CN110276228A - Multiple features fusion video fire hazard recognizer - Google Patents
Multiple features fusion video fire hazard recognizer Download PDFInfo
- Publication number
- CN110276228A CN110276228A CN201810208155.2A CN201810208155A CN110276228A CN 110276228 A CN110276228 A CN 110276228A CN 201810208155 A CN201810208155 A CN 201810208155A CN 110276228 A CN110276228 A CN 110276228A
- Authority
- CN
- China
- Prior art keywords
- image
- flame
- fire
- doubtful
- value
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- 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
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Fire-Detection Mechanisms (AREA)
Abstract
The invention discloses a kind of multiple features fusion video fire hazard recognizers.The target of doubtful flame in Infrared video image is obtained first with improved adaptive RTS threshold adjustment model, eliminate background interference, the static state and behavioral characteristics of flame are analyzed again, guarantee the accuracy of feature extraction, finally the flame multiple features fusion method based on step analysis carries out fire identification, the algorithm can effectively improve the recognition efficiency of flame, guarantee the real-time accurate judgement to fire.
Description
Technical field
The present invention relates to fire-fighting equipment technical field more particularly to a kind of multiple features fusion video fire hazard recognizers.
Background technique
Fire brings bright and warm to the mankind, while also promoting social progress.However, with the development of society,
The use of fire also constantly upgrades to mankind's bring disaster, and fire has become most serious in the world today and threatens human survival and hair
One of the normal hair disaster of exhibition has the characteristics that occurrence frequency is high, space-time span is big, be easy to cause a large amount of property loss and sternly
The personal injury of weight.For a long time, the mankind never stop the research to fire detection technology, and non-contact fire detection is wherein one
The highly important technology of item, but conventional contactless fire detector detection range is small, it is affected by environment larger, it is based on video
The fire defector of image can effectively avoid the above problem, however the presence of existing video flame detection algorithm is easy by complicated field
The problems such as influence of scape, illumination condition, the real-time of algorithm is not poor high with reliability, is easy to produce erroneous judgement and fails to judge, so that
Related algorithm less effective in actual use.
Summary of the invention
The purpose of the present invention is in view of the deficiencies of the prior art, and proposes a kind of multiple features fusion video fire hazard identification and calculate
Method can effectively improve the recognition efficiency of flame, guarantee the real-time accurate judgement to fire.
To achieve the above object, the invention adopts the following technical scheme: a kind of multiple features fusion video fire hazard recognizer,
It includes following processing step: S1: acquiring the infrared video information of monitoring area in real time by CCD camera and optical filter;It is described
Before optical filter is sheathed on CCD camera camera lens;CCD camera is connected by USB data line with computer after described, in real time acquisition monitoring
The video information in region;S2: the adaptive background that computer is combined with Background difference using video image interframe calculus of finite differences is more
New method extracts doubtful conflagration area;S3: computer is filtered doubtful flame region, edge enhances, opening operation is located in advance
Reason;S4: the static state of flame is carried out to doubtful flame region and behavioral characteristics extract;S5: using analytic hierarchy process (AHP) to the more of fire
A flame characteristic carries out weight assignment;S6: being weighted the doubtful probability of flame to multiple features, then with it is set by user
Flame characteristic overall situation assessed value is compared, and when doubtful probability is greater than feature overall situation assessed value, assert fire, otherwise not
Fire occurs.
In an embodiment of the present invention, the cutoff wavelength of the optical filter is 800nm.
In an embodiment of the present invention, the adaptive RTS threshold adjustment method process is as follows: BS1: opening up two pieces of memories point
It Wei not dynamic memory, static memory;BS2: starting Image Acquisition, defines image sequence count parameter j, acquires the 1st width image, sentences
Whether disconnected j is equal to 1, and static memory is arrived in storage if equal, as background template, the otherwise next 3 width figure of Coutinuous store
As into dynamic memory;BS3: it is poor that each image in dynamic memory and background template image are made;BS4: to gray scale difference value image
Carry out binaryzation, edge enhancing pretreatment;BS5: by obtaining gray scale difference after dynamic memory image and background template image subtraction
The pixel value of value image is compared to calculate the size of doubtful conflagration area, then with presetting pixel threshold G1, threshold value G2,
Then context update is carried out according to comparison result or alarm is the doubtful area of fire.
