CN110533872A - A kind of fire alarm method for compartment - Google Patents
A kind of fire alarm method for compartment Download PDFInfo
- Publication number
- CN110533872A CN110533872A CN201810560772.9A CN201810560772A CN110533872A CN 110533872 A CN110533872 A CN 110533872A CN 201810560772 A CN201810560772 A CN 201810560772A CN 110533872 A CN110533872 A CN 110533872A
- Authority
- CN
- China
- Prior art keywords
- image
- flare
- smog
- doubtful
- segmentation
- 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
- 238000000034 method Methods 0.000 title claims abstract description 48
- 230000011218 segmentation Effects 0.000 claims abstract description 22
- 238000000605 extraction Methods 0.000 claims abstract description 11
- 239000000779 smoke Substances 0.000 claims description 21
- 238000012549 training Methods 0.000 claims description 8
- 238000010606 normalization Methods 0.000 claims description 7
- 238000002474 experimental method Methods 0.000 claims description 3
- 239000000284 extract Substances 0.000 claims description 3
- 238000003384 imaging method Methods 0.000 claims description 3
- 238000001514 detection method Methods 0.000 abstract description 15
- 235000019504 cigarettes Nutrition 0.000 description 10
- 230000008569 process Effects 0.000 description 7
- 230000008859 change Effects 0.000 description 6
- 238000012544 monitoring process Methods 0.000 description 5
- 238000004422 calculation algorithm Methods 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 239000003595 mist Substances 0.000 description 3
- 239000000443 aerosol Substances 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000009434 installation Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 230000001960 triggered effect Effects 0.000 description 2
- 230000002776 aggregation Effects 0.000 description 1
- 238000004220 aggregation Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000007493 shaping process Methods 0.000 description 1
- 230000001360 synchronised effect Effects 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
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
-
- 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
-
- 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
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B17/00—Fire alarms; Alarms responsive to explosion
- G08B17/12—Actuation by presence of radiation or particles, e.g. of infrared radiation or of ions
- G08B17/125—Actuation by presence of radiation or particles, e.g. of infrared radiation or of ions by using a video camera to detect fire or smoke
-
- 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
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
- G06V10/467—Encoded features or binary features, e.g. local binary patterns [LBP]
Abstract
The invention discloses a kind of fire alarm methods for compartment, which comprises acquisition compartment inner video image；Image pre-segmentation is carried out to the video image, obtains doubtful flare image；Confidence declaration is carried out to the doubtful flare image based on flare anticipation Gauss model；When the doubtful flare image passes through the confidence declaration, flare texture feature extraction is extracted from the doubtful flare image；Judge that current scene with the presence or absence of flare, issues early warning when there are flare according to the flare textural characteristics.Method according to the invention it is possible to carry out early detection to the fire in compartment；Compared to the prior art, fire alarm of the invention not only more quick and precisely, but also has strong anti-interference ability.
Description
Technical field
The present invention relates to field of track traffic, and in particular to a kind of fire alarm method for compartment.
Background technique
In various disasters, fire be most frequently, most generally threaten public security and social development major casualty it
One.Especially, in field of track traffic, due to the closure of compartment, once extensive fire occurs, caused by life wealth
Producing loss is even more that can not estimate.But conventionally, as not can guarantee vehicle hardware facility is in ideal shape always
State cannot achieve perfect monitoring along with passenger is numerous, this results in the inducement for preventing to cause fire that can not be absolute.
Therefore, fire alarm is carried out in time when fire is just induced just into the guarantee personnel security of the lives and property, prevention fire disaster
One important means of accident.
In the prior art, fire alarm usually is realized using smoke detector in compartment.But due to train vehicle
Compartment space is larger, and the installation site of smoke detector is higher, and smog needs to reach a certain concentration, and the ability that reach a certain height
Smoke alarm is triggered, is easily caused for fire judgement not in time；Also, the passenger due to not complying with the rules in compartment smokes
Cause triggering alarm that can also cause mistake early warning；Simultaneously smoke alarm can quickly attract crew's note that still without
Method quickly informs the specific location of fire hazard aerosol fog.Therefore, the fire not being able to satisfy in compartment using smoke detector is pre-
It is alert to require.
