CN110812753B - Artificial intelligent fire extinguishing method with open fire point identification function and fire extinguisher equipment - Google Patents
Artificial intelligent fire extinguishing method with open fire point identification function and fire extinguisher equipment Download PDFInfo
- Publication number
- CN110812753B CN110812753B CN201910901117.XA CN201910901117A CN110812753B CN 110812753 B CN110812753 B CN 110812753B CN 201910901117 A CN201910901117 A CN 201910901117A CN 110812753 B CN110812753 B CN 110812753B
- Authority
- CN
- China
- Prior art keywords
- open fire
- fire
- fire point
- candidate
- point
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- A—HUMAN NECESSITIES
- A62—LIFE-SAVING; FIRE-FIGHTING
- A62C—FIRE-FIGHTING
- A62C31/00—Delivery of fire-extinguishing material
- A62C31/02—Nozzles specially adapted for fire-extinguishing
-
- A—HUMAN NECESSITIES
- A62—LIFE-SAVING; FIRE-FIGHTING
- A62C—FIRE-FIGHTING
- A62C37/00—Control of fire-fighting equipment
- A62C37/04—Control of fire-fighting equipment with electrically-controlled release
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/20—Scenes; Scene-specific elements in augmented reality scenes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Public Health (AREA)
- Business, Economics & Management (AREA)
- Emergency Management (AREA)
- Fire-Detection Mechanisms (AREA)
- Fire Alarms (AREA)
Abstract
The invention has proposed a artificial intelligence with open fire point recognition function puts out a fire method and fire extinguisher apparatus, said method comprises S1, shoot the consecutive image frame; s2, converting each image frame into an HSI component map; s3, setting distinguishing threshold values of saturation S and brightness I respectively, comparing the distinguishing threshold values with corresponding components of each image frame, and analyzing whether a candidate open fire point area exists or not; s4, when a candidate open fire point region exists, calculating to obtain a characteristic value array sequence containing a series of candidate open fire point region characteristic values; s5, judging the ignition condition of the candidate open fire point area and quantitatively generating an open fire ignition condition grade; and S6, when the actual open fire point exists in the candidate open fire point area, opening a safety mechanism of the fire extinguisher, and spraying the open fire point in a proper amount, wherein the spraying angle is consistent with the visual field direction of the camera. The invention is beneficial to avoiding the occurrence of casualties and fire extinguishing substance waste caused by error spraying, makes up the defects of the existing fire extinguisher and embodies the superiority of artificial intelligence in the aspect of fire fighting.
Description
Technical Field
The invention relates to the technical field of intelligent fire fighting, in particular to an artificial intelligent fire extinguishing method with an open fire point identification function and fire extinguisher equipment.
Background
The fire extinguisher is one of common fire-proof facilities, is convenient for people to take and use when a fire disaster happens, and is generally stored in public places, private cars or houses; fire extinguishers containing different chemical substances are different in properties, for example, a dry powder fire extinguisher contains 50% of ammonium dihydrogen phosphate and 25% of ammonium sulfate, and is very corrosive, and in addition, sprayed dust can be rapidly settled, a large amount of particles are inhaled into the lung, so that people can be suffocated, and therefore, when the fire extinguisher is used, the fire extinguisher is forbidden to be sprayed to a human body, and casualties are avoided.
However, in case of serious fire, the user may mistakenly spray the fire to the human body without adjusting the spraying direction due to scaring or spray the fire to the burn personnel because the composition and property of the fire extinguisher are unknown, which may cause wound infection and secondary damage to the burn personnel.
Therefore, under the large environment of artificial intelligence, how to design an artificial intelligence fire extinguishing apparatus capable of automatically identifying the open fire point, the situation that the fire extinguisher is mistakenly sprayed due to the error of a user is prevented, and the technical personnel in the field need to solve the problem urgently.
Disclosure of Invention
In view of the above, the invention provides an artificial intelligence fire extinguishing method and fire extinguisher equipment with an open fire point identification function, wherein the common fire extinguisher equipment is combined with a camera, a digital processing chip and a control interface, and the visual field direction shot by the camera is consistent with the spraying direction of the fire extinguisher; the digital processing chip carries out artificial intelligence analysis to the image of camera shooting, judges whether have the naked light point, if exist the naked light point then open fire extinguisher's safety mechanism through the control interface, allows the fire extinguisher to spray, avoids using the fire extinguisher in violation of rules and regulations, causes the emergence of mistake spraying condition.