Further, step BS5 is comprising the following specific steps if the 1st width image of dynamic memory and background template figure
Pixel bj1 as obtaining gray scale difference value image after subtracting each other is less than G1, then proving do not have fire in monitor area, this time
Judgement terminates, and continues to read next picture and is judged;If bj1 is greater than G1, the storage of this picture into dynamic memory, after
The 2nd width image of continuous acquisition;2nd width image and background template image are subtracted each other to obtain the pixel value difference bj2 of gray scale difference value image
Less than G1, any response is not done, continues to read next, if difference is greater than G1, is stored in another dynamic memory;
If the 1st width image and the 2nd width image are consecutive images in dynamic memory, the 2nd width image and background template image are made
Pixel value bj2 after difference is recorded;Alarm immediately is the doubtful area of fire if discontinuous;The 3rd is judged in the same way
Width image whether with first two it is continuous, if continuous the 3rd width image with subtracted each other to obtain gray scale with background template image
The pixel value difference bj3 of error image is recorded, and otherwise alarm is the doubtful area of fire;The 3rd width image and the 2nd width image, the 2nd
Width image and the 1st width image make difference respectively, and record the pixel value after making difference;If the two values are both greater than threshold value G2,
Using the 3rd width image is stored in dynamic memory as background template image, original background image, no in Lai Gengxin static memory
Then alarm is the doubtful area of fire.
The static state for extracting flame described in an embodiment of the present invention and behavioral characteristics include: circularity, Sharp features, matter
Heart movement, similarity, roughness, area increase.
In an embodiment of the present invention, weight assignment procedure of the analytic hierarchy process (AHP) to multiple flame characteristics are as follows: CS1:
Flame characteristic importance assessment table is established, and flame characteristic qualitative rule judgment matrix is defined according to assessment table;CS2: calculating is sentenced
The consistency ration of disconnected matrix, it is that judgment matrix meets consistency check condition which, which is less than specified value, otherwise adjustment assessment table
Numerical value repeats CS1 and CS2 step, until meeting test condition;CS3: the corresponding spy of judgment matrix maximum eigenvalue is sought
Vector is levied, and obtains weight vector after being standardized, the weight of each numerical value in weight vector and multiple flame characteristics is one by one
It is corresponding.
Further, CS2 is the following steps are included: the maximum eigenvalue for calculating judgment matrix A is λmax;Calculate judgment matrix A
Coincident indicator CI:
N is matrix dimension, obtains the one of judgment matrix A by inquiring N-dimensional vector Aver-age Random Consistency Index inquiry table
Cause property standard RI;Calculate the consistency ration CR of judgment matrix A:
If when CR < 0.1, it is believed that matrix A meets test condition, otherwise adjustment assessment table numerical value, repeats CS1 and CS2 step
Suddenly, until meeting test condition.
In an embodiment of the present invention, the calculating process of the doubtful probability of the flame are as follows: DS1: by each feature u of flame
A logger I (u) is matched, when video image Flame meets some flame characteristic, otherwise it is 0 that I (u), which is 1,;DS2: every
The weight of a flame characteristic is added after being multiplied with corresponding logger value, obtains the doubtful probability value of flame.
