CN108648406A - A kind of indoor smog detection alarm system based on machine vision - Google Patents
A kind of indoor smog detection alarm system based on machine vision Download PDFInfo
- Publication number
- CN108648406A CN108648406A CN201810265915.3A CN201810265915A CN108648406A CN 108648406 A CN108648406 A CN 108648406A CN 201810265915 A CN201810265915 A CN 201810265915A CN 108648406 A CN108648406 A CN 108648406A
- Authority
- CN
- China
- Prior art keywords
- smog
- detection device
- image
- machine vision
- feature
- 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
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B17/00—Fire alarms; Alarms responsive to explosion
- G08B17/10—Actuation by presence of smoke or gases, e.g. automatic alarm devices for analysing flowing fluid materials by the use of optical means
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
Abstract
The invention discloses a kind of, and the indoor smog based on machine vision detects alarm system, belongs to fire-fighting domain.Present system includes video image video camera, color of image space convertor, preprocessor, sport foreground detection device, behavioral characteristics detection device and static nature detection device, lossless process method based on high-definition image, pass through the high concurrent algorithm ability of image processor, intelligent analysis is carried out to smog in real time, and there is certain adaptive learning ability, wherein alarm susceptibility can effectively be adjusted according to site environment.The more Gaussian Backgrounds of mixing that we apply model sport foreground detection algorithm it can be found that small smog waves, there is pre-alarm ability for smog flame, the fire starting stage is effectively alarmed, the fire for leading to large area is prevented, causes the loss that can not be retrieved.
Description
Technical field
The present invention relates to a kind of, and the indoor smog based on machine vision detects alarm system, belongs to fire-fighting domain.
Background technology
The prevention and detection of fire are always people's emphasis of interest.And fire is when occurring, the phenomenon that occurring earlier is
Smog.Therefore, seem more important for the detection of smog.Used detection means is divided into mostly at present：Smoke detection
Device and video monitoring.Various sense cigarette type detectors need smog to reach a certain concentration and reach in detector a certain range, ability
It is effectively alarmed, if at tall and big spacious place, the effect of seeming of such detector is insufficient.And with the high speed of modern society
Development, large-diameter building are increasing.Once fire occurs, when such detector sends out alarm, very big loss is had resulted in,
Lose effect and the purpose of original detection.
Current indoor smog alarm mainly uses the methods of the alarm of pole early-stage smog and indoor smog alarm, pole early stage cigarette
Mist alarm is called aspirating smoke detection or pole early fire detection, is exactly that the air of protection zone is inhaled by air sampling tube
Enter detector and analyzed to carry out the early warning of fire, which there are at high price and installing engineering is expensive,
And need periodic cleaning and replace probe etc., later maintenance cost is high, and wrong report etc. is easy under the messy environment of environment and is asked
Topic.
Indoor smog alarm is broadly divided into photoelectricity and ion smoke alarm, and the sensitivity of such smoke alarm is low,
It could alarm after reaching certain to the concentration of smog and requiring, being unable to reach smog in the early stage and just having generated the stage effectively to alarm
Function, fire is uncontrollable when often alarming, and cannot find dangerous situation in time at smog flame generation initial stage, danger is strangled
In cradle.There are one optics labyrinths in photoelectric mist alarm, are equipped with infrared tube, infrared receiving tube can not receive when smokeless
The infrared light that infrared transmitting tube is sent out, when flue dust enters optics labyrinth, by reflecting, reflecting, reception pipe receives infrared light,
Whether intelligent alarm circuit judges are more than threshold value, if it exceeds sending out alarm.There are one ionisation chamber, ions for ion smoke alarm
Radioactive element used in room -- americium 241 (Am241), about 0.8 micromicrocurie of intensity or so are in the equilibrium-like of electric field under normal condition
State can destroy this equilibrium relation when there is flue dust to enter ionisation chamber, and warning circuit detects meeting when concentration is more than the threshold value of setting
Send out alarm.
And need a large amount of personnel that could carry out comprehensive, round-the-clock real time monitoring common video monitoring.This
Kind scheme is even more to be difficult to carry out.First, needing a large amount of personnel that could in real time be observed numerous videos；Second is that monitoring personnel is not
It may round-the-clock energy concentration.
Invention content
To solve the above problems, the present invention provides a kind of indoor smog detection alarm system based on machine vision.