In order to achieve the purpose, the invention adopts the following technical scheme:
an artificial intelligence fire extinguishing method with an open fire point identification function comprises the following steps:
s1, shooting continuous image frames;
s2, converting each image frame into an HSI component diagram from an RGB model;
s3, setting distinguishing threshold values of saturation S and brightness I respectively, comparing the saturation S and the brightness I in the component image of each image frame with the corresponding distinguishing threshold values, and analyzing whether a candidate open fire point area exists or not;
s4, respectively acquiring a characteristic value array of the candidate open fire point region of each image frame when the candidate open fire point region exists, and acquiring a characteristic value array sequence containing a series of candidate open fire point region characteristic values;
s5, judging the ignition condition of the candidate open fire point area according to the characteristic value array sequence and quantitatively generating an open fire ignition condition grade;
and S6, when the candidate open fire point area has a real open fire point, opening a safety mechanism of the fire extinguisher, and carrying out proper injection on the open fire point area according to the level of the open fire ignition condition, wherein the injection position is consistent with the shooting angle of the image frame.
Preferably, the HSI component map in S2 is a component map composed of components of hue H, saturation S, and brightness I, and since the HSI model has basic attributes that completely reflect colors perceived by humans and corresponds to the results of colors perceived by humans one to one, the HSI model is applied to a visual recognition system, so as to facilitate artificial intelligence perception of colors of images and process and analyze the images.
Preferably, in S3, when both the saturation S and the brightness I of each image frame meet the threshold limit, the image frame region is defined as a candidate ignition point region, and when the saturation S and the brightness I of each image frame do not meet the threshold limit, the candidate ignition point region does not exist; specifically, the saturation S and the brightness I of the flame are obviously different from the saturation S and the brightness I of an unfired area, the red saturation of a flame image pixel is reduced along with the increase of the brightness, and the brightness of the flame image pixel is gradually reduced from the center to the edge of the flame, so that the threshold values of the saturation S and the brightness I are set, the anti-interference performance is strong in the open flame identification process, and whether the image area has an open flame point or not can be accurately judged.
Preferably, in S4, the area a of the candidate open fire region and the coordinates C (x, y) of the area center of gravity of the candidate open fire region in each image frame are calculated, the acquisition time T of each image frame is obtained, and the area a, the coordinates C (x, y) of the center point and the acquisition time T are used as the multivariate array of the characteristic values of the candidate open fire region: and (A, C (x, y), T), combining the multiple arrays of the candidate fire point region characteristic values of each image frame to form a characteristic value array sequence:
<A1，C1(x，y)，T1>，<A2，C2(x，y)，T2>，......，<An，Cn(x，y)，Tn>；
wherein n is more than or equal to 24;
the camera can shoot more than 24 frames of images within one second, so n is more than or equal to 24, a characteristic value array sequence is set, parameters of candidate open fire point areas of a series of collected picture frames are conveniently integrated, and the subsequent analysis of the fire conditions of the candidate open fire point areas is provided.
Preferably, in S5, the feature value array sequence is substituted into a trained BP neural network, and the BP neural network outputs fire levels of the candidate open fire regions, where the fire levels are divided into six levels of 0 to 5; wherein, the 0 grade represents the non-fire state, and the 1-5 grades are arranged according to the severity degree of the fire condition in an increasing way.
A BP neural network is a learning process of an error back propagation algorithm and consists of two processes of information forward propagation and error back propagation, firstly, a characteristic value array sequence is input through an input layer, input sample information is transmitted to each neuron of an intermediate layer, the intermediate layer is an internal information processing layer and takes charge of information transformation, so that the intermediate layer processes and analyzes the input characteristic value array sequence information into fire levels, the fire levels are transmitted to each neuron of an output layer through a last hidden layer of the intermediate layer, the fire levels are output by the output layer and are compared with the fire levels in the sample, if the comparison result is not consistent, the BP neural network enters a back propagation stage of errors, weights of all layers are corrected through the output layer in a mode of error gradient reduction, the processes are repeated until the errors of the output result are reduced to an acceptable degree, and a trained BP neural network is obtained, the BP neural network has the capability of identifying similar flame images, so that the BP neural network is applied to identification of open fire grades, specifically, a characteristic value array sequence of actually shot images is input, the characteristic value array sequence is processed and analyzed by the BP neural network, a numerical value 0-5 representing the fire grade is output, when the output is 0, the fire does not fire, and when the output is 1-5, the fire severity increases progressively according to the increasing of the numerical value, so that the flow of the sprayed fire extinguishing substance pair is controlled, the aim of extinguishing fire is fulfilled, resources are saved, and waste is avoided.
Preferably, the spraying position is consistent with the shooting angle of the image frame, and the fire extinguisher body can be accurately started or prohibited from spraying according to the fire condition after the work is finished.
An artificial intelligence fire extinguisher device with open fire point recognition function, comprising: the fire extinguisher comprises a fire extinguisher body, a camera, a digital processing chip and a control interface; wherein the content of the first and second substances,
the digital processing chip includes: the device comprises a component diagram conversion module, a flame identification module, a characteristic value array sequence acquisition module and a fire condition grade acquisition module;
the camera is used for shooting continuous image frames;
the component map conversion module is used for converting each image frame into an HSI component map from an RGB model;
the flame identification module is used for respectively setting a distinguishing threshold of saturation S and brightness I, comparing the saturation S and the brightness I in each image frame component image with the corresponding distinguishing threshold, and analyzing whether a candidate open fire area exists or not;
when the candidate open fire point areas exist, the characteristic value array sequence obtaining module is used for respectively obtaining the characteristic value arrays of the candidate open fire point areas of each image frame and obtaining a characteristic value array sequence containing a series of candidate open fire point area characteristic values;
the fire condition grade acquisition module is used for judging the fire conditions of the candidate open fire point areas according to the characteristic value array sequence and quantitatively generating open fire ignition condition grades;
when the candidate open fire point area has a real open fire point, the control interface opens a safety mechanism of the fire extinguisher body, and proper amount of injection is carried out on the open fire point area according to the level of the open fire ignition condition, and the injection position is consistent with the shooting angle of the image frame.
Preferably, the HSI component map in the component map conversion module is a component map composed of components of hue H, saturation S, and brightness I, and conforms to the way in which the human visual system perceives colors.
Preferably, in the flame identification module, when the saturation S and the brightness I of each image frame both meet the threshold limit, the image frame region is defined as a candidate ignition point region, and when the saturation S and the brightness I of each image frame do not meet the threshold limit, the candidate ignition point region does not exist.
Preferably, the eigenvalue array sequence obtaining module calculates an area a of the candidate open fire region of each image frame and a region gravity center point coordinate C (x, y) of the candidate open fire region, respectively, obtains an acquisition time T of each image frame, and takes the area a, the center point coordinate C (x, y), and the acquisition time T as a multivariate array of the eigenvalues of the candidate open fire region: and (A, C (x, y), T), combining the multiple arrays of the candidate fire point region characteristic values of each image frame to form a characteristic value array sequence:
<A1，C1(x，y)，T1>，<A2，C2(x，y)，T2>……<An，Cn(x，y)，Tn>；
wherein n is more than or equal to 24.
Preferably, the fire level acquisition module brings the characteristic value array sequence into a trained BP neural network, the BP neural network outputs fire levels of the candidate open fire point areas, and the fire levels are divided into 0-5 levels; wherein, the 0 grade represents the non-fire state, and the 1-5 grades are arranged according to the severity degree of the fire condition in an increasing way.
Preferably, the spraying direction of the fire extinguisher body is consistent with the visual field direction of the camera, so that whether the control interface is started or not can be accurately determined according to the fire condition by the fire extinguisher body after the work is finished.
The invention has the following beneficial effects:
based on the technical scheme, the invention provides the artificial intelligent fire extinguishing method with the open fire point identification function based on the prior art, and the artificial intelligent fire extinguisher equipment with the open fire point identification function is designed according to the method.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the provided drawings without creative efforts.
FIG. 1 is a flow chart of a method of the present invention;
fig. 2 is a block diagram of the apparatus of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
As shown in fig. 1, the present invention proposes the following method:
an artificial intelligence fire extinguishing method with an open fire point identification function comprises the following steps:
s1, shooting continuous image frames;
s2, converting each image frame into an HSI component diagram from an RGB model;
specifically, the HSI component diagram is a component diagram formed by components of hue H, saturation S and brightness I, and accords with the mode of human visual system color perception, so that the HSI model is applied to a visual recognition system, the image color is perceived conveniently and intelligently, and the efficiency of processing and analyzing the image is improved.
S3, setting distinguishing threshold values of saturation S and brightness I respectively, comparing the saturation S and the brightness I in each image frame component image with the corresponding distinguishing threshold values, and analyzing whether a candidate open fire point area exists or not;
specifically, when the saturation S and the brightness I of each image frame both meet the threshold limit, the image frame region is defined as a candidate open fire region, and when the saturation S and the brightness I of each image frame do not all meet the threshold limit, the candidate open fire region does not exist; the saturation S and the brightness I of the flame are obviously different from the saturation S and the brightness I of an ignition area, the red saturation of a flame image pixel is reduced along with the increase of the brightness, and the brightness of the flame image pixel is gradually reduced from the center to the edge of the flame, so that the threshold values of the saturation S and the brightness I are set, and whether the image area has a candidate open fire point or not is judged according to whether the threshold values accord with the limit or not.
S4, respectively acquiring a characteristic value array of the candidate open fire point region of each image frame when the candidate open fire point region exists, and acquiring a characteristic value array sequence containing a series of candidate open fire point region characteristic values;
specifically, the area a of the candidate open fire point region of each image frame and the region center of gravity point coordinates C (x, y) of the candidate open fire point region are respectively calculated, the acquisition time T of each image frame is obtained, and the area a, the center point coordinates C (x, y) and the acquisition time T are used as the multi-element array of the characteristic values of the candidate open fire point region: < A, C (x, y), T >.