Compared with prior art, the invention has the following advantages that
1. by carrying out adaptive RTS threshold adjustment, flame multi-feature extraction to monitoring area video and based on step analysis
Flame multiple features fusion method carries out fire disaster flame identification, realizes that video fire hazard rapidly and efficiently identifies, can when fire occurs
And alarm, it effectively reduces property loss and avoids casualties;
2. eliminating influence of the extraneous background to fire disaster flame by improved adaptive RTS threshold adjustment model, improving may
The integrity degree and accuracy that flame region extracts, provide more accurate video image information for subsequent fire identification;
3. extracting the multiple static and behavioral characteristics of flame, flame information degree of covering height;
4. relevant fire expertise quantify and move multiple features of fire using analytic hierarchy process (AHP)
State assigns weight, and final to provide the decision-making foundation of quantitative terms for fire differentiation, this method is simple and effective, real-time is good;
5. background interference can be effectively reduced by using infrared fileter, and then improve fire identification efficiency.
Detailed description of the invention
Fig. 1 is fire identification process figure in the multiple features fusion video fire hazard recognizer of the embodiment of the present invention.
Fig. 2 is that Adaptive background subtraction updates in the multiple features fusion video fire hazard recognizer of the embodiment of the present invention and fire is doubted
Like the extraction flow chart in region.
Specific embodiment
Explanation is further explained to the present invention in the following with reference to the drawings and specific embodiments.
The present invention provides a kind of multiple features fusion video fire hazard recognizer, and main flow schematic diagram is referring to Fig. 1.It includes
Following steps:
S1: the video information of monitoring area is acquired in real time by CCD camera and optical filter.
Specifically, the infrared fileter for being 800nm by using cutoff wavelength, the optical filter set that it is adapted with size
Ring is fixed in CCD camera.Then CCD camera is connected by USB data line with computer, can acquire monitoring area in real time
Video information;
S2: the adaptive RTS threshold adjustment method that is combined using frame differential method with Background difference extracts doubtful fire zone
Domain;Main flow schematic diagram is referring to fig. 2.
Specifically, steps are as follows:
BS1: opening up two pieces of memories is respectively dynamic memory, static memory;
BS2: starting Image Acquisition, defines image sequence count parameter j, acquires the 1st width image, judges whether j is equal to 1,
If equal storage arrive static memory, be denoted as BT1 as background template, otherwise the next 3 width image of Coutinuous store to move
RT is denoted as in state memoryi, wherein i=1,2,3;
BS3: it is poor that each image in dynamic memory and background template image are made, and obtains gray scale difference value image CZi:
CZi=| RTi-BT1|
BS4: to gray scale difference value image CZiCarry out binaryzation, edge enhancing pretreatment;
BS5: it is counted by obtaining the pixel bj1 of gray scale difference value image after dynamic memory image and background template image subtraction
Calculate the size of doubtful conflagration area, then be compared with presetting pixel threshold G1, threshold value G2, then according to comparison result into
Row context update or alarm are the doubtful area of fire.
Specifically, steps are as follows:
If bj1 is less than G1, prove there is no fire in monitor area, this time judgement terminates, and continues to read
Next picture is judged;If bj1 is greater than G1, the storage of this picture into dynamic memory, continue to acquire the 2nd width image,
With background template image carry out additive operation, if pixel value difference be less than G1, it may be possible to moment have infrared source enter or
The sunlight of mirror-reflection enters monitoring area, does not do any response, continues to read next, if difference is greater than G1, deposits
Storage is in another dynamic memory.If the 1st width image and the 2nd width image are consecutive images in dynamic memory, the 2nd
Width image is recorded with the pixel value bj2 after background template image work difference.Alarm immediately is that fire is doubtful if discontinuous
Area, it may be possible to caused by the area fluctuating change of flame;Judge in the same way the 3rd width image whether with first two it is continuous,
3rd width image and the pixel value bj3 after background image work difference are recorded if continuous, otherwise alarm is that fire is doubtful
Area;
The 3rd width image and the 2nd width image, the 2nd width image and the 1st width image make difference respectively, and record the picture after making difference
Element value.If the two values are both greater than threshold value G2, using being stored in dynamic memory the 3rd width image as background template figure
Picture, original background image in Lai Gengxin static memory, otherwise alarm is the doubtful area of fire.
