value, obtain the color inspection with flame color feature
Survey target area；
Step 3, extract the i-th frame video F in video to be detectediBefore adjacent 2 frame Fi-1、Fi-2, use accumulation difference to enter
Row motion detection, obtains motion detection target area, will be color detection target area and motion detection target area simultaneously
Region is as flame suspicious region；
In one embodiment, described motion detection comprises the following steps:
Step 3.1, the i-th frame of video F to video to be detectedi2 frame Fs adjacent with iti-1、Fi-2Carry out gray processing process；
Step 3.2, ask for gray processing process after the i-th frame of video Fi2 frame Fs adjacent with iti-1、Fi-2The most adjacent two interframe
Frame difference figure, respectively FD, i=| Fi-Fi-1|, FD, i-1=| Fi-1-Fi2|；
Step 3.3, calculate the gray average M of two width frame difference figures respectivelyi、Mi-1, the most right using this average as segmentation threshold
The frame difference figure F that step 3.2 obtainsD, i、FD, i-1Carry out binary segmentation, obtain the two width frame corresponding binary map of difference figure；
Step 3.4, by two width binary map make and operations, with result be move detect target area；
Step 4, the shape and structure feature of calculating flame suspicious region, including circularity, rectangular degree, length-width ratio, to flame
Suspicious region carries out spatial domain Morphological Identification, if these eigenvalues of flame suspicious region meet sets threshold value, thinks and meets fire
The shape of flame, proceeds subsequent step, otherwise it is assumed that this region is not flame region, stops subsequent step；
In one embodiment, each shape and structure feature calculation method is as follows:
Circularity:Rectangular degree:Length-width ratio:
Wherein S is the area of flame suspicious region, and L is the girth of flame suspicious region, SRMinimum for flame suspicious region
The area of boundary rectangle, LRFor the length of flame suspicious region minimum enclosed rectangle, WRFor flame suspicious region minimum enclosed rectangle
Width；
Step 5, on the basis of step 4, statistical analysis frame to be detected FiIn each flame suspicious region pixel gray scale
Value is at its adjacent 3 frame Fi-1、Fi-2、Fi-3On situation of change, obtain the dynamic blinking characteristics of time domain of each flame suspicious region, warp
The comparison crossed and set threshold value, synthetic determination realizes the detection to flame, and wherein, the dynamic blinking characteristics of described time domain extracts and includes
Following steps:
Step 5.1, calculating frame F to be detected after step 4 is screenedi3 frame Fs adjacent with iti-1、Fi-2、Fi-3Difference exhausted
To value, respectively FD, i=| Fi-Fi-1|, FD, i-1=| Fi-1-Fi-2|, FD, i-2=| Fi-2-Fi-3|；
Step 5.2, calculate the cumulative mean value F of above-mentioned difference absolute valueM, i,
Step 5.3, extract frame F to be detectediMiddle all flame suspicious region L after step 4 is screenedt, and labelling is corresponding
The coordinate of pixel, first flame suspicious region L1Pixel be designated as F respectivelyI, L1(x11, y11), FI, L1(x12, y12) ...,
FI, L1(x1n1, y1n1), FI, L1Refer to first flame suspicious region L1, F in the i-th frameI, L1(x11, y11) refer to first fire in the i-th frame
First pixel in the L1 of flame suspicious region by that analogy, the t flame suspicious region LtPixel be designated as F respectivelyI, Lt
(xt1, yt1), FI, Lt(xt2, yt2) ..., FI, Lt(xtnt, ytnt)
Step 5.4, average F obtained according to step 5.2M, iCalculate the pixel grey scale average of each flame suspicious region
FAL, t, i, first flame suspicious region gray averageBy that analogy, the t flame is doubted
Like area grayscale averageWherein, FM, i, L1Refer to the average image F shown in step 5.2M, iIn
First flame suspicious region L1；N1 refers to pixel number in first flame suspicious region；
Step 5.5, extract adjacent interframe grey scale change more than the pixel of its place doubtful flame region gray average 2 times
As flame flicking pixel, and add up Fi、Fi-1、Fi-2、Fi-3Flame flicker pixel number NUMT, i.I.e. for Fi、Fi-1、
Fi-2、Fi-3Four frames, obtain adjacent interframe grey scale change respectively, obtain 3 grey scale change values, the flame flicking pixel that statistics obtains
Point number is following pixel number sum: Fi、Fi-1Between grey scale change value more than gray average 2 times pixel number,
Fi-1、Fi-2Between grey scale change value more than the pixel number of gray average 2 times, Fi-2、Fi-3Between grey scale change value more than ash
The pixel number of degree average 2 times.
Step 5.6, calculating FiIn the flame flicking eigenvalue feature of each flame suspicious regionT, i, work as featureT, i
More than setting threshold value fthTime judge that this region, as real flame, as a example by the t flame suspicious region, calculates this region area
AT, i, flame flicking eigenvalueIn one embodiment, flame flicking characteristic threshold value fthValue is 2.
The present invention, in the software design of embodiment, is provided with import feature and can provide importing test video, be provided with inspection
Brake includes showing fire defector result character information, display fire defector position coordinates, display fire defector design sketch, with
Time the video needed can be carried out storage and preserve, use the INTERFACE DESIGN of hommization to make overall use of software be more prone to behaviour
Make.
The above, the only detailed description of the invention of the present invention, but protection scope of the present invention is not limited thereto, and any
Those familiar with the art is in scope disclosed by the invention, it is possible to the change readily occurred in or replacement, all should contain
Lid is within the scope of the invention as claimed.
Claims (10)
1. the dynamic flame detection method merging multiple features, it is characterised in that said method comprising the steps of:
Step 1, sets up flame color model based on RGB, HSI and YUV color space；
Step 2, the color model that applying step 1 is set up carries out color detection to current frame to be detected, obtains having flame color
The color detection target area of feature；
Step 3, extracts adjacent 2 frames of frame to be detected, uses accumulation difference to carry out motion detection, obtains motion detection target area
Territory, will be that the color detection target region with motion detection target area is as flame suspicious region simultaneously；
Step 4, calculates the shape and structure feature of flame suspicious region, flame suspicious region carries out spatial domain Morphological Identification, rejects
Do not meet the flame suspicious region of spatial domain Morphological Identification condition；
Step 5, on the basis of step 4, carries out statistical analysis to the situation of change of flame suspicious region pixel gray value,
To the flame suspicious region dynamic blinking characteristics of time domain, differentiate that whether described flame suspicious region is according to the dynamic blinking characteristics of time domain
Real flame region.
2. the method for claim 1, it is characterised in that the motion detection in described step 3, specifically includes following step
Rapid:
Adjacent 2 frames before current frame to be detected and its are carried out gray processing process by step 31；
Step 32, the most adjacent two frames of 3 two field pictures after processing gray processing subtract each other, and obtain two width differential charts；
Step 33, calculates the average of two width differential charts respectively, as segmentation threshold, two width difference drawing is carried out two-value using this average
Segmentation, obtains two width binary map；
Two width binary map are made and operation, are moving target with result by step 34.
3. the method for claim 1, it is characterised in that the dynamic blinking characteristics of time domain in described step 5 extracts, specifically
Comprise the following steps:
Step 51, calculates difference absolute value of adjacent 3 frames before current frame to be detected and its；
Step 52, calculates the cumulative mean value of above-mentioned difference absolute value；
Step 53, extracts the flame suspicious region in current frame to be detected, and the coordinate of labelling respective pixel；
Step 54, according to the pixel grey scale average of the current frame Flame suspicious region to be detected of described accumulative mean value computation；
Step 55, makees adjacent interframe grey scale change more than the pixel of its flame suspicious region, place pixel grey scale average 2 times
For flame flicking pixel, add up current frame to be detected and each flame suspicious region Flame in this 4 frame of adjacent 3 frames before it
Flicker pixel number；
Step 56, calculates the area of each flame suspicious region, using the ratio of flame flicking pixel number and area as this
The dynamic blinking characteristics of time domain of flame suspicious region.
4. the method for claim 1, it is characterised in that described step 1 Flame color model is expressed as:
Flame color model:
Wherein, R, G, B are the red, green, blue color components of flame；RtRed component threshold value for pixel；Y is the brightness of flame
Component, YtLuminance threshold for pixel；Trg=| R-G |, Tgb=| G-B |, Thrg、ThgbIt is T respectivelyrg、TgbThreshold value.
5. the method for claim 1, it is characterised in that in described step 4, shape and structure feature includes: circularity, square
Shape degree, length-width ratio, computational methods are as follows:
Circularity:Rectangular degree:Length-width ratio:
Wherein S is the area of flame suspicious region, and L is the girth of flame suspicious region, SRFor the minimum external square in flame suspicious region
The area of shape, LRFor the length of flame suspicious region minimum enclosed rectangle, WRWidth for flame suspicious region minimum enclosed rectangle
Degree.
6. the method for claim 1, it is characterised in that described in step 4, flame suspicious region is carried out spatial domain form
Differentiate and refer to that described shape and structure feature and form are set threshold value compares differentiation；Described in step 5 dynamic according to time domain
Blinking characteristics differentiates whether described flame suspicious region is that real flame region refers to dynamic for described time domain blinking characteristics and sudden strain of a muscle
Bright feature-set threshold value compares differentiation.
7. the dynamic flame detection device merging multiple features, it is characterised in that described device includes:
Flame color model building module, for setting up flame color model based on RGB, HSI and YUV color space；
Color target detection module, for current frame to be detected being carried out color detection according to the color model set up, obtains
There is the color detection target area of flame color feature；
Moving object detection module, for extracting adjacent 2 frames of frame to be detected, uses accumulation difference to carry out motion detection, obtains
Motion detection target area, will be that the color detection target region with motion detection target area is as the doubtful district of flame simultaneously
Territory；
Spatial domain Morphological Identification module, for calculating the shape and structure feature of flame suspicious region, carries out sky to flame suspicious region
Territory Morphological Identification, rejects the flame suspicious region not meeting spatial domain Morphological Identification condition；
Time domain dynamic discriminant module, for the situation of change of flame suspicious region pixel gray value is carried out statistical analysis,
To the flame suspicious region dynamic blinking characteristics of time domain, differentiate that whether described flame suspicious region is according to the dynamic blinking characteristics of time domain
Real flame region.
8. device as claimed in claim 7, it is characterised in that described moving object detection module includes:
Gray proces submodule, for carrying out gray processing process to adjacent 2 frames before current frame to be detected and its；
Mathematic interpolation submodule, the most adjacent two frames of 3 two field pictures after processing gray processing subtract each other, and obtain two width differential charts；
Binary segmentation submodule, for calculating the average of two width differential charts respectively, poor to two width using this average as segmentation threshold
Drawing carries out binary segmentation, obtains two width binary map；
Moving target output sub-module, for two width binary map being made and operation, is moving target with result.
9. device as claimed in claim 7, it is characterised in that described time domain dynamic discriminant module includes:
Difference absolute value calculating sub module, for calculating difference absolute value of adjacent 3 frames before current frame to be detected and its；
Cumulative mean value calculating sub module, for calculating the cumulative mean value of above-mentioned difference absolute value；
Pixel coordinate labelling submodule, for extracting the flame suspicious region in current frame to be detected, and labelling respective pixel
Coordinate；
Gray average calculating sub module, for the picture according to the current frame Flame suspicious region to be detected of described accumulative mean value computation
Element gray average；
Flame flicking pixel statistics submodule, for being more than its flame suspicious region, place pixel by adjacent interframe grey scale change
The pixel that gray average is 2 times, as flame flicking pixel, adds up current frame to be detected and before it in this 4 frame of adjacent 3 frames
Each flame suspicious region Flame flicker pixel number；
Time domain dynamic blinking characteristics submodule, for calculating the area of each flame suspicious region, by flame flicking pixel
The ratio of number and area is as the dynamic blinking characteristics of time domain of this flame suspicious region.
10. device as claimed in claim 7, it is characterised in that described flame color model is expressed as:
Flame color model:
Wherein, R, G, B are the red, green, blue color components of flame；RtRed component threshold value for pixel；Y is the brightness of flame
Component, YtLuminance threshold for pixel；Trg=| R-G |, Tgb=| G-B |, Thrg、ThgbIt is T respectivelyrg、TgbThreshold value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610382004.XA CN106096603A (en) | 2016-06-01 | 2016-06-01 | A kind of dynamic flame detection method merging multiple features and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610382004.XA CN106096603A (en) | 2016-06-01 | 2016-06-01 | A kind of dynamic flame detection method merging multiple features and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106096603A true CN106096603A (en) | 2016-11-09 |
Family
ID=57446875
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610382004.XA Pending CN106096603A (en) | 2016-06-01 | 2016-06-01 | A kind of dynamic flame detection method merging multiple features and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106096603A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106845443A (en) * | 2017-02-15 | 2017-06-13 | 福建船政交通职业学院 | Video flame detecting method based on multi-feature fusion |
| CN107220628A (en) * | 2017-06-06 | 2017-09-29 | 北京环境特性研究所 | The method of infrared jamming source detection |
| CN107609603A (en) * | 2017-10-09 | 2018-01-19 | 济南大学 | A kind of image matching method of multiple color spaces difference fusion |
| CN108038867A (en) * | 2017-12-22 | 2018-05-15 | 湖南源信光电科技股份有限公司 | Fire defector and localization method based on multiple features fusion and stereoscopic vision |
| CN108229458A (en) * | 2017-12-22 | 2018-06-29 | 湖南源信光电科技股份有限公司 | A kind of intelligent flame recognition methods based on motion detection and multi-feature extraction |
| CN108334855A (en) * | 2018-02-24 | 2018-07-27 | 南瑞集团有限公司 | A kind of substation's flame identification algorithm using enhancing RGB component feature |
| CN109472192A (en) * | 2018-09-20 | 2019-03-15 | 国网江苏省电力有限公司检修分公司 | A kind of outside transformer substation has the flame image recognition methods of anti-interference ability |
| CN109726620A (en) * | 2017-10-31 | 2019-05-07 | 北京国双科技有限公司 | A kind of video flame detecting method and device |
| CN110263696A (en) * | 2019-06-17 | 2019-09-20 | 沈阳天眼智云信息科技有限公司 | Flame detection method based on infrared video |
| CN111539239A (en) * | 2019-01-22 | 2020-08-14 | 杭州海康微影传感科技有限公司 | Method, device and storage medium for open fire detection |
| CN114882401A (en) * | 2022-04-29 | 2022-08-09 | 清远蓄能发电有限公司 | Flame detection method and system based on RGB-HSI model and flame initial growth characteristics |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104598895A (en) * | 2015-02-10 | 2015-05-06 | 天津艾思科尔科技有限公司 | Method and device for flame detection based on video image analysis |
-
2016
- 2016-06-01 CN CN201610382004.XA patent/CN106096603A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104598895A (en) * | 2015-02-10 | 2015-05-06 | 天津艾思科尔科技有限公司 | Method and device for flame detection based on video image analysis |
Non-Patent Citations (4)
| Title |
|---|
| 吴茜茵等: "多特征融合的火焰检测算法", 《智能系统学报》 * |
| 徐铭铭: "基于多特征融合的森林火灾火焰视频识别研究", 《中国优秀硕士学位论文全文数据库农业科技辑》 * |
| 杜静: "基于视觉显著性分析的视频火焰检测", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
| 谢伟: "多特征融合的视频火焰检测算法研究", 《电脑知识与技术》 * |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106845443A (en) * | 2017-02-15 | 2017-06-13 | 福建船政交通职业学院 | Video flame detecting method based on multi-feature fusion |
| CN106845443B (en) * | 2017-02-15 | 2019-12-06 | 福建船政交通职业学院 | Video flame detection method based on multi-feature fusion |
| CN107220628A (en) * | 2017-06-06 | 2017-09-29 | 北京环境特性研究所 | The method of infrared jamming source detection |
| CN107220628B (en) * | 2017-06-06 | 2020-04-07 | 北京环境特性研究所 | Method for detecting infrared interference source |
| CN107609603A (en) * | 2017-10-09 | 2018-01-19 | 济南大学 | A kind of image matching method of multiple color spaces difference fusion |
| CN109726620A (en) * | 2017-10-31 | 2019-05-07 | 北京国双科技有限公司 | A kind of video flame detecting method and device |
| CN109726620B (en) * | 2017-10-31 | 2021-02-05 | 北京国双科技有限公司 | Video flame detection method and device |
| CN108038867A (en) * | 2017-12-22 | 2018-05-15 | 湖南源信光电科技股份有限公司 | Fire defector and localization method based on multiple features fusion and stereoscopic vision |
| CN108229458A (en) * | 2017-12-22 | 2018-06-29 | 湖南源信光电科技股份有限公司 | A kind of intelligent flame recognition methods based on motion detection and multi-feature extraction |
| CN108334855A (en) * | 2018-02-24 | 2018-07-27 | 南瑞集团有限公司 | A kind of substation's flame identification algorithm using enhancing RGB component feature |
| CN109472192A (en) * | 2018-09-20 | 2019-03-15 | 国网江苏省电力有限公司检修分公司 | A kind of outside transformer substation has the flame image recognition methods of anti-interference ability |
| CN111539239A (en) * | 2019-01-22 | 2020-08-14 | 杭州海康微影传感科技有限公司 | Method, device and storage medium for open fire detection |
| CN111539239B (en) * | 2019-01-22 | 2023-09-22 | 杭州海康微影传感科技有限公司 | Open fire detection method, device and storage medium |
| CN110263696A (en) * | 2019-06-17 | 2019-09-20 | 沈阳天眼智云信息科技有限公司 | Flame detection method based on infrared video |
| CN114882401A (en) * | 2022-04-29 | 2022-08-09 | 清远蓄能发电有限公司 | Flame detection method and system based on RGB-HSI model and flame initial growth characteristics |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106096603A (en) | A kind of dynamic flame detection method merging multiple features and device | |
| CN107705288B (en) | Infrared video detection method for dangerous gas leakage under strong interference of pseudo-target motion | |
| CN109859171A (en) | A kind of flooring defect automatic testing method based on computer vision and deep learning | |
| CN102567727B (en) | Method and device for replacing background target | |
| Recky et al. | Windows detection using k-means in cie-lab color space | |
| CN106408846A (en) | Image fire hazard detection method based on video monitoring platform | |
| CN106408015A (en) | Road fork identification and depth estimation method based on convolutional neural network | |
| CN111339883A (en) | Method for identifying and detecting abnormal behaviors in transformer substation based on artificial intelligence in complex scene | |
| CN111047655A (en) | High-definition camera cloth defect detection method based on convolutional neural network | |
| CN112560675B (en) | Bird visual target detection method combining YOLO and rotation-fusion strategy | |
| CN105678318B (en) | The matching process and device of traffic sign | |
| CN106483143A (en) | A kind of solar energy photovoltaic panel dust stratification on-Line Monitor Device and its detection method | |
| CN105069816B (en) | A kind of method and system of inlet and outlet people flow rate statistical | |
| CN105223208B (en) | A kind of circuit board detecting template and preparation method thereof, circuit board detecting method | |
| CN112183472A (en) | Method for detecting whether test field personnel wear work clothes or not based on improved RetinaNet | |
| CN106228541A (en) | The method and device of screen location in vision-based detection | |
| CN105096305A (en) | Method and device for analyzing state of insulators | |
| CN103489012A (en) | Crowd density detecting method and system based on support vector machine | |
| CN108898098A (en) | Early stage video smoke detection method based on monitor supervision platform | |
| CN104282013B (en) | A kind of image processing method and device for foreground target detection | |
| Shengze et al. | Research based on the HSV humanoid robot soccer image processing | |
| CN115359406A (en) | Post office scene figure interaction behavior recognition method and system | |
| CN108334823A (en) | High-resolution remote sensing image container area area detecting method based on machine learning | |
| CN113869245A (en) | Method and device for identifying safety region | |
| Lee et al. | Contaminated facade identification using convolutional neural network and image processing |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20161109 |