Summary of the invention
The present invention provides a kind of fire alarm methods for compartment, which comprises
Acquire compartment inner video image；
Image pre-segmentation is carried out to the video image, obtains doubtful flare image；
Confidence declaration is carried out to the doubtful flare image based on flare anticipation Gauss model；
When the doubtful flare image passes through the confidence declaration, flare line is extracted from the doubtful flare image
Manage feature extraction；
Judge that current scene with the presence or absence of flare, issues early warning when there are flare according to the flare textural characteristics.
In one embodiment, based on the comparison with fixed background, image pre-segmentation is carried out by background subtraction.
In one embodiment, image pre-segmentation is carried out to the video image, obtains doubtful flare image, comprising:
Judge whether the video image includes that color and/or luminance information meet flare color and/or luminance information
Region.
In one embodiment, image pre-segmentation is carried out to the video image, obtains doubtful flare image, in which:
When there are the doubtful flare image, the fixed background is not updated
When the doubtful flare image is not present, fixed background described in real-time update.
In one embodiment, the method also includes:
Flare sample image is obtained, according to the flare sample image training acquisition flare anticipation model and/or really
The fixed confidence threshold value, the flare sample image includes positive sample and/or negative sample, in which:
Acquisition flare image simultaneously marks the region for picking out flare in every frame image, generates the positive sample after size normalization
This；
And/or
Model is prejudged using flare described in flare imaging experiments, obtains the image for being not recognized as flare, mark is picked out
The region of flare in every frame image generates the negative sample after size normalization.
In one embodiment, flare texture feature extraction is carried out for the doubtful flare image, in which:
Utilize the textural characteristics of LBP operator extraction image.
In one embodiment, the method also includes:
Image pre-segmentation is carried out to the video image, obtains doubtful smog image；
Smoke characteristics are extracted from the doubtful smog image；
Judge that current scene with the presence or absence of smog, issues early warning when there are smog according to the smoke characteristics.
In one embodiment, image pre-segmentation is carried out to the video image, obtains doubtful smog image, comprising:
Judge whether the video image includes region that dark channel value meets smog dark model；
Judge that dark channel value meets the region of smog dark model and whether meets smog dynamic characteristic.
In one embodiment, the method also includes:
Obtain smog sample image, according to smog sample image training obtain the smog dark model and/or
The smog dynamic characteristic.
In one embodiment, smoke characteristics are extracted from the doubtful smog image, wherein calculate the doubtful smog figure
As the entropy and profile gray level co-occurrence matrixes of before and after frames.
Method according to the invention it is possible to carry out early detection to the fire in compartment；Compared to the prior art, of the invention
Fire alarm not only more quick and precisely, but also have strong anti-interference ability.
Other feature or advantage of the invention will illustrate in the following description.Also, Partial Feature of the invention or
Advantage will be become apparent by specification, or be appreciated that by implementing the present invention.The purpose of the present invention and part
Advantage can be realized or be obtained by step specifically noted in the specification, claims and drawings.
Detailed description of the invention
Attached drawing is used to provide further understanding of the present invention, and constitutes part of specification, with reality of the invention
It applies example and is used together to explain the present invention, be not construed as limiting the invention.In the accompanying drawings:
Fig. 1 is method flow diagram according to an embodiment of the invention；
Fig. 2 is method partial process view according to an embodiment of the invention；
Fig. 3 is system structure schematic diagram according to an embodiment of the invention.
Specific embodiment
Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings and examples, implementation personnel of the invention whereby
Can fully understand that how the invention applies technical means to solve technical problems, and reach technical effect realization process and according to
The present invention is embodied according to above-mentioned realization process.As long as each embodiment it should be noted that do not constitute conflict, in the present invention
And each feature in each embodiment can be combined with each other, be formed by technical solution protection scope of the present invention it
It is interior.
In the prior art, fire alarm usually is realized using smoke detector in compartment.But due to train vehicle
Compartment space is larger, and the installation site of smoke detector is higher, and smog needs to reach a certain concentration, and the ability that reach a certain height
Smoke alarm is triggered, is easily caused for fire judgement not in time；Also, the passenger due to not complying with the rules in compartment smokes
Cause triggering alarm that can also cause mistake early warning；Simultaneously smoke alarm can quickly attract crew's note that still without
Method quickly informs the specific location of fire hazard aerosol fog.Therefore, the fire not being able to satisfy in compartment using smoke detector is pre-
It is alert to require.
For fire alarm process the defects of of the prior art in compartment, the invention proposes one kind for arranging
The fire alarm method in vehicle compartment.Method of the invention, can be right in such a way that video monitoring and image procossing combine
Fire in compartment carries out early detection.Further, in the method for the invention, it by the way of multiple-authentication, compares
In the prior art, fire alarm of the invention not only more quick and precisely, but also is had strong anti-interference ability.
Next the implementation process based on flow chart the present invention is described in detail embodiment.It is walked shown in the flow chart of attached drawing
Suddenly it can be executed in the computer system comprising such as a group of computer-executable instructions.Although showing in flow charts each
The logical order of step, but in some cases, it can be with the steps shown or described are performed in an order that is different from the one herein.
As shown in Figure 1, in one embodiment, fire alarm method includes the following steps.
It acquires compartment inner video image (S110)；
Image pre-segmentation is carried out to video image, obtains doubtful flare image (S120)；
Confidence declaration (S130) is carried out to doubtful flare image based on flare anticipation Gauss model；
When doubtful flare image passes through confidence declaration, flare textural characteristics are extracted from doubtful flare image
(S140)；
Judge that current scene with the presence or absence of flare (S150), issues early warning when there are flare according to flare textural characteristics
(S151)。
Specifically, in one embodiment, in the step s 120, it includes a determination steps in logic, i.e., sentence first
Whether include doubtful flare image in disconnected video image, continue video monitoring if not including, if will be doubted comprising if
It is split from video image like flare image.
Further, in one embodiment, the judgment step and segmentation step of step S120 is synchronous in execution executes.
The differentiation that doubtful flare image is directly carried out to video image is directly divided if there is meeting doubtful flare image condition
Out.
Further, since in the most of the time, the background image that camera acquires in compartment is all more consistent.Cause
This in the step s 120, based on the comparison with fixed background, carries out image by background subtraction and divides in advance in one embodiment
It cuts.
Specifically, in one embodiment, two field pictures pair are directly compared in the variation between adjacent two frame of detection image sequence
Answer pixel.
Further, in one embodiment, background difference is realized according to flare color (RGB) and the characteristic of brightness (HSV)
Method.Specifically, in one embodiment, judge video image whether include color and/or luminance information meet flare color and/or
The region of luminance information.That is, the RGB relationship and HSV brightness relationship by practical flare image pass through background subtraction when detecting
Point-score carries out dynamical object segmentation, judges whether RGB the and HSV information of dynamic change target (video image) meets flare feature
Condition rejects ineligible region, retains qualified region to realize image pre-segmentation.
Specifically, in one embodiment, the RGB relationship of flare characteristic condition is R >=G >=B.
Specifically, in one embodiment, the HSV brightness relationship of flare characteristic condition is 0 °≤H≤60 °, 0.2≤S≤1,
100≤I≤255。
Further, in one embodiment, in order to improve the validity that background subtraction is implemented, to realization background contrast's
Fixed background is updated.Specifically, in one embodiment, judging whether to be partitioned into doubtful flare image；When there are doubtful fire
When seedling image, fixed background is not updated, is directly determined further；When doubtful flare image is not present, real-time update
Fixed background continues video monitoring.
Further, in one embodiment, in step s 130, based on flare anticipation Gauss model to doubtful flare image
Each of pixel judged, obtain the confidence level of doubtful flare image.It is sentenced to each pixel
It is disconnected, determine that it belongs to the probability of flare image pixel, so that finally obtaining doubtful flare image is flare image on the whole
Confidence level (probability).When the confidence level is more than some threshold value (confidence threshold value), so that it may judge doubtful flare image roughly
For flare image.
Further, in one embodiment, the flare anticipation Gauss model for being prejudged to doubtful flare image is
It is obtained by the training to flare sample image.Specifically, in one embodiment, method further include:
Flare sample image is obtained, flare is obtained according to the training of flare sample image and prejudges model.
Further, in one embodiment, confidence threshold value is determined also according to flare sample image.
Further, in one embodiment, flare sample image includes positive sample.Specifically, in one embodiment, acquisition
Flare image simultaneously marks the region for picking out flare in every frame image, generates the positive sample after size normalization.
Further, in one embodiment, flare sample image includes negative sample.Specifically, in one embodiment, using
Flare described in flare imaging experiments prejudges model, obtains the image for being not recognized as flare, and mark picks out every frame image moderate heat
The region of seedling generates negative sample after size normalization.
Specifically, in one embodiment, by way of the flare image of above-mentioned steps is searched for collection in worksite and/or on the net
It obtains.
Further, in one embodiment, it during extracting flare textural characteristics from doubtful flare image, uses
Local binary patterns (Local Binary Patterns, LBP).Specifically, in one embodiment, utilizing LBP operator extraction
The textural characteristics of image are to obtain flare textural characteristics.Then it is input to pair using the textural characteristics of LBP operator extraction image
The support vector machines (Support Vector Machines, SVM) answered, thus it is exportable whether be flare result.
Specifically, in one embodiment, in LBP operator, the more complicated sub- predicting shape of recurrence has been used, each
Predicting unit uses random tree, and carrys out predicting shape using random forest and change.
Specifically, in one embodiment, being input to corresponding svm classifier using the textural characteristics of LBP operator extraction image
Device, SVM are separated data by construction divisional plane, thus it is exportable whether be flare result.
Further, in one embodiment, it before extracting flare textural characteristics in doubtful flare image, needs to pass through
Flare prejudges Gauss model and carries out size normalization to doubtful flare image.
Further, it is contemplated that in actual scene, the mark of scene of fire not merely includes flare, further includes smog,
In order to improve early warning accuracy, in one embodiment, the invention also provides the processes of smog early warning.Specifically, such as Fig. 2 institute
Show, on the basis of flare identifies early warning, method further include:
Based on image pre-segmentation is carried out to video image is compared with fixed background, doubtful smog image (S260) is obtained；
Smoke characteristics (S270) is extracted from doubtful smog image；
Judge that current scene with the presence or absence of smog (S280), issues early warning when there are smog according to smoke characteristics
(S281)。
Further, in actual scene, since flare feature is more significant.Therefore, in one embodiment, algorithm is preferential
It detects flare and otherwise continues the judgement of cigarette without the monitoring of cigarette when there are flare, this detection mode can be improved
The accuracy rate of fire is detected, false detection rate caused by reducing due to smog interference.
That is, in one embodiment, the image pre-segmentation for flare is carried out to video image first, if there is no doubtful
Flare image or later flare anticipation Gauss model judge or flare textural characteristics judge in result be it is no just progress needle
Image pre-segmentation to smog.
Further, since in the most of the time, the background image that camera acquires in compartment is all more consistent.Cause
This in step S260, based on the comparison with fixed background, carries out image by background subtraction and divides in advance in one embodiment
It cuts.
Specifically, in one embodiment, two field pictures pair are directly compared in the variation between adjacent two frame of detection image sequence
Answer pixel.
Further, in one embodiment, the feature based on smog image realizes that the background difference for smog compares.In
In one embodiment, the feature instantiation of smog image is smog dark feature and smog dynamic characteristic.Specifically, for a width
Clear fog free images are constantly present some pixels, their at least one Color Channels have very low pixel value, these pixel values
Referred to as dark pixel, the Color Channel are also referred to as the dark of the figure.Foggy image is whiter than fog free images, that is, secretly
It is brighter outside channel.Smog dark feature namely refers to that mist is denseer, and dark is whiter.Specifically, smog dynamic characteristic is specifically
Refer to cigarette when waving, profile variations are random, and being presented as has the area change in cigarette district domain little, and perimeter change is obvious
Region retains the detection for meeting the region progress next stage of cigarette dynamic characteristic according to this characteristic.
Establish smog dark model, based on fixed background to the mistake for comparing video image and carrying out image pre-segmentation
Cheng Zhong:
Judge whether video image includes region that dark channel value meets smog dark model；
Meet the region of smog dark model if there is dark channel value, then judges that dark channel value meets smog dark
Whether the region of model meets smog dynamic characteristic.
Ineligible region is finally rejected, retains qualified region to realize image pre-segmentation.
Further, in one embodiment, smog dark model and/or smog dynamic characteristic are by smog sample
What the training of image obtained.Specifically, in one embodiment, method further include:
Smog sample image is obtained, smog dark model and/or smog dynamic are obtained according to the training of smog sample image
Characteristic.
Further, in one embodiment, the smog picture searched out by collection in worksite and on the net pre-processes, to smog
Contours extract is carried out, binaryzation contour images is generated, generates smoke sample image after mark.
Further, in one embodiment, during extracting flare textural characteristics from doubtful flare image, due to
After shaping, its position and area change is little for cigarette, before the entropy acquisition based on this by calculating doubtful smog image before and after frames
The aggregation characteristic and profile gray level co-occurrence matrixes of image grayscale distribution obtain two pictures for being separated by certain distance in image space afterwards
Existing Gray Correlation between element is waved the description factor of feature using them as cigarette.The knot that final basis is calculated
Fruit judges whether the region has the possibility of cigarette.
Further, in one embodiment, it before extracting flare textural characteristics in doubtful flare image, needs to pass through
The smog overall size being partitioned into is normalized in dynamic characteristic and dark characteristic.
Further, based on method of the invention, the invention also provides a kind of fire early-warning systems.As shown in figure 3, In
In one embodiment, system includes camera 310 and warning module 320 (software processing platform), fire alarm software load (fortune
Row) on warning module 320 (software processing platform).
Specifically, in one embodiment, software processing platform is embedded-type ARM processing platform, and the processor performance is more
Stablize, is suitable for vehicle environment.
Specifically, in one embodiment, camera is 360 degree of IP Cameras, full-shape in compartment can be acquired in real time
The video pictures of degree, and send processor in time and carry out image characteristic analysis, thus to the timely early warning of unusual condition in compartment.
Specifically, in one embodiment, fire alarm software frame is realized based on QT, it is real that vision algorithm is based on openCV
It is existing；So that exploitation is more efficient, convenient for debugging.
In one embodiment, fire alarm software operates on I.MX6 embedded development platform, by utilizing motor train compartment
It looks around IP Camera for interior 360 degree and reads data, motor train compartment real-time status is monitored, software mainly uses computer to regard
To normal scene and there is the field of fiery point/have obvious smoke by the method for machine learning by extracting characteristics of image in feel method
Scape is classified.To realize timely fire alarm.
Specifically, in one embodiment, carrying out scene first by the live video that 360 degree of IP Cameras obtain and sentencing
It is disconnected, then start to carry out flare detection and Smoke Detection if it is RGB image, if being found to have flare or cigarette after software detection
The characteristic information of mist makes early warning in time.
The software frame of fire alarm software realizes that bottom vision algorithm is realized based on openCV, the stream of software based on QT
Journey includes video acquisition, video decoding, pretreatment, image characteristics extraction, feature detection, the links such as the result is shown.
IP Camera acquires video image, then carries out video decoding to the video image of acquisition.Fire alarm software
The video carried out first after video image decoding is obtained by web camera, analysis and video data including video format
It obtains, includes stretching, denoising and the extraction of area-of-interest followed by pretreatment；Then the flame and cigarette of image are carried out
Mist feature extraction finally carries out feature detection and early warning.Since image detection is divided into the detection of flare and smog, due to flare
Feature is more significant, and algorithm preferentially detects flare, without the detection of smog when there are flare, otherwise continues whether there is
The judgement of smog.Final result shows positioning and early warning response including suspicious object.
Further, in one embodiment, software processing platform is X 86 processor processing platform.
Further, in one embodiment, camera is multiple cameras for being installed on different location to obtain in compartment
Panorama.
While it is disclosed that embodiment content as above but described only to facilitate understanding the present invention and adopting
Embodiment is not intended to limit the invention.Method of the present invention can also have other various embodiments.Without departing substantially from
In the case where essence of the present invention, those skilled in the art make various corresponding changes or change in accordance with the present invention
Shape, but these corresponding changes or deformation all should belong to scope of protection of the claims of the invention.
Claims (10)
1. a kind of fire alarm method for compartment, which is characterized in that the described method includes:
Acquire compartment inner video image；
Image pre-segmentation is carried out to the video image, obtains doubtful flare image；
Confidence declaration is carried out to the doubtful flare image based on flare anticipation Gauss model；
When the doubtful flare image passes through the confidence declaration, it is special that flare texture is extracted from the doubtful flare image
Sign is extracted；
Judge that current scene with the presence or absence of flare, issues early warning when there are flare according to the flare textural characteristics.
2. the method according to claim 1, wherein passing through background subtraction based on the comparison with fixed background
Carry out image pre-segmentation.
3. obtaining and doubting the method according to claim 1, wherein carrying out image pre-segmentation to the video image
Like flare image, comprising:
Judge whether the video image includes region that color and/or luminance information meet flare color and/or luminance information.
4. obtaining and doubting the method according to claim 1, wherein carrying out image pre-segmentation to the video image
Like flare image, in which:
When there are the doubtful flare image, the fixed background is not updated
When the doubtful flare image is not present, fixed background described in real-time update.
5. the method according to claim 1, wherein the method also includes:
Flare sample image is obtained, the flare anticipation model is obtained according to flare sample image training and/or determines institute
Confidence threshold value is stated, the flare sample image includes positive sample and/or negative sample, in which:
Acquisition flare image simultaneously marks the region for picking out flare in every frame image, generates the positive sample after size normalization；
And/or
Model is prejudged using flare described in flare imaging experiments, obtains the image for being not recognized as flare, mark picks out every frame
The region of flare in image generates the negative sample after size normalization.
6. the method according to claim 1, wherein carrying out flare textural characteristics for the doubtful flare image
It extracts, in which:
Utilize the textural characteristics of LBP operator extraction image.
7. method described according to claim 1~any one of 6, which is characterized in that the method also includes:
Image pre-segmentation is carried out to the video image, obtains doubtful smog image；
Smoke characteristics are extracted from the doubtful smog image；
Judge that current scene with the presence or absence of smog, issues early warning when there are smog according to the smoke characteristics.
8. obtaining and doubting the method according to the description of claim 7 is characterized in that carrying out image pre-segmentation to the video image
Like smog image, comprising:
Judge whether the video image includes region that dark channel value meets smog dark model；
Judge that dark channel value meets the region of smog dark model and whether meets smog dynamic characteristic.
9. according to the method described in claim 8, it is characterized in that, the method also includes:
Smog sample image is obtained, the smog dark model and/or described is obtained according to smog sample image training
Smog dynamic characteristic.
10. the method according to the description of claim 7 is characterized in that extract smoke characteristics from the doubtful smog image,
In, calculate the entropy and profile gray level co-occurrence matrixes of the doubtful smog image before and after frames.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810560772.9A CN110533872A (en) | 2018-05-25 | 2018-05-25 | A kind of fire alarm method for compartment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810560772.9A CN110533872A (en) | 2018-05-25 | 2018-05-25 | A kind of fire alarm method for compartment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110533872A true CN110533872A (en) | 2019-12-03 |
Family
ID=68657363
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810560772.9A Pending CN110533872A (en) | 2018-05-25 | 2018-05-25 | A kind of fire alarm method for compartment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110533872A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115294718A (en) * | 2022-08-09 | 2022-11-04 | 九江职业技术学院 | Fire early warning system based on multisource data fusion |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN202838583U (en) * | 2012-10-17 | 2013-03-27 | 无锡嘉华智慧交通科技有限公司 | Track traffic vehicle-mounted fire disaster image alarming device |
| CN105046868A (en) * | 2015-06-16 | 2015-11-11 | 苏州华启智能科技股份有限公司 | Fire early warning method based on infrared thermal imager in narrow environment |
| CN105225235A (en) * | 2015-09-18 | 2016-01-06 | 北京航空航天大学 | A kind of video flame detecting method based on multispectral characteristic |
| CN206639341U (en) * | 2017-02-24 | 2017-11-14 | 郭敏强 | Fire alarm installation in novel and multifunctional subway carriage |
| CN107944359A (en) * | 2017-11-14 | 2018-04-20 | 中电数通科技有限公司 | Flame detecting method based on video |
-
2018
- 2018-05-25 CN CN201810560772.9A patent/CN110533872A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN202838583U (en) * | 2012-10-17 | 2013-03-27 | 无锡嘉华智慧交通科技有限公司 | Track traffic vehicle-mounted fire disaster image alarming device |
| CN105046868A (en) * | 2015-06-16 | 2015-11-11 | 苏州华启智能科技股份有限公司 | Fire early warning method based on infrared thermal imager in narrow environment |
| CN105225235A (en) * | 2015-09-18 | 2016-01-06 | 北京航空航天大学 | A kind of video flame detecting method based on multispectral characteristic |
| CN206639341U (en) * | 2017-02-24 | 2017-11-14 | 郭敏强 | Fire alarm installation in novel and multifunctional subway carriage |
| CN107944359A (en) * | 2017-11-14 | 2018-04-20 | 中电数通科技有限公司 | Flame detecting method based on video |
Non-Patent Citations (5)
| Title |
|---|
| XIAO-HAN CHEN: "Fire Alarm Using Multi-rules Detection and Texture Features Classification in Video Surveillance", 《2014 7TH INTERNATIONAL CONFERENCE ON INTELLIGENT COMPUTATION TECHNOLOGY AND AUTOMATION》 * |
| 姜东海: ""基于图像处理的火灾烟雾识别技术研究"", 《中国优秀硕士学位论文全文数据库（电子期刊） 信息科技辑》 * |
| 文泽波: "基于视频图像特征提取的烟雾检测算法研究", 《中国优秀硕士学位论文全文数据库（电子期刊） 信息科技辑》 * |
| 杨亚洁 等: ""基于多特征匹配的视频图像火灾火焰检测方法研究"", 《电子设计工程》 * |
| 邓兴 等: ""基于灰度共生矩阵的烟雾图像纹理分析"", 《六盘水师范学院学报》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115294718A (en) * | 2022-08-09 | 2022-11-04 | 九江职业技术学院 | Fire early warning system based on multisource data fusion |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Çetin et al. | Video fire detection–review | |
| CN106373320B (en) | Method for recognizing fire disaster based on flame color dispersion and sequential frame image similarity | |
| CN110516609A (en) | A kind of fire video detection and method for early warning based on image multiple features fusion | |
| KR101822924B1 (en) | Image based system, method, and program for detecting fire | |
| CN106897720A (en) | A kind of firework detecting method and device based on video analysis | |
| CN109637068A (en) | Intelligent pyrotechnics identifying system | |
| CN112069975A (en) | Comprehensive flame detection method based on ultraviolet, infrared and vision | |
| CN107301405A (en) | Method for traffic sign detection under natural scene | |
| CN103839085B (en) | A kind of detection method of compartment exception crowd density | |
| KR100918436B1 (en) | Fire detection system and method basedon visual data | |
| CN108038867A (en) | Fire defector and localization method based on multiple features fusion and stereoscopic vision | |
| CN107437318B (en) | Visible light intelligent recognition algorithm | |
| Khalil et al. | Fire detection using multi color space and background modeling | |
| WO2013101460A2 (en) | Clustering-based object classification | |
| CN109409186A (en) | Driver assistance system and method for object detection and notice | |
| CN103903020B (en) | A kind of fire image recognition methods and device based on CodeBook | |
| CN112183472A (en) | Method for detecting whether test field personnel wear work clothes or not based on improved RetinaNet | |
| CN107688793A (en) | A kind of outside transformer substation fire automatic monitoring method for early warning | |
| CN109961042A (en) | Smoke detection method combining deep convolutional neural network and visual change diagram | |
| Dung et al. | Algorithm for fire detection using a camera surveillance system | |
| CN112287823A (en) | Facial mask identification method based on video monitoring | |
| CN109034038B (en) | Fire identification device based on multi-feature fusion | |
| CN111553214B (en) | Method and system for detecting smoking behavior of driver | |
| CN114202646A (en) | Infrared image smoking detection method and system based on deep learning | |
| CN109274945B (en) | Method and system for self-adaptively performing true color restoration on image |
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: 20191203 |
|
| RJ01 | Rejection of invention patent application after publication |