S3: fire suspicious region is filtered, edge enhancing, opening operation pretreatment;
S4: the static state of flame is carried out to doubtful flame region and behavioral characteristics extract;
Specifically, flame static state and behavioral characteristics include the following: circularity, Sharp features, mass center movement, similarity,
Roughness, area increase.
S5: weight assignment procedure is carried out using multiple features of the analytic hierarchy process (AHP) to fire are as follows:
CS1: flame characteristic importance assessment table is established, and flame characteristic qualitative rule is defined according to assessment table and judges square
Battle array;
Specifically, according to the research experience about fire, show that flame characteristic importance assessment table is as follows:
Flame characteristic | Circularity | Sharp features | Mass center is mobile | Similarity | Roughness | Area increases |
Circularity | 1 | 3 | 3 | 3 | 5 | 5 |
Sharp features | 1/3 | 1 | 3/2 | 2 | 3 | 3 |
Mass center is mobile | 1/3 | 2/3 | 1 | 3/2 | 2 | 3 |
Similarity | 1/3 | 1/2 | 2/3 | 1 | 3/2 | 2 |
Roughness | 1/5 | 1/3 | 1/2 | 2/3 | 1 | 1 |
Area increases | 1/5 | 1/3 | 1/3 | 1/2 | 1 | 1 |
Then it is as follows judgment matrix A to be converted by importance assessment table:
CS2: calculating the consistency ration of judgment matrix, and it is that judgment matrix meets consistency check item which, which is less than specified value,
Part, otherwise adjustment assessment table numerical value, repeats CS1 and CS2 step, until meeting test condition;
Specifically, the maximum eigenvalue for calculating judgment matrix A is λmax=6.0685;
Calculate the coincident indicator CI of judgment matrix A:
The consistency criterion RI=of judgment matrix A is obtained by inquiring N-dimensional vector Aver-age Random Consistency Index inquiry table
1.24。
N (matrix dimension) | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 |
RI | 0 | 0 | 0.58 | 0.90 | 1.12 | 1.24 | 1.32 | 1.41 |
Calculate the consistency ration CR of judgment matrix A:
If when CR < 0.1, it is believed that the inconsistent degree of matrix A can receive, and otherwise adjustment assessment table numerical value, repeats CS1
With CS2 step, until meeting test condition;
CS3: acquiring the corresponding feature vector of judgment matrix maximum eigenvalue, and obtains weight vector after being standardized,
The weight of each numerical value and multiple flame characteristics in weight vector corresponds.
Specifically, acquire judgment matrix A the corresponding feature vector of maximum eigenvalue be V=[0.815,0.3937,
0.2256,0.1432,0.1286], then feature vector is standardized, is had: V=[0.4053,0.1958,0.1515,
0.1122,0.0712,0.0640].The result for assigning weight automatically based on analytic hierarchy process (AHP) is as follows: circularity weight is 0.4053,
Fire angle weight is 0.1958, and the mobile weight of mass center is 0.1515, flame roughness weight 0.1122, flame similarity weight
It is 0.0712, it is 0.0640 that area, which increases weight,.
S6: the doubtful probability of flame is weighted to multiple features, is then commented with the flame characteristic overall situation set by user
Valuation is compared, and when doubtful probability is greater than feature overall situation assessed value, assert fire, fire does not otherwise occur.
Specifically, the calculating process of the doubtful probability of flame is as follows:
DS1: each feature u of flame is matched into a logger I (u), when video image Flame meets some flame
When feature, otherwise it is 0 that I (u), which is 1,;
DS2: the weight W of each flame characteristic be multiplied with corresponding logger value I after be added, it is doubtful general to obtain flame
Rate value IF:
When the doubtful probability of flame is greater than feature overall situation assessed value, assert fire, fire does not otherwise occur.