Further, for two adjacent frames of images, candidate open fire point areas are respectively extracted as GiAnd Gi+1Then, subtraction processing is performed on the two frame images (0 is set for pixel values smaller than 0), and an image GD (x, y) is generated:
GD(x，y)＝Gi-Gi+1
and (3) applying a Wiener filter to carry out smooth denoising treatment on the GD (x, y) to remove isolated point noise. Image G is then determined using the following methodi+1Open fire area of (1):
1) the image GD (x, y) is set as a start detection target pixel.
2) The new set of pixels is taken as the starting point for detection.
3) A clustering condition is determined as a pixel growth criterion.
(1) Adjacent to the seed point.
(2) The brightness value of the pixel point is larger than T. The value of T is about 0.9, which is suitable.
(3) Definition Gi+1The maximum value R (x, y) of the gradient in each direction in the image is generally small in value, so that the open fire region can be effectively detected basically by setting R (x, y) to be less than 0.1. Wherein R (x, y) is defined as
R(x，y)＝max(|Gi+1(x，y+1)-Gi+1(x，y)|，|Gi+1(x，y-1)-Gi+1(x，y)|，|Gi+1(x+1，y)-Gi+1(x，y)|，|Gi+1(x-1，y)-Gi+1(x，y)|，|Gi+1(x+1，y+1)-Gi+1(x，y)|，|Gi+1(x+1，y-1)-Gi+1(x，y)|，|Gi+1(x-1，y+1)-Gi+1(x，y)|，|Gi+1(x-1，y-1)-Gi+1(x，y)|)
4) Searching the whole image, finding out the pixels which accord with the step 3), and adding the pixels into the pixel set. Generating a new pixel set, judging whether the pixel set is increased, if so, returning to the step 2), and if not, executing the step 5).
5) Generation of image G 'from pixels in a set of pixels'2。
Calculating an area value A of the open fire point region: the area value a is defined as the sum of the points in the target extracted binary image with pixel 1:
defining the coordinates of the gravity center point of the target image by considering the change of the position of the gravity center point, and then binarizing the image G'2Center of gravity point coordinate C (x)c，yc) Comprises the following steps:
further, a multivariate array of the eigenvalues of the fire zone is generated<A，C(xc，yc)，T>。
Combining the multi-element arrays of the candidate fire point region characteristic values of each image frame to form a characteristic value array sequence:
<A1，C1(x，y)，T1>，<A2，C2(x，y)，T2>，......，<An，Cn(x，y)，Tn>；
wherein n is more than or equal to 24;
the camera can shoot images more than 24 frames in one second, so n is more than or equal to 24, the image frames are obtained by the camera for identification, the image frame obtaining speed is high, the open fire judgment time is shortened, the fire extinguishment is not delayed, the fire aggravation is caused, a large number of parameters can be collected, the follow-up accurate analysis on the fire in the candidate open fire area is provided, and the accurate fire grade is obtained.
S5, judging the ignition condition of the candidate open fire point area according to the characteristic value array sequence and quantitatively generating an open fire ignition condition grade;
specifically, the characteristic value array sequence is brought into a trained BP neural network, the BP neural network outputs fire levels of the candidate open fire point areas, and the fire levels are divided into six levels of 0-5; wherein, the 0 grade represents the non-fire state, and the 1-5 grades are arranged according to the severity degree of the fire condition in an increasing way.
Because the BP neural network has self-learning capability, the similar real characteristic value array sequence can be quickly and accurately analyzed and the accurate fire level can be output by adopting the BP neural network trained by a large number of characteristic value array sequence samples and fire level samples, so that the fire extinguishing operation can be carried out according to the fire level.
And S6, when the candidate open fire point area has a real open fire point, opening a safety mechanism of the fire extinguisher, and carrying out proper injection on the open fire point area according to the level of the open fire ignition condition, wherein the injection position is consistent with the shooting angle of the image frame.
Specifically, the spraying position is consistent with the shooting angle of the image frame, and the fire extinguisher body can be accurately started or prohibited from spraying according to the fire condition after the work is finished.
As shown in fig. 2, according to the above method, the following apparatus is designed:
an artificial intelligence fire extinguisher device with open fire point recognition function, comprising: the fire extinguisher comprises a fire extinguisher body 4, a camera 1, a digital processing chip 2 and a control interface 3; wherein the content of the first and second substances,
the digital processing chip 2 includes: the device comprises a component diagram conversion module 21, a flame identification module 22, a characteristic value array sequence acquisition module 23 and a fire situation grade acquisition module 24;
the camera 1 is used for shooting continuous image frames;
the component map conversion module 21 is configured to convert each image frame from an RGB model to an HSI component map;
the flame identification module 22 is configured to set a distinguishing threshold of saturation S and brightness I, compare the saturation S and the brightness I in each image frame component map with the corresponding distinguishing threshold, and analyze whether a candidate open fire region exists;
when the candidate open fire point region exists, the characteristic value array sequence obtaining module 23 is configured to obtain a characteristic value array of the candidate open fire point region of each image frame, and obtain a characteristic value array sequence including a series of candidate open fire point region characteristic values;
the fire situation grade acquisition module 24 is used for judging the fire situation of the candidate open fire point region according to the characteristic value array sequence and quantitatively generating an open fire ignition situation grade;
when the candidate open fire point area has a real open fire point, the control interface 3 opens a safety mechanism of the fire extinguisher body 4, and proper amount of injection is carried out on the open fire point area according to the level of the open fire ignition condition, and the injection position is consistent with the shooting angle of the image frame.
In order to further optimize the technical characteristics, the fire extinguisher body 4 comprises a fire extinguishing nozzle and a safety device, the opening of the safety device is determined by the control interface according to the fire condition grade, and when the fire extinguisher body 4 is in a state of spraying fire extinguishing substances, the spraying direction of the fire extinguishing nozzle needs to be ensured to be consistent with the visual field direction of the camera 1 so as to avoid that the judgment of the fire condition is not consistent with the actual condition.
In order to further optimize the above technical features, the HSI component map in the component map conversion module 21 is a component map composed of components of hue H, saturation S, and brightness I, and conforms to the way in which the human visual system perceives colors.
In order to further optimize the above technical features, in the flame identification module 22, when the saturation S and the brightness I of each image frame both meet the threshold limit, the image frame region is defined as a candidate open fire point region, and when the saturation S and the brightness I of each image frame do not meet the threshold limit, the candidate open fire point region does not exist.
In order to further optimize the above technical features, the feature value array sequence obtaining module 23 respectively calculates an area a of the candidate open fire region and a region gravity center point coordinate C (x, y) of the candidate open fire region of each image frame, obtains an acquisition time T of each image frame, and takes the area a, the center point coordinate C (x, y), and the acquisition time T as a multivariate array of feature values of the candidate open fire region: < A, C (x, y), T >.
Further, for two adjacent frames of images, candidate open fire point areas are respectively extracted as GiAnd Gi+1Then, subtraction processing is performed on the two frame images (0 is set for pixel values smaller than 0), and an image GD (x, y) is generated:
GD(x，y)＝Gi-Gi+1
and (3) applying a Wiener filter to carry out smooth denoising treatment on the GD (x, y) to remove isolated point noise. Image G is then determined using the following methodi+1Open fire area of (1):
1) the image GD (x, y) is set as a start detection target pixel.
2) The new set of pixels is taken as the starting point for detection.
3) A clustering condition is determined as a pixel growth criterion.
(1) Adjacent to the seed point.
(2) The brightness value of the pixel point is larger than T. The value of T is about 0.9, which is suitable.
(3) Definition Gi+1Maximum value of gradient R (x, y) in each direction in the image because of brightnessR (x, y) in the fire point area is generally small in value, so that the open fire point area can be effectively detected basically by setting R (x, y) to be less than 0.1. Wherein R (x, y) is defined as
R(x，y)＝max(|Gi+1(x，y+1)-Gi+1(x，y)|，|Gi+1(x，y-1)-Gi+1(x，y)|，|Gi+1(x+1，y)-Gi+1(x，y)|，|Gi+1(x-1，y)-Gi+1(x，y)|，|Gi+1(x+1，y+1)-Gi+1(x，y)|，|Gi+1(x+1，y-1)-Gi+1(x，y)|，|Gi+1(x-1，y+1)-Gi+1(x，y)|，|Gi+1(x-1，y-1)-Gi+1(x，y)|)
4) Searching the whole image, finding out the pixels which accord with the step 3), and adding the pixels into the pixel set. Generating a new pixel set, judging whether the pixel set is increased, if so, returning to the step 2), and if not, executing the step 5).
5) Generation of image G 'from pixels in a set of pixels'2。
Calculating an area value A of the open fire point region: the area value a is defined as the sum of the points in the target extracted binary image with pixel 1:
defining the coordinates of the gravity center point of the target image by considering the change of the position of the gravity center point, and then binarizing the image G'2Center of gravity point coordinate C (x)c，yc) Comprises the following steps:
further, a multivariate array of the eigenvalues of the fire zone is generated<A，C(xc，yc)，T>. Combining the multi-element arrays of the candidate fire point region characteristic values of each image frame to form a characteristic value array sequence:
<A1，C1(x，y)，T1>，<A2，C2(x，y)，T2>，......，<An，Cn(x，y)，Tn>；
wherein n is more than or equal to 24.
In order to further optimize the technical characteristics, the fire level acquisition module 24 brings the characteristic value array sequence into a trained BP neural network, and the BP neural network outputs fire levels of the candidate open fire point areas, wherein the fire levels are divided into 0-5 levels; wherein, the 0 grade represents the non-fire state, and the 1-5 grades are arranged according to the severity degree of the fire condition in an increasing way.