Indoor smog provided by the invention based on machine vision detects alarm system：Video image video camera, figure
As color space converter, preprocessor, sport foreground detection device, behavioral characteristics detection device, static nature detection device；
The video image video camera obtains video flowing using standard Streaming transfer protocol RTSP and is decoded as YUV420
Format-pattern.
Described image color space converter, by YUV420 format conversions at rgb format or HSV formats.
The preprocessor carries out preprocessing process to video image.Preprocessing process includes memory application, image turn
It changes, image processor initialization.
The sport foreground detection device obtains smog using more Gaussian Background modeling sport foreground detection algorithms are mixed
Shape, and acceleration optimization is carried out using image processor.It can be arranged according to the class requirement of different scenes and mix more Gausses back ofs the body
Scape models the sensitivity level of sport foreground detection algorithm.The sport foreground detection device is from Thread-Level Parallelism and asynchronous flow
It manages parallel two aspects and parallel optimization is carried out to the more Gaussian Background modeling sport foreground detection algorithms of the mixing.
The behavioral characteristics detection device detects the morphological feature and motion feature of smog.
The static nature detection device detects the color characteristic and shape feature of smog.Wherein, the shape of smog is detected
Feature refers to the ratio of the perimeter and area in the region that smog is calculated according to the non-regular shape of smog.
In one embodiment of the invention, the morphological feature of the behavioral characteristics detection device detection smog is：Root
According to the shape of smog is obtained, the metamorphosis of a time series is counted, while extracting smog sample and non-smog sample, is utilized
Linear SVM trains to obtain puff profile variation model, filtering non-smoke region.
In one embodiment of the invention, the motion feature of the behavioral characteristics detection device detection smog is：It adopts
The direction of motion that smog is obtained with image light flow calculation methodologies, is filtered non-smoke region.
In one embodiment of the invention, the behavioral characteristics detection device is obtained using 2-d wavelet computational methods
The same area formed smog before with formed smog after texture energy, formed smog before texture energy background modeling background
It is calculated on image, the texture energy formed after smog calculates on present image.
In one embodiment of the invention, the color characteristic of the static nature detection device detection smog refers to cigarette
Mist rgb space color model, smog rgb space color model are：
R ± α=G ± α=B ± α
80≤G≤220
B ＞ G ＞ R
Wherein, α is a constant for value very little, we generally take α=5, due in the application system acquisition and storage side
Formula is YUV models, so need to convert rgb space color model, there are following transformational relations with yuv space for rgb space：
R=Y+1.402 × (Cr-128)
G=Y-0.34414 × (Cb-128-0.71414 × (Cr-128)
B=Y+1.772 × (Cb-128)
Wherein, Y indicates that brightness, that is, grayscale value, Cr indicate RGB input signals RED sector and rgb signal brightness
Difference between value, Cb indicate the difference between RGB input signals blue portion and rgb signal brightness value；So cigarette can be obtained
Mist YUV color space models：
(Cr-128)=1.26 × (Cb-128)
80≤Y-0.3×(Cb-128)-0.7×(Cr-128)≤220
1.7 × (Cb-128) ＞ -0.3 × (Cb-128) -0.7 × (Cr-128) ＞ 1.4* (Cr-128).
The present invention also provides the indoor smogs based on machine vision to detect alarm method, includes the following steps：
Step 1：The smog image of video image video camera collection site；
Step 2：Collected smog image is converted into rgb format or HSV formats by color of image space convertor；
Step 3：Preprocessor carries out preprocessing process to the video image after conversion format；
Step 4：The sport foreground detection device models sport foreground detection algorithm using more Gaussian Backgrounds are mixed, and
Acceleration optimization is carried out using image processor；
Step 5：Behavioral characteristics detection device detects the morphological feature and motion feature of smog, static nature detection device
Detect the color characteristic and shape feature of smog；
Step 6：It determines whether to send out alarm in conjunction with alarm level.
Mist detecting device and method provided by the invention based on video image, the lossless place based on current high-definition image
Reason method carries out intelligent analysis to smog in real time by the high concurrent algorithm ability of image processor, and with it is certain from
Adaptive learning ability, wherein can be according to the sensitivity level of different scenes environmental requirement flexible modulation detection algorithm, sensitivity etc.
Grade is divided into high, medium and low three grades (generally taking middle rank).The more Gaussian Background modeling sport foreground detections of mixing that the present invention applies
Algorithm can reach 0.002%obs/m, have pre-alarm ability for smog flame it can be found that small smog waves,
The fire starting stage is effectively alarmed, the fire for leading to large area is prevented, causes the loss that can not be retrieved.
Description of the drawings
Fig. 1 is the mist detecting device structural schematic diagram based on video image that the embodiment of the present invention one provides；
Fig. 2 is the smog detection method flow diagram provided by Embodiment 2 of the present invention based on video image.
Specific implementation mode
Below in conjunction with the drawings and specific embodiments to a kind of mist detecting device based on video image proposed by the present invention
And method is described in further detail.According to following explanation and claims, advantages and features of the invention will become apparent from.It needs
Illustrate, attached drawing is all made of very simplified form and uses non-accurate ratio, only to convenient, lucidly auxiliary is said
The purpose of the bright embodiment of the present invention.
Embodiment one
A kind of mist detecting device based on video image that the present embodiment one provides, including video image video camera, figure
As color space converter, preprocessor, sport foreground detection device, behavioral characteristics detection device and static nature detection dress
It sets.
Specifically, the video image video camera encodes high-definition network camera using H264, and use standard Streaming Media
Transport protocol RTSP obtains video flowing and is decoded as YUV420 format-patterns；Described image color space converter, by YUV420
Format conversion carries out image analysis at rgb format or HSV formats；The preprocessor, pre-processes video image
Process；The sport foreground detection device obtains the shape of smog using more Gaussian Background modeling sport foreground detection algorithms are mixed
Shape, and carry out acceleration optimization by a large amount of stream calculation unit using image processor, the sport foreground detection device is from line
Journey grade parallel carries out the more Gaussian Background modeling sport foreground detection algorithms of the mixing with parallel two aspects of asynchronous stream process
Parallel optimization.
The Thread-Level Parallelism is mainly the kernel function using CUDA, by the context update of each pixel of former algorithm
It is handled on map procedures a to stream processing unit of GPU, calculating speed is accelerated by the parallel execution of multithreading；Institute
The concept in asynchronous process Optimum utilization CUDA programming models is stated, the data transmission between each stream is made and calculates overlapping progress,
Time delay caused by hiding data transmission, to accelerate calculating process.Also, the more Gaussian Background modeling sport foreground detections of mixing
Sensitivity level can be arranged in algorithm itself, before mixing more Gaussian Background modeling movements according to the setting of the class requirement of different scenes
The sensitivity level of scape detection algorithm, heavy smog can also be adapted to by being adapted to thin smog.
The morphological feature and motion feature of the behavioral characteristics detection device detection smog.The morphological character refers to,
The forming process of smog is all that irregular variable condition is presented, obtains the shape of smog according to sport foreground detection device, unite
The metamorphosis of a time series is counted, while extracting a large amount of smog samples and non-smog sample, is trained using Linear SVM
To puff profile variation model, to filter non-smoke region；Motion feature master refers to that smog is integrally presented one and floats generally upwards
Dynamic situation using this feature, while using image light flow calculation methodologies that can obtain the direction of motion of smog, carried out
Filter non-smoke region.With the difference after formation before motion feature also has smoke region to be formed simultaneously, counts while caning be found that field
The texture energy of smoke region is substantially less than the texture energy to form smoke foreground in scape, using this feature, the dynamic
Feature detection device simultaneously using 2-d wavelet computational methods can obtain the same area formed smog before with formed smog after
Texture energy, the texture energy before formation calculate on the background image of background modeling, and texture energy is in present image after formation
Upper calculating.
The color characteristic and shape feature of the static nature detection device detection smog.The color characteristic refers to cigarette
Mist is generally divided into green cigarette and white cigarette, is obtained by many experiments, R, G, the B of smog (including green cigarette and white cigarette) in rgb space
Value is not much different, and gray value, substantially within the scope of 80-220, smog is light blue color, i.e. B under the irradiation of light>G>R.Therefore,
Show that smog rgb space color model is：
R ± α=G ± α=B ± α
80≤G≤220
B ＞ G ＞ R
Wherein, α is a constant (concrete numerical value range) for value very little, due in the application system acquisition and storage mode
It is YUV models, so need to convert rgb space color model, there are following transformational relations with yuv space for rgb space：
R=Y+1.402 × (Cr-128)
G=Y-0.34414 × (Cb-128-0.71414 × (Cr-128)
B=Y+1.772 × (Cb-128)
Wherein, Y indicates that brightness, that is, grayscale value, Cr indicate RGB input signals RED sector and rgb signal brightness
Difference between value, Cb indicate the difference between RGB input signals blue portion and rgb signal brightness value；So cigarette can be obtained
Mist YUV color space models：
(Cr-128)=1.26 × (Cb-128)
80≤Y-0.3×(Cb-128)-0.7×(Cr-128)≤220
1.7 × (Cb-128) ＞ -0.3 × (Cb-128) -0.7 × (Cr-128) ＞ 1.4* (Cr-128).
Shape feature refers to then that the non-regular shape of smog calculates the ratio of the perimeter and area in the region of smog.
Embodiment two
The present invention also provides a kind of smog detection methods based on video image, include the following steps：
Step 1：The smog image of video image video camera collection site；
Step 2：Collected smog image is converted into rgb format or HSV formats by color of image space convertor；
Step 3：Preprocessor carries out preprocessing process to the video image after conversion format；
Step 4：The sport foreground detection device models sport foreground detection algorithm using more Gaussian Backgrounds are mixed, and
Acceleration optimization is carried out using image processor；
Step 5：Behavioral characteristics detection device detects the morphological feature and motion feature of smog, static nature detection device
Detect the color characteristic and shape feature of smog.
The morphological character refers to, is all that irregular variable condition is presented, before movement in the forming process of smog
Scape detection device obtains the shape of smog, counts the metamorphosis of a time series, at the same extract a large amount of smog samples with it is non-
Smog sample trains to obtain puff profile variation model using Linear SVM, to filter non-smoke region；Motion feature master is
Refer to smog and the situation waved generally upwards is integrally presented, is using this feature, while using image light flow calculation methodologies
The direction of motion that smog can be obtained is filtered non-smoke region.Motion feature also has smoke region to form preceding and shape simultaneously
Difference after counts while caning be found that the texture energy of smoke region in scene is substantially less than the line to form smoke foreground
Energy is managed, using this feature, the behavioral characteristics detection device can obtain using 2-d wavelet computational methods same simultaneously
Region is formed before smog with the texture energy after formation smog, and the texture energy before formation is counted on the background image of background modeling
It calculates, texture energy calculates on present image after formation.
The color characteristic refers to that smog is generally divided into green cigarette and white cigarette, is obtained by many experiments, smog (including blueness
Cigarette and white cigarette) R, G, B value in rgb space is not much different, and gray value is substantially within the scope of 80-220, photograph of the smog in light
It penetrates down, is light blue color, i.e. B>G>R.Therefore, show that smog rgb space color model is：
R ± α=G ± α=B ± α
80≤G≤220
B ＞ G ＞ R
Wherein, α is a constant (generally taking α=5) for value very little, since system acquisition in the application is with storage mode
YUV models, so need to convert rgb space color model, there are following transformational relations with yuv space for rgb space：
R=Y+1.402 × (Cr-128)
G=Y-0.34414 × (Cb-128-0.71414 × (Cr-128)
B=Y+1.772 × (Cb-128)
Wherein, Y indicates that brightness, that is, grayscale value, Cr indicate RGB input signals RED sector and rgb signal brightness
Difference between value, Cb indicate the difference between RGB input signals blue portion and rgb signal brightness value；So cigarette can be obtained
Mist YUV color space models：
(Cr-128)=1.26 × (Cb-128)
80≤Y-0.3×(Cb-128)-0.7×(Cr-128)≤220
1.7 × (Cb-128) ＞ -0.3 × (Cb-128) -0.7 × (Cr-128) ＞ 1.4* (Cr-128).
Shape feature refers to then the ratio of perimeter and area that the non-regular shape of smog calculates the region of smog.
Although the present invention has been described by way of example and in terms of the preferred embodiments, it is not limited to the present invention, any to be familiar with this skill
The people of art can do various change and modification, therefore the protection model of the present invention without departing from the spirit and scope of the present invention
Enclosing be subject to what claims were defined.
Claims (8)
1. a kind of indoor smog based on machine vision detects alarm system, which is characterized in that including：Video image video camera,
Color of image space convertor, preprocessor, sport foreground detection device, behavioral characteristics detection device, static nature detection dress
It sets；
The video image video camera obtains video flowing using standard Streaming transfer protocol RTSP and is decoded as YUV420 formats
Image；
Described image color space converter, by YUV420 format conversions at rgb format or HSV formats；
The preprocessor carries out preprocessing process to video image, and preprocessing process includes memory application, image conversion, figure
As processor initializes；
The sport foreground detection device obtains the shape of smog using more Gaussian Background modeling sport foreground detection algorithms are mixed
Shape, and acceleration optimization is carried out using image processor；
The behavioral characteristics detection device detects the morphological feature and motion feature of smog；
The static nature detection device detects the color characteristic and shape feature of smog；Wherein, the shape feature of smog is detected
Refer to the ratio of the perimeter and area in the region that smog is calculated according to the non-regular shape of smog.
2. a kind of indoor smog based on machine vision according to claim 1 detects alarm system, which is characterized in that institute
The more Gaussian Backgrounds modeling sport foregrounds inspections of mixing can be arranged according to the class requirement of different scenes by stating sport foreground detection device
The sensitivity level of method of determining and calculating；The sport foreground detection device is right in terms of parallel two of Thread-Level Parallelism and asynchronous stream process
The more Gaussian Background modeling sport foreground detection algorithms of mixing carry out parallel optimization.
3. a kind of indoor smog based on machine vision according to claim 1 detects alarm system, which is characterized in that institute
It is the shape according to smog to state behavioral characteristics detection device, counts the metamorphosis of a time series, while extracting smog sample
Sheet and non-smog sample, train to obtain puff profile variation model, to filter non-smoke region using Linear SVM.
4. a kind of indoor smog based on machine vision according to claim 1 or 3 detects alarm system, feature exists
In the behavioral characteristics detection device obtains the direction of motion of smog using image light flow calculation methodologies, is filtered non-smog
Region.
5. a kind of indoor smog based on machine vision according to claim 1 or 4 detects alarm system, feature exists
In, the behavioral characteristics detection device obtained using 2-d wavelet computational methods the same area formed before smog with form smog after
Texture energy, formed smog before texture energy calculated on the background image of background modeling, formed smog after texture energy
Amount calculates on present image.
6. detecting alarm system, feature according to a kind of any indoor smog based on machine vision of Claims 1 to 5
It is, the color characteristic of the static nature detection device detection smog refers to：
Smog rgb space color model is：
R ± α=G ± α=B ± α
80≤G≤220
B ＞ G ＞ R
Wherein, α is a constant for value very little, generally takes α=5, there are following transformational relations with yuv space for rgb space：
R=Y+1.402 × (Cr-128)
G=Y-0.34414 × (Cb-128-0.71414 × (Cr-128)
B=Y+1.772 × (Cb-128)
Wherein, Y indicates that brightness, Cr indicate the difference between RGB input signals RED sector and rgb signal brightness value, Cb tables
Show the difference between RGB input signals blue portion and rgb signal brightness value, smog YUV color space models can be obtained：
(Cr-128)=1.26 × (Cb-128)
80≤Y-0.3×(Cb-128)-0.7×(Cr-128)≤220
1.7 × (Cb-128) ＞ -0.3 × (Cb-128) -0.7 × (Cr-128) ＞ 1.4* (Cr-128).
7. a kind of indoor smog based on machine vision according to claim 2 detects alarm system, which is characterized in that spirit
Sensitivity grade takes precautions against grade setting according to the use environment of user and to smog, is divided into high, medium and low three sensitivity levels, and one
As be defaulted as middle rank.
8. a kind of indoor smog based on machine vision detects alarm method, which is characterized in that any using claim 1~7
The system, includes the following steps：
Step 1：The smog image of video image video camera collection site；
Step 2：Collected smog image is converted into rgb format or HSV formats by color of image space convertor；
Step 3：Preprocessor carries out preprocessing process to the video image after conversion format；
Step 4：The sport foreground detection device models sport foreground detection algorithm using more Gaussian Backgrounds are mixed, and uses
Image processor carries out acceleration optimization；
Step 5：Behavioral characteristics detection device detects the morphological feature and motion feature of smog, the detection of static nature detection device
The color characteristic and shape feature of smog；
Step 6：It determines whether to send out alarm in conjunction with alarm level.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810265915.3A CN108648406A (en) | 2018-03-28 | 2018-03-28 | A kind of indoor smog detection alarm system based on machine vision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810265915.3A CN108648406A (en) | 2018-03-28 | 2018-03-28 | A kind of indoor smog detection alarm system based on machine vision |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108648406A true CN108648406A (en) | 2018-10-12 |
Family
ID=63745106
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810265915.3A Pending CN108648406A (en) | 2018-03-28 | 2018-03-28 | A kind of indoor smog detection alarm system based on machine vision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108648406A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109522845A (en) * | 2018-11-19 | 2019-03-26 | 国网四川省电力公司电力科学研究院 | Distribution transformer based on intelligent robot tests safety work measure of supervision |
| CN109637073A (en) * | 2019-01-08 | 2019-04-16 | 重庆水利电力职业技术学院 | A kind of fire hazard aerosol fog identifying system based on optical flow method |
| CN110796822A (en) * | 2019-09-15 | 2020-02-14 | 杭州拓深科技有限公司 | Method for prejudging fire point direction based on improved smoke alarm |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104408745A (en) * | 2014-11-18 | 2015-03-11 | 北京航空航天大学 | Real-time smog scene detection method based on video image |
| CN106897720A (en) * | 2017-01-11 | 2017-06-27 | 济南中维世纪科技有限公司 | A kind of firework detecting method and device based on video analysis |
-
2018
- 2018-03-28 CN CN201810265915.3A patent/CN108648406A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104408745A (en) * | 2014-11-18 | 2015-03-11 | 北京航空航天大学 | Real-time smog scene detection method based on video image |
| CN106897720A (en) * | 2017-01-11 | 2017-06-27 | 济南中维世纪科技有限公司 | A kind of firework detecting method and device based on video analysis |
Non-Patent Citations (4)
| Title |
|---|
| 姚太伟: "图像型火灾烟雾探测技术研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 张广渊等: "《数字图像处理基础及OPENCV实现》", 31 December 2014, 知识产权出版社 * |
| 王俊然等: "基于模式识别的火灾烟雾探测中抗水雾干扰研究", 《长春理工大学学报(自然科学版)》 * |
| 郭红星等: "自适应混合高斯背景建模算法的GPU并行优化", 《第21届全国多媒体技术、第8届全国普适计算、第8届全国人机交互联合学术会议论文集》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109522845A (en) * | 2018-11-19 | 2019-03-26 | 国网四川省电力公司电力科学研究院 | Distribution transformer based on intelligent robot tests safety work measure of supervision |
| CN109637073A (en) * | 2019-01-08 | 2019-04-16 | 重庆水利电力职业技术学院 | A kind of fire hazard aerosol fog identifying system based on optical flow method |
| CN110796822A (en) * | 2019-09-15 | 2020-02-14 | 杭州拓深科技有限公司 | Method for prejudging fire point direction based on improved smoke alarm |
| CN110796822B (en) * | 2019-09-15 | 2021-07-20 | 杭州拓深科技有限公司 | Method for prejudging fire point direction based on improved smoke alarm |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106600888B (en) | Automatic forest fire detection method and system | |
| CN109165577B (en) | Early forest fire detection method based on video image | |
| CN109637068A (en) | Intelligent pyrotechnics identifying system | |
| CN108648406A (en) | A kind of indoor smog detection alarm system based on machine vision | |
| CN108389359B (en) | Deep learning-based urban fire alarm method | |
| CN106897720A (en) | A kind of firework detecting method and device based on video analysis | |
| CN103778418A (en) | Mountain fire image identification method of image monitoring system of electric transmission line tower | |
| JP5218906B2 (en) | Smoke detection device and smoke detection method | |
| CN111899452A (en) | Forest fire prevention early warning system based on edge calculation | |
| CN107679471A (en) | Indoor occupant sky hilllock detection method based on video monitoring platform | |
| CN110334660A (en) | A kind of forest fire monitoring method based on machine vision under the conditions of greasy weather | |
| CN201259717Y (en) | Fire disaster probe system | |
| CN102881106A (en) | Dual-detection forest fire identification system through thermal imaging video and identification method thereof | |
| CN107301375A (en) | A kind of video image smog detection method based on dense optical flow | |
| CN108492517A (en) | A kind of fire monitoring system for building | |
| CN107688793A (en) | A kind of outside transformer substation fire automatic monitoring method for early warning | |
| CN111476964A (en) | Remote forest fire prevention monitoring system and method | |
| KR101196678B1 (en) | Real-time fire detection device and method | |
| CN201091014Y (en) | Fire detecting device | |
| CA3081967C (en) | Method and system for connected advanced flare analytics | |
| CN108830161A (en) | Smog recognition methods based on video stream data | |
| CN102999994A (en) | Flame detection device based on video image analysis | |
| CN110930632A (en) | Early warning system based on artificial intelligence | |
| CN103020588A (en) | Flame detection method based on video image analysis | |
| CN107967448B (en) | Early fire smoke real-time detection method and system |
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 |
Application publication date: 20181012 |
|
| RJ01 | Rejection of invention patent application after publication |