The infrared fileter can reduce fire hazard environment background interference, and then improve fire identification efficiency.
The flame characteristic overall situation assessed value can be set according to practical situations, so that this algorithm has well
Adaptability.
In conclusion the present invention provides the multiple features fusion video fire hazard identifications that one kind can efficiently carry out fire detection
Algorithm.
Although the present invention is described referring to multiple embodiments, the present invention is not limited to above-described embodiments, answer
When understand those skilled in the art the component involved in above-described embodiment can be carried out it is appropriate combination form new embodiment,
The various apparent modifications and variations carried out on the basis of not departing from the principle of the invention should all fall into protection scope of the present invention
Within.
Claims (8)
1. a kind of multiple features fusion video fire hazard recognizer, which is characterized in that including following processing step:
S1: the infrared video information of monitoring area is acquired in real time by CCD camera and optical filter;The optical filter is sheathed on CCD
Before camera lens;CCD camera is connected by USB data line with computer after described, acquires the video information of monitoring area in real time;
S2: adaptive RTS threshold adjustment method that computer is combined using video image interframe calculus of finite differences with Background difference is extracted
Doubtful conflagration area;
S3: computer is filtered doubtful flame region, edge enhances, opening operation pretreatment;
S4: the static state of flame is carried out to doubtful flame region and behavioral characteristics extract;
S5: weight assignment is carried out using multiple flame characteristics of the analytic hierarchy process (AHP) to fire;
S6: being weighted the doubtful probability of flame to multiple features, then with flame characteristic overall situation assessed value set by user
It is compared, when doubtful probability is greater than feature overall situation assessed value, assert fire, fire does not otherwise occur.
2. multiple features fusion video fire hazard recognizer according to claim 1, it is characterised in that: the optical filter is cut
Only wavelength is 800nm.
3. multiple features fusion video fire hazard recognizer according to claim 1, it is characterised in that: the adaptive background
Update method process is as follows:
BS1: opening up two pieces of memories is respectively dynamic memory, static memory;
BS2: starting Image Acquisition, defines image sequence count parameter j, acquires the 1st width image, judges whether j is equal to 1, if
Static memory is arrived in equal then storage, and as background template, otherwise the next 3 width image of Coutinuous store is into dynamic memory;
BS3: it is poor that each image in dynamic memory and background template image are made;
BS4: binaryzation, edge enhancing pretreatment are carried out to gray scale difference value image;
BS5: doubtful to calculate by obtaining the pixel value of gray scale difference value image after dynamic memory image and background template image subtraction
It is compared like the size of conflagration area, then with presetting pixel threshold G1, threshold value G2, is then carried on the back according to comparison result
Scape updates or alarm is the doubtful area of fire.
4. multiple features fusion video fire hazard recognizer according to claim 3, it is characterised in that: step BS5 include with
Lower specific steps: if obtaining the pixel of gray scale difference value image after the 1st width image of dynamic memory and background template image subtractionbj1 is less thanG1, then prove do not have fire in monitor area, this time judgement terminates, continue to read next picture into
Row judgement;Ifbj1 is greater than G1, the storage of this picture into dynamic memory, continues to acquire the 2nd width image;
2nd width image and background template image are subtracted each other to obtain the pixel value difference of gray scale difference value imagebj2 are less thanG1, do not appoint
What is responded, and continues to read next, if difference is greater thanG1, then it is stored in another dynamic memory;If in dynamic
The 1st width image and the 2nd width image are consecutive images in depositing, then the 2nd width image and background template image are made the pixel value after differencebj2 record;Alarm immediately is the doubtful area of fire if discontinuous;Judge in the same way the 3rd width image whether with
First two continuous, is subtracted each other the 3rd width image and with background template image to obtain the picture of gray scale difference value image if continuous
Plain differencebj3 record, and otherwise alarm is the doubtful area of fire;
The 3rd width image and the 2nd width image, the 2nd width image and the 1st width image make difference respectively, and record the pixel value after making difference;
If the two values are both greater than threshold valueG2, then coming more using the 3rd width image in dynamic memory is stored in as background template image
Original background image in new static memory, otherwise alarm is the doubtful area of fire.