In order to further optimize the technical characteristics, the fire extinguishing substance contained in the fire extinguisher body 4 is dry powder, and the fire extinguisher body is a conventional dry powder fire extinguisher, so that when the artificial intelligent fire extinguishing method and the fire extinguisher equipment provided by the invention are used, the dry powder fire extinguisher can be prevented from being sprayed to people to cause skin corrosion or suffocation, the flow can be controlled according to the fire level, and the dry powder waste is avoided.
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. The device disclosed by the embodiment corresponds to the method disclosed by the embodiment, so that the description is simple, and the relevant points can be referred to the method part for description.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (9)
1. An artificial intelligence fire extinguishing method with an open fire point identification function is characterized by comprising the following steps:
s1, shooting continuous image frames;
s2, converting each image frame into an HSI component diagram from an RGB model;
s3, setting distinguishing threshold values of saturation S and brightness I respectively, comparing the saturation S and the brightness I in each image frame component image with the corresponding distinguishing threshold values, and analyzing whether a candidate open fire point area exists or not;
s4, respectively acquiring a characteristic value array of the candidate open fire point region of each image frame when the candidate open fire point region exists, and acquiring a characteristic value array sequence containing a series of candidate open fire point region characteristic values;
s5, judging the ignition condition of the candidate open fire point area according to the characteristic value array sequence and quantitatively generating an open fire ignition condition grade;
s6, when the candidate open fire point area has a real open fire point, opening a safety mechanism of the fire extinguisher, and carrying out proper injection on the open fire point area according to the level of the open fire ignition condition, wherein the injection position is consistent with the shooting angle of the image frame; the S4 includes the steps of,
respectively calculating the area A of a candidate open fire point region of each image frame and the region gravity center point coordinates C (x, y) of the candidate open fire point region, acquiring the acquisition time T of each image frame, and taking the area A, the center point coordinates C (x, y) and the acquisition time T as a multi-element array of the characteristic values of the candidate open fire point region: < A, C (x, y), T >,
for two adjacent frames of images, respectively extracting candidate open fire point areas as GiAnd Gi+1Then, subtraction processing is performed on the two frame images, and an image GD (x, y) is generated with a pixel value smaller than 0 set to 0:
GD(x,y)= Gi- Gi+1
carrying out smooth denoising treatment on GD (x, y) by using a Wiener filter to remove isolated point noise; image G is then determined using the following methodi+1Open fire area of (1):
1) with the image GD (x, y) as a start detection target pixel set;
2) taking the new pixel set as a detection starting point;
3) determining a clustering condition as a pixel growth criterion;
(1) adjacent to the seed point;
(2) the brightness value of the pixel point is larger than T, and the value of T is 0.9;
(3) definition Gi+1The maximum value R (x, y) of the gradient in each direction in the image is set to be less than 0.1, wherein R (x, y) is defined as
4) Searching the whole image, finding out the pixels which accord with the step 3), and adding the pixels into the pixel set; generating a new pixel set, judging whether the pixel set is increased, if so, returning to the step 2), and if not, executing the step 5);
5) generation of image G from pixels in a set of pixels2’；
Calculating an area value A of the open fire point region: the area value a is defined as the sum of the points in the target extracted binary image with pixel 1:
considering the change of the position of the gravity center point, defining the gravity center point coordinate of the target image, and then binarizing the image G2' center of gravity point coordinate C (x)c,yc) Comprises the following steps:
further, a multivariate array of the eigenvalues of the fire zone is generated
(ii) a For each image frameThe multivariate arrays of the characteristic values of the candidate open fire point areas are combined to form a characteristic value array sequence:
wherein n is more than or equal to 24.
2. The artificial intelligence fire extinguishing method with the function of open fire point identification according to claim 1, wherein the HSI component map in S2 is a component map composed of components of hue H, saturation S and brightness I, which is in accordance with the way that the human visual system perceives color.
3. The artificial intelligence fire extinguishing method with the open fire point identification function according to claim 1, wherein in S3, when the saturation S and the brightness I of each image frame both meet the threshold limits, the image frame area is defined as a candidate open fire point area, and when the saturation S and the brightness I of each image frame do not all meet the threshold limits, the candidate open fire point area does not exist.
4. The artificial intelligence fire extinguishing method with the open fire point recognition function according to claim 1, wherein in S5, the characteristic value array sequence is brought into a trained BP neural network, and the BP neural network outputs fire levels of candidate open fire point areas, wherein the fire levels are divided into six levels of 0-5; wherein, the 0 grade represents the non-fire state, and the 1-5 grades are arranged according to the severity degree of the fire condition in an increasing way.
5. An artificial intelligence fire extinguishing method with open fire point recognition function according to claim 1, characterized by using an artificial intelligence fire extinguisher apparatus with open fire point recognition function, said apparatus comprising: the fire extinguisher comprises a fire extinguisher body (4), a camera (1), a digital processing chip (2) and a control interface (3); wherein the content of the first and second substances,
the digital processing chip (2) comprises: the device comprises a component diagram conversion module (21), a flame identification module (22), a characteristic value array sequence acquisition module (23) and a fire situation grade acquisition module (24);
the camera (1) is used for shooting continuous image frames;
the component map conversion module (21) is used for converting each image frame into an HSI component map from an RGB model;
the flame identification module (22) is used for respectively setting a distinguishing threshold of saturation S and brightness I, comparing the saturation S and the brightness I in each image frame component image with the corresponding distinguishing threshold, and analyzing whether a candidate open fire point area exists or not;
when the candidate open fire point areas exist, the characteristic value array sequence obtaining module (23) is used for respectively obtaining the characteristic value arrays of the candidate open fire point areas of each image frame and obtaining a characteristic value array sequence containing a series of candidate open fire point area characteristic values;
the fire condition grade acquisition module (24) is used for judging the fire conditions of the candidate open fire point areas according to the characteristic value array sequence and generating the open fire ignition condition grade in a quantitative mode;
when the candidate open fire point area has a real open fire point, the control interface (3) opens a safety mechanism of the fire extinguisher body (4), and proper injection is carried out on the open fire point area according to the level of the open fire ignition condition, and the injection position is consistent with the shooting angle of the image frame.
6. The artificial intelligence fire extinguishing method with the open fire point recognition function according to claim 5, wherein the HSI component map in the component map conversion module (21) is a component map composed of components of hue H, saturation S and brightness I, and conforms to the way that the human visual system perceives colors.
7. The artificial intelligence fire extinguishing method with the open fire point identification function according to claim 5, characterized in that in the flame identification module (22), when the saturation S and the brightness I of each image frame both meet the threshold limit, the image frame area is defined as a candidate open fire point area, and when the saturation S and the brightness I of each image frame do not all meet the threshold limit, the candidate open fire point area does not exist.
8. The artificial intelligence fire extinguishing method with an open fire point recognition function according to claim 5, wherein the characteristic value array sequence obtaining module (23) calculates an area A of a candidate open fire point region and a region center of gravity point coordinate C (x, y) of the candidate open fire point region for each image frame, respectively, obtains an acquisition time T of each image frame, and takes the area A, the center point coordinate C (x, y), and the acquisition time T as a multivariate array of the characteristic values of the candidate open fire point region: and (A, C (x, y), T), combining the multiple arrays of the candidate fire point region characteristic values of each image frame to form a characteristic value array sequence:
wherein n is more than or equal to 24.
9. The artificial intelligence fire extinguishing method with the open fire point recognition function according to claim 5, characterized in that the fire level acquisition module (24) brings the array sequence of the characteristic values into a trained BP neural network, the BP neural network outputs the fire levels of the candidate open fire point areas, and the fire levels are divided into six levels of 0-5; wherein, the 0 grade represents the non-fire state, and the 1-5 grades are arranged according to the severity degree of the fire condition in an increasing way.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910901117.XA CN110812753B (en) | 2019-09-23 | 2019-09-23 | Artificial intelligent fire extinguishing method with open fire point identification function and fire extinguisher equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910901117.XA CN110812753B (en) | 2019-09-23 | 2019-09-23 | Artificial intelligent fire extinguishing method with open fire point identification function and fire extinguisher equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110812753A CN110812753A (en) | 2020-02-21 |
| CN110812753B true CN110812753B (en) | 2021-07-30 |
Family
ID=69548153
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910901117.XA Active CN110812753B (en) | 2019-09-23 | 2019-09-23 | Artificial intelligent fire extinguishing method with open fire point identification function and fire extinguisher equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110812753B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112861754A (en) * | 2021-02-23 | 2021-05-28 | 车主邦（北京）科技有限公司 | Abnormity processing method and device for electric energy supply station |
| CN113177496A (en) * | 2021-05-10 | 2021-07-27 | 浙江大华技术股份有限公司 | Fire point detection method, device, equipment and storage medium |
| CN115526896A (en) * | 2021-07-19 | 2022-12-27 | 中核利华消防工程有限公司 | Fire prevention and control method and device, electronic equipment and readable storage medium |
| CN114283367B (en) * | 2021-12-26 | 2022-09-06 | 特斯联科技集团有限公司 | Artificial intelligent open fire detection method and system for garden fire early warning |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0036071A1 (en) * | 1980-03-13 | 1981-09-23 | Caterpillar Tractor Co. | Fire suppression system |
| CN101577032A (en) * | 2009-06-02 | 2009-11-11 | 汕头大学 | Wireless fire detector for early fire recognition |
| CN101609589A (en) * | 2008-06-17 | 2009-12-23 | 侯荣琴 | Multi-frequency image fire detection system |
| CN101625789A (en) * | 2008-07-07 | 2010-01-13 | 北京东方泰坦科技有限公司 | Method for monitoring forest fire in real time based on intelligent identification of smoke and fire |
| CN103761529A (en) * | 2013-12-31 | 2014-04-30 | 北京大学 | Open fire detection method and system based on multicolor models and rectangular features |
| CN104299351A (en) * | 2014-10-22 | 2015-01-21 | 常州大学 | Intelligent early warning and fire extinguishing robot |
| CN106355809A (en) * | 2016-11-09 | 2017-01-25 | 宁波大红鹰学院 | Early warning and emergent processing system for forest fire |
| CN106650594A (en) * | 2016-10-09 | 2017-05-10 | 北方民族大学 | Video fire detection method, device and system |
| KR101749000B1 (en) * | 2016-08-11 | 2017-06-27 | 중앙대학교 산학협력단 | Fire detection method and apparatus using HD image |
| CN109985333A (en) * | 2019-04-10 | 2019-07-09 | 杭州富阳和翔模具有限公司 | It is a kind of for fire plant capable of automatic alarm in factory position warehouse |
-
2019
- 2019-09-23 CN CN201910901117.XA patent/CN110812753B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0036071A1 (en) * | 1980-03-13 | 1981-09-23 | Caterpillar Tractor Co. | Fire suppression system |
| CN101609589A (en) * | 2008-06-17 | 2009-12-23 | 侯荣琴 | Multi-frequency image fire detection system |
| CN101625789A (en) * | 2008-07-07 | 2010-01-13 | 北京东方泰坦科技有限公司 | Method for monitoring forest fire in real time based on intelligent identification of smoke and fire |
| CN101577032A (en) * | 2009-06-02 | 2009-11-11 | 汕头大学 | Wireless fire detector for early fire recognition |
| CN103761529A (en) * | 2013-12-31 | 2014-04-30 | 北京大学 | Open fire detection method and system based on multicolor models and rectangular features |
| CN104299351A (en) * | 2014-10-22 | 2015-01-21 | 常州大学 | Intelligent early warning and fire extinguishing robot |
| KR101749000B1 (en) * | 2016-08-11 | 2017-06-27 | 중앙대학교 산학협력단 | Fire detection method and apparatus using HD image |
| CN106650594A (en) * | 2016-10-09 | 2017-05-10 | 北方民族大学 | Video fire detection method, device and system |
| CN106355809A (en) * | 2016-11-09 | 2017-01-25 | 宁波大红鹰学院 | Early warning and emergent processing system for forest fire |
| CN109985333A (en) * | 2019-04-10 | 2019-07-09 | 杭州富阳和翔模具有限公司 | It is a kind of for fire plant capable of automatic alarm in factory position warehouse |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110812753A (en) | 2020-02-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110812753B (en) | Artificial intelligent fire extinguishing method with open fire point identification function and fire extinguisher equipment | |
| CN109522793B (en) | Method for detecting and identifying abnormal behaviors of multiple persons based on machine vision | |
| CN107944359B (en) | Flame detecting method based on video | |
| CN106600888B (en) | Automatic forest fire detection method and system | |
| Yang et al. | Research on a skin color detection algorithm based on self-adaptive skin color model | |
| CN110728223A (en) | Helmet wearing identification method based on deep learning | |
| US6421462B1 (en) | Technique for differencing an image | |
| CN108960185A (en) | Vehicle target detection method and system based on YOLOv2 | |
| JP2021089717A (en) | Method of subject re-identification | |
| CN112818953A (en) | Mask wearing state identification method, device, equipment and readable storage medium | |
| CN102609710B (en) | Smoke and fire object segmentation method aiming at smog covering scene in fire disaster image video | |
| CN110660187B (en) | Forest fire alarm monitoring system based on edge calculation | |
| CN114241383A (en) | Method and device for identifying building fire by image | |
| CN112949536A (en) | Fire alarm method based on cloud platform | |
| Goodall et al. | No-reference task performance prediction on distorted LWIR images | |
| CN111091586A (en) | Rapid smoke dynamic shielding area detection and positioning method and application thereof | |
| Munshi | Fire detection methods based on various color spaces and gaussian mixture models | |
| Kaliyev et al. | Computer Vision-based Fire Detection using Enhanced Chromatic Segmentation and Optical Flow Model | |
| CN115331152A (en) | Fire fighting identification method and system | |
| CN110245590B (en) | Product recommendation method and system based on skin image detection | |
| Akash et al. | Robust face detection using hybrid skin color matching under different illuminations | |
| CN109410169B (en) | Image background interference degree identification method and device | |
| CN112052912A (en) | Intelligent flame combustion state identification method for fire-fighting robot | |
| Ye et al. | A blurry image recognition method for straw burning detection | |
| Sun et al. | Flame Image Segmentation Method Based on Color Features |
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 |