5. multiple features fusion video fire hazard recognizer according to claim 1, it is characterised in that: the extraction flame
Static and behavioral characteristics include: circularity, Sharp features, mass center movement, similarity, roughness, area growth.
6. multiple features fusion video fire hazard recognizer according to claim 1, it is characterised in that: the analytic hierarchy process (AHP)
To the weight assignment procedure of multiple flame characteristics are as follows:
CS1: flame characteristic importance assessment table is established, and flame characteristic qualitative rule judgment matrix A is defined according to assessment table;
CS2: calculating the consistency ration of judgment matrix, and it is that judgment matrix meets consistency check condition which, which is less than specified value,
Otherwise adjustment assessment table numerical value, repeats CS1 and CS2 step, until meeting test condition;
CS3: seeking the corresponding feature vector of judgment matrix maximum eigenvalue, and weight vector, Quan Xiang are obtained after being standardized
The weight of each numerical value and multiple flame characteristics in amount corresponds.
7. multiple features fusion video fire hazard recognizer according to claim 6, it is characterised in that: CS2 includes following step
It is rapid:
Calculate judgment matrix A maximum eigenvalue be;
Calculate the coincident indicator of judgment matrix ACI:
NFor matrix dimension, pass through inquiryNDimensional vector Aver-age Random Consistency Index inquiry table obtains judgment matrixAConsistency
StandardRI;Calculate the consistency ration of judgment matrix ACR:
If when CR < 0.1, it is believed that matrix A meets test condition, and otherwise adjustment assessment table numerical value, repeats CS1 and CS2 step, directly
Until meeting test condition.
8. multiple features fusion video fire hazard recognizer according to claim 1, which is characterized in that the flame is doubtful general
The calculating process of rate are as follows:
DS1: each feature u of flame is matched into a loggerI(u), when video image Flame meets some flame characteristic
When,I(u) it is 1, is otherwise 0;
DS2: the weight of each flame characteristic is added after being multiplied with corresponding logger value, obtains the doubtful probability value of flame.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810208155.2A CN110276228B (en) | 2018-03-14 | 2018-03-14 | Multi-feature fusion video fire disaster identification method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810208155.2A CN110276228B (en) | 2018-03-14 | 2018-03-14 | Multi-feature fusion video fire disaster identification method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110276228A true CN110276228A (en) | 2019-09-24 |
CN110276228B CN110276228B (en) | 2023-06-20 |
Family
ID=67958258
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810208155.2A Active CN110276228B (en) | 2018-03-14 | 2018-03-14 | Multi-feature fusion video fire disaster identification method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110276228B (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111710125A (en) * | 2020-05-26 | 2020-09-25 | 上饶市中科院云计算中心大数据研究院 | Intelligent scenic spot fire prevention early warning method and system |
WO2021077472A1 (en) * | 2019-10-22 | 2021-04-29 | 北京航天常兴科技发展股份有限公司 | Comprehensive pyrolysis particle electrical fire monitoring method, device and system |
CN113516091A (en) * | 2021-07-27 | 2021-10-19 | 福建工程学院 | Method for identifying electric spark image of transformer substation |
CN115359615A (en) * | 2022-08-15 | 2022-11-18 | 北京飞讯数码科技有限公司 | Indoor fire alarm early warning method, system, device, equipment and medium |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20110064264A1 (en) * | 2008-05-08 | 2011-03-17 | Utc Fire & Security | System and method for video detection of smoke and flame |
CN105740866A (en) * | 2016-01-22 | 2016-07-06 | 合肥工业大学 | Rotary kiln sintering state recognition method with artificial feedback regulation mechanism |
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 |
-
2018
- 2018-03-14 CN CN201810208155.2A patent/CN110276228B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20110064264A1 (en) * | 2008-05-08 | 2011-03-17 | Utc Fire & Security | System and method for video detection of smoke and flame |
CN105740866A (en) * | 2016-01-22 | 2016-07-06 | 合肥工业大学 | Rotary kiln sintering state recognition method with artificial feedback regulation mechanism |
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 |
Non-Patent Citations (1)
Title |
---|
金肖: "多特征融合视频火灾识别研究", 《机械制造与自动化》 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2021077472A1 (en) * | 2019-10-22 | 2021-04-29 | 北京航天常兴科技发展股份有限公司 | Comprehensive pyrolysis particle electrical fire monitoring method, device and system |
CN111710125A (en) * | 2020-05-26 | 2020-09-25 | 上饶市中科院云计算中心大数据研究院 | Intelligent scenic spot fire prevention early warning method and system |
CN113516091A (en) * | 2021-07-27 | 2021-10-19 | 福建工程学院 | Method for identifying electric spark image of transformer substation |
CN113516091B (en) * | 2021-07-27 | 2024-03-29 | 福建工程学院 | Method for identifying electric spark image of transformer substation |
CN115359615A (en) * | 2022-08-15 | 2022-11-18 | 北京飞讯数码科技有限公司 | Indoor fire alarm early warning method, system, device, equipment and medium |
CN115359615B (en) * | 2022-08-15 | 2023-08-04 | 北京飞讯数码科技有限公司 | Indoor fire alarm early warning method, system, device, equipment and medium |
Also Published As
Publication number | Publication date |
---|---|
CN110276228B (en) | 2023-06-20 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109934176B (en) | Pedestrian recognition system, recognition method, and computer-readable storage medium | |
CN110276228A (en) | Multiple features fusion video fire hazard recognizer | |
EP3869459B1 (en) | Target object identification method and apparatus, storage medium and electronic apparatus | |
US9639769B2 (en) | Liveness detection | |
CN110674748B (en) | Image data processing method, apparatus, computer device, and readable storage medium | |
CN111626188B (en) | Indoor uncontrollable open fire monitoring method and system | |
TWI505200B (en) | Method,system,computer program product and process for determining parts and associated attributes of an object in video | |
CN111062429A (en) | Chef cap and mask wearing detection method based on deep learning | |
CN105389562A (en) | Secondary optimization method for monitoring video pedestrian re-identification result based on space-time constraint | |
CN113807276A (en) | Smoking behavior identification method based on optimized YOLOv4 model | |
CN105022999A (en) | Man code company real-time acquisition system | |
CN107067412A (en) | A kind of video flame smog detection method of Multi-information acquisition | |
CN109934047A (en) | Face identification system and its face identification method based on deep learning | |
CN111401310B (en) | Kitchen sanitation safety supervision and management method based on artificial intelligence | |
CN109325462A (en) | Recognition of face biopsy method and device based on iris | |
CN111862413A (en) | Method and system for realizing epidemic situation resistant non-contact multidimensional identity rapid identification | |
CN103903020B (en) | A kind of fire image recognition methods and device based on CodeBook | |
CN111639602A (en) | Pedestrian shielding and orientation detection method | |
CN112541403A (en) | Indoor personnel falling detection method utilizing infrared camera | |
JPWO2018173947A1 (en) | Image retrieval device | |
CN115937971B (en) | Method and device for identifying hand-lifting voting | |
WO2023093241A1 (en) | Pedestrian re-identification method and apparatus, and storage medium | |
CN106557523A (en) | Presentation graphics system of selection and equipment and object images search method and equipment | |
Yang et al. | Intelligent collection of rice disease images based on convolutional neural network and feature matching | |
CN115170059A (en) | Intelligent safety monitoring system for outdoor construction site and working method |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |