TW201140505A - Surveillance video fire detecting and extinguishing system - Google Patents

Abstract

The present invention is a surveillance video fire detecting and extinguishing system having a control module, a surveillance video detector, an alarm module and an extinguishing robot. The control module continuously reads a video from the surveillance video detector and performs a surveillance video fire detecting method to determine and trace a fire-image in the video. When the fire-image is detected, the control module relocates the extinguishing robot in a new location in real space that is corresponding to the fire-image in the video, and the extinguishing robot executes a fire extinguish method to put the fire off.

Description

201140505 VI. Description of the Invention: [Technical Field] The present invention relates to a system for fire detection and automatic fire extinguishing, and more particularly to a system for determining fire occurrence and automatic fire extinguishing using captured images. L no prior technology] In recent years, several major fires have occurred in public places in succession, causing major casualties and losses of property of hundreds of millions of 7C, so that the public deeply felt that public places and residences are eves, not women, #能在第一Time, to warn the fire of the stars enough to poke the original, will greatly reduce the loss of life and property. , t, looking at the technology of the fire sensing device used and sold in the market: use such as particle sampling analysis, temperature sampling analysis or environmental analysis, so it is necessary to wait for a fire or Accumulation of particles caused by smoke: When the spread reaches the sensing range I of the sensed farm, the fire alarm signal will be released. The fire sensory gathering of the 詈, Γ 并不 does not provide the fire that causes the fire. '敁 size, flame burning Level and other information. The purpose of the day, the use of video surveillance, the use of security personnel to use a large amount of money:::: lending to achieve the fire prevention skills, in addition to the need to consume, in practice, the use of Bu 'Μ to store a large amount of video data It is a bit unrealistic to use the post-mortem inspection service. In order to overcome the fire 4, the eye is strong and so people can raise the V, and can be limited to the detection of the spot-like range." · The detection range of the smoke: π non-column optical sensing device to increase the fire system and It is necessary to install an infrared array sensing device, σ, and text to achieve the effect. Ε S3 3 201140505 [Summary of the Invention] In order to solve the aforementioned problem that the existing fire detection technology or method is too slow to detect the exact location of the fire and the degree of fire development, the present invention combines the video system ( Video surveillance) and visual sensing (Visual Sen/oing) to determine the location and state of the fire, can solve the traditional use of fixed-point sensing devices to detect fires such as Flame Sensor, temperature sensing (Temperature Sens〇), gas

The sensor detection (Gas Sensor) and other fire detection means are too slow to detect the exact location of the fire and the technical level of development, and the present invention further combines the results of the judgment with the automatic fire extinguishing system, The goal of the invention is to achieve automatic fire detection and automatic fire extinguishing. The invention provides a machine vision fire intrusion detection and automatic fire extinguishing system, which comprises a processing control terminal, an image 撷趑扨y, an alarm, and an alarm. The group and a remote fire extinguishing module, the center: the processing control image, and whether there is an image in the image to control the alarm mode visual fire sfl image position control to the physical space, . The infra-red image capturing device continuously reads a video and video, and the fire and the crying method determine the captured video-fire feature. When the system determines that the fire image has the fire feature, the group sends a 4e. · ° ' The processing control terminal continues with the machine: the method continuously tracks the state of the fire feature in the continuous reading; and the terminal controls the remote fire extinguishing module to move to the video shadow through the benefit, the, and the line signal佶* u S3 generates a fire k feature position corresponding to the adjacent 4 201140505 position 'automatically determines the control or remote control mode to perform a fire extinguishing means on the physical space corresponding position of the fire feature. The 'Shai remote control fire extinguishing module is a remote control platform' which carries a fire extinguishing device or material for performing the fire extinguishing means, and the remote fire extinguishing module generates a feedback signal to the processing control module to enable the processing control module The positional relationship between the remote fire extinguishing module and the fire feature is known to calculate the motion vector and the rotation vector of the remote fire extinguishing module to approximate the fire feature. Wherein the feedback signal is a light source signal. The step of the machine visual fire detection method includes: dividing a possible range of fire features in the video image: the video image is segmented by the moving object judgment means to be one of the video images a range in which the moving object determining means discriminates the target object having the movement and the swaying phenomenon in the video image as a possible range of the fire feature; ^ Sweater transition and fire feature correlation comparison: the fire feature may be first After a color conversion, and according to the result of color conversion and a comparison of the sample t-line comparison analysis, and generate - color spectrum phase relationship comparison value; analysis of the dynamic behavior of the fire characteristics _ determine whether the fire feature possible range - Geometric topology shape irregularity and a sudden movement characteristic, 2 calculate the possible range of the fire characteristic - turbulence ratio, the color spectrum correlation value and the turbulence ratio through a fuzzy logic calculation to obtain a probability index, After that, the death will be more likely to have the most southern possibility indicator of a fire feature. You will find out the existence and location of the fire features in the video shirt image from the &# ^ second position analysis; and S], the ruler shirt image area tracking: a motion tracking algorithm for video 5 201140505 The image tracks the change and location of the fire feature in the video image. Wherein, in the possible range step of the fire feature in the video image of the camera, the step-by-step color box directly identifies the fireman's special puzzle in the video image to complete the analysis of the dynamic behavior of the fire feature. After the step, the fire feature found is directly indicated in the video image. The 'discriminating moving object judgment means is a background segmentation algorithm for moving the historical state image. The motion tracking algorithm is a continuous adaptive mean tracking algorithm. The motion tracking algorithm is a continuous adaptive mean tracking algorithm. Therefore, the machine vision fire detection and automatic fire reduction system provided by the invention can analyze and judge whether the read video image has a fire feature in real time. 'Because there is no need to pass through a complicated detector, 4 does not need to wait for smoke or flame to approach. Detecting n', you can quickly perform fire detection and determine the exact location of the fire. After the fire location is determined, you can immediately dispatch a remote fire-extinguishing module to extinguish the fire to achieve immediate detection of the disaster. Technical effect. Please refer to the first figure, which is a system block diagram of a preferred embodiment of the machine vision fire detection and automatic fire extinguishing system of the present invention, comprising: a strict control terminal 10' - an image capturing device 2 〇, an alarm module 3〇: = remote fire extinguishing module 40. The processing control terminal 报 reports the image with the image capturing device 2〇201140505, and 30 Ray, 14, * · 罨丨 罨丨 ,, the processing control terminal 1Q is set by 2 0 拄 ut 田 々 々 々 Pick up, "3 Jia take - video image, and perform - machine vision fire detection method to judge the digging & s fire detection side breaks yell against you you especially feature 'by the judgment" The physical space like the corresponding shot is in the middle, the * «· 牯 * * * 玍 X fire, its fire X feature refers to the flame or smoke. Among them, 1 0 can be β _ / Θ processing control terminal 疋 0 PC, Server and other devices, I a search for the German, 4f this has to renew the entry into the visual ~ Xiangya implementation of the machine visual fire detection method! 4 whether there is a fire feature in the image of the thorn. Go ... i break far video" = After the discrimination of the fire detection method, it is determined that when the view == fire feature, the process control terminal 10 controls the core group 30 to issue an alarm, and the process control terminal visualizes the fire-inflammation - the target I, the + j Shi, the cavity 'only μ 〇 机 益 益 持续 持续 持续 持续 持续 持续 持续 持续 持续 持续 持续 持续 持续 持续 持续 持续The location and status of the video video " plastic, & u Λ s reported the sound signal (alarm sound; wireless fire notification signal (passed to the building administrator, fire station, etc.) ^ not in the process control (four) 1◦ At the same time, the visual control signal of the human-machine interface simultaneously controls the remote control fire-extinguishing module 40 to move into the physical space and the video image through the wireless signal by the processing control line 眭彳η, 乐ι ρμ, j, The adjacent position corresponding to the location of the fire feature is generated, and the physical space corresponding to the fire feature is automatically determined by the control or remote control mode. The direct control means. The so-called automatic judgment control or remote control means ## ^b, the remote fire extinguishing module 40 executes the control mode of the fire reduction means, and automatically determines the gas #4 t . · ' w remote fire extinguishing module 40 can automatically detect the fire feature close enough to automatically heat < The fire extinguishing means, on the other hand, the remote control mode means that the remote fire extinguishing module 40 is connected to the 兮 油 oil Λ Λ and the 5 hai processing control terminal 1 〇 is controlled to execute the fire extinguishing means. [S3: 7 201140505 Among them, the shai remote control fire extinguishing module 4〇 can It is a remote control platform (remote control car, machine benefit person, crawling car...), which carries a fire extinguishing device or a fire extinguishing material to carry out the fire extinguishing means. The type of the fire extinguishing means is not limited, and may be dry powder, / bag / In order to allow the process control module 1 to drive the @remote fire extinguishing module 40 into the fire, the video image captured by the image capture device 20 is continuously controlled to control the remote fire extinguishing module. The group 4 〇 continuously moves the position to approach the fire feature, and the remote fire extinguishing module 40 can further have a signal generating unit, wherein the feedback signal generating unit can generate a light source generating component (for example, a light emitting diode)羌源 (丨 jght emjttjng diode, LED)) or a wireless position signal generator (such as GpS positioning), the edge feedback signal generating unit generates a feedback signal for the processing control group 10 to know the remote fire extinguishing The positional relationship between the module 4〇 and the fire feature enables the process control module 1 to continuously adjust the relative position between the remote fire extinguishing module 4〇 and the “Xuanhuohuo”, and Driven by the remote module 40 to execute the fire extinguishing means. The feedback signal generating unit of the present embodiment is a light emitting diode. When the remote fire extinguishing module 40 enters the fire field, the video image captured by the image capturing device 2 can not only capture the fire feature. At the same time, it can be photographed. The remote control fire extinguishing module 40 and the light emitting diode. Since the characteristics such as the wavelength and shape of the light-emitting diode are controllable factors and can be set and planned in advance, the processing control terminal 10 can discriminate the video from the video image through simple image processing and execution of the discrimination program. The position signal relationship between the feedback signal generating unit and the fire feature is determined by the light source signal generated by the signal generating unit, and thus the processing control terminal can be returned through the remote fire extinguishing module 40. The light source signal is given to adjust and control the 8 201140505 remote fire extinguishing module 40. The direct fire extinguishing module 40 approaches the fire J = suspected vector, and drives the remote control to achieve the purpose of fire extinguishing. In terms of the actual test user % + a, use, the present embodiment uses a board-mounted surveillance camera as the 兮&## 'to page stencil" shirt-like capture device 20 to capture the surrounding one-view shirt For example, the video image transmission v^, U and the processing control terminal 20 are further analyzed, and then the fuzzy theory is used to push '^^ ώ which pushes the linear motion of the 3H remote fire extinguishing module 40. Vector and rotation vector, while paying J and establishing correction angle with shovel &

The rule base of the fuzzy system; therefore, when the image capture device: the current control device and the smoke, the process control terminal 1G can calculate the remote control mode, and 40 and the distance and direction of the fire feature, seeking to correct the positive angle of the vector disk And according to the analysis correction vector and the correction angle, the speed of the remote fire extinguishing module 40 is planned, and the command is transmitted to the remote fire extinguishing module 40 by wireless communication. 〃In order to further explain the visual fire detection method of the machine, please refer to the second figure and the third article, which are examples of the preferred implementation process of the machine vision fire detection method of the present invention, and the steps thereof include: (51) Reading the video image: The video image 60 is continuously and instantly read by the image capturing device 2Q. In actual execution, the image capturing device 2〇 may be a monitor having a color recording effect, a camera, or a CCD (Charge_COUp|ed device) or a light sensing element such as a CM〇s. (53) The possible range of the fire feature in the divided video image: the video image 60 is read by the multi-animal image _+ segment and the fire feature range 68 in the captured video image 60 is divided. Since the fire is caused by a burning of the substance and a phenomenon of turbulence, the flame of the fire itself has a special color, shape and spread type. 9 201140505 State These models provide an important reference for fire identification. The process of burning flames involves the behavior of chemical changes and nep flow, and the fire talks and has the characteristics of flashing. Another characteristic of the flame is that the shape of the flame changes with the flow of the air stroke, and there will be sharp jitter. And sudden movements, etc., so the smoke generated by the flame will also produce jitter as the flame moves. Therefore, the present embodiment utilizes the fire characteristics (ie, fire and smoke) that may occur in a fire, and the movement, shaking, and the like of the fire characteristics of the sea, and finds the object with movement and jitter in the image of the image. And the subject matter is segmented and selected as the fire feature possible range 681. There are many calculation methods for finding and discriminating the moving object of the afl image. The moving object determining means of this embodiment is a background segmentation algorithm of a Motion History Image (MHI). The MHI algorithm is mainly used for computer vision in the analysis and movement of gesture behavior [j. Davis, Recognizing movement using motion histograms, " Technical Report 487, MIT Media Lab, 1999.; JW Davis • and A. Bobick, &quot The Representation and Recognition of Action Using Temporal Templates," IEEE Transactions on Pattern Analysis and Machine, Intelligence, Vol. 23, No. 3, pp. 257-267, 2001. ; GR Bradski and JW Davis, "Motion segmentation and Pose recognition with motion history gradients," Machine Vision and Applications, vol. 13, pp. 174-184, 2002.], the MHI algorithm can be used not only to determine the position of the current object, but also to utilize the object in the video image 60 The motion information in the scene in the signal to segment and measure these motions. The regions that are segmented 201140505 are not motion blocks " but are naturally connected to the motion part of the object. The Moving History State Image (MHI) algorithm is mainly used. To describe the state of motion of an object in an image, and to use color depth for each pixel This example uses the MH丨 algorithm to have the following characteristics: () The moving history state image that occurs within a certain period of time can be represented by a single grayscale image; (2) can be directly The area marked by the movement is; (3) The amount of CPU calculation using MHI is not large, so dynamic detection of fire and smoke can be realized.

Further, after the step is completed, the selected fire feature possible range 681 can be directly marked in the video image 6〇 by a color frame, so that the user can directly see the result after the calculation, the selected result, and the warning is used. The main/video video 6〇 corresponds to various possible conditions on the scene, as shown in the third figure. The video image 6〇 shown in the third figure contains two possible fire feature possible ranges 681, a flame 62 (right pattern) and a red leaf pot 65 (left graph). For example, the red leaf pot may be judged by the MH丨 algorithm after being swayed by the wind, and is marked as a fire feature with a range of 6 81. (55) Correlation comparison of color conversion and fire characteristics. In order to segment the image pixels of flame and smoke possible in the fire feature, this embodiment marks the area in which the previous step (53) is moved (ie, the foregoing: the fire feature) Possible range 681) 'First through a color conversion process, the color image information of the fire feature 681 of the selected Z can be converted into a color space mode that is relatively easy to process, thereby improving the performance and effect of subsequent operations. For example, the color conversion processing of the present embodiment converts the color image information of the possible range 681 of the 2 disaster features into a %(1) 201140505

Saturation, Intensity) color space mode. Choosing the HIS color space Because the HIS color space is derived from the human visual system, the color is described and defined by Hue, Saturation or Chroma, and Brightness (丨ntensity or Brightness). This conical model describing the HSI color space is quite complex, but it does show the changes in hue, brightness, and color saturation. Since human vision is much more sensitive to brightness than sensitivity to color shading, in order to facilitate color processing and recognition, the human visual system often uses the HSI color space, which is more in line with human visual characteristics than the RGB color space. A large number of algorithms in image processing and machine vision are conveniently used in the H SI color space, which can be processed separately and independently of each other. Therefore, the H 丨 color space in this embodiment can greatly simplify the workload of image analysis and processing, and can realize dynamic detection of flame and smoke. During the calculation, the parameters of the HS丨 color space can be normalized to the following ranges: 〇° hues 360° ; saturation^ 255 ; A 0 S intensity S 255. In order to find out the HSI color distribution range of the flame and the smoke, the fire feature possible range 681 of the HIS conversion can be compared and analyzed with a comparison template to define whether the captured image information of the possible range of the fire feature 681 is It is indeed a flame or smoke. The establishment of the comparison template may be to capture a set of video images 60 stream pictures in advance for fire characteristics generated by different environments and conditions. For example, the comparison template can be taken from a HIS color space database of fire features generated under ambient or outdoor conditions such as indoors and different materials. 201140505 In other words, after analyzing the samples comparing flames and smoke, just. The real-time video input image screen and the experimental statistics are built. See the 戎 comparison template, and the histogram color similarity comparison meter is different, and the correlation comparison value is obtained to represent It may be the score of the similarity of the flame or the smoke. & (57) Analysis of the dynamic behavior of fire characteristics 1 Analysis of the turbulence characteristics of fire characteristics In order to reduce the misjudgment caused by the area or object similar to the color of the flame and smoke, it is possible to analyze the fire characteristics extracted by the range 68彳The dynamic behavior is used to distinguish image areas in video with the important means of similar color characteristics to determine the correct flame and smoke. Among them, the dynamic behavior of the fire feature includes the geometrical topology irregularity of one of the flame and the smoke, and the characteristic of a sudden movement. Since both the flame and the smoke are fluids, there will be irregularities (irrigation of geometric topological shapes) and turbulent flow phenomena. In this embodiment, the turbulence phenomenon is used to determine and distinguish between the realities. The important basis for the occurrence of flames and smoke and similar backgrounds and objects (such as people moving in red clothes), when the turbulence of flames and smoke increases, the turbulence ratio Ω of the video image will also follow Increase. Where ‘the turbulence ratio Ω is defined as: Ω —, --(1-1) where ' P is the perimeter of the selected area (perj meter) and A is the area of the selected area. When the shape complexity of the selected area increases (ie, p becomes larger, A becomes smaller), [S3 13 201140505 Abandoned flow is increased by Ω] Therefore, the extravagant flow ratio analysis can be used to determine whether the selected fire feature range 681 is indeed For flames or smoke. 2. Time-domain analysis of fire characteristics In addition, the characteristics of fire talk and smoke will be used over time, and it is also used as an important basis for measuring fire and smoke, because the characteristics of flash can cause fire talk in video. With smoke images, intermittent irregularities appear and disappear; in order to distinguish similar image areas from flames and smoke and analyze the time-domain nature of flame and smoke flashing over time, this embodiment uses fuzzy logic (fuZZy-|0giC -enhanced app "〇ach" to find a possible index of flame and smoke. 67, as shown in the fourth figure. The fuzzy logic used in this embodiment has two inputs, one for the color spectrum of the flame and the smoke. The sexual comparison value and the turbulence ratio Ω, which first normalize SCQrr and Ω to the numerical range of 〇~彳, and generate a probability index u of the fire characteristic after a fuzzy logic broadly. The calculation of Lu correlation is based on the following formula: · ^(,4, B) = - B) (12) \ίΣ((Αί ~ - B)2 where 'A and B respectively represent the video image to be analyzed Fire characteristics can be used in this garden and the comparison template statistics Fruit, Ai values representative of statistical results' A bar represents the mean. 201140505

The fourth block of the work blocks Scorr and Ω normalizes the values of the rounds into GsScorr and 〇/; to the range of [〇,1] and produces a flame and smoke probability index w. GsScorr and G. The values are composed of four regions and use the two-angle fuzzy function, which are ZE (zero, none), pS (p0Sitive sman, small), PM (positive middle, middle), and PL (p〇sitjve | arge). , large), and based on the experimental analysis of the image database, establish a fuzzy (Ruzzy) rule base, as shown in Table 1. The output of the fuzzy logic calculus contains four output values (singletons) {#1, v 2, #3, v 4}. This embodiment uses weighted average defuzzmcati〇n

Flanie-and-smokp index us ΖΕ PM PS PM PL Ω ΖΕ Hi il2 PS /// Hi li2 PM Ul lh U4 PL /// ... "4 u4 Calculate the probability ^ standard ^. The probability indicator ^ The calculation formula is: if = Gi Σ/=ι ... (2) where Gu is another constant (sca|ing constant), and A (t// ) is the result of the above several relational parameters ( In the case of real (four) calculation, this embodiment takes about 20 historical image frames (historical frames) possibility index u to find the most π-like area of the video image in the video image. The region has the highest probability', then the image region continues to perform a time domain analysis 201140505 (temporal analysis) ° In order to better distinguish objects in the video image 60 that have similar characteristics to the fire feature, this embodiment utilizes an average level Leap rate (Leve| Cr〇ssjng

The Rate ' LCR ) algorithm is used to distinguish objects that have similar properties to flames and smoke. Each pixel (pjxe|) in the selected image region is subjected to LCR calculation for time domain analysis: 1 7'-' LCR(a·, y) = — [ II(w, > /ij}, · · (3) /=0 # where Ut is the probability of length ( (a Probability 〇f length T). ΙΚΦ} is an indication function indicating the function &amp when the element (a "gument" is true (true) 1, otherwise, it is 〇. 々1 is the threshold value. In this embodiment, the image length T used for the lcr calculation is 40 frames, and the judgment threshold Μ is set. value.

Therefore, the spatial position of the fire feature of the video image 6〇 of the present embodiment can be determined by the following formula (4):

Flume Sm〇ke(.v. v) = lTRUE· if LCR ^ 1 FALSE, if lcr<a-2. ...(4) ” The critical value of the mussel test, (χ, y) represents the fire characteristic The k2 threshold value of the vacant embodiment is an experimental result obtained by observing the video image 6 of the h feature, and the value of the embodiment is 〇.〇5 when performing the smoke determination. When the flame is judged, the value is 〇. 〇7 5. By the above judgment step 6〇, whether there is flame or smoke, 'the position of the video image and the flame and smoke can be effectively judged. 16 201140505 (59}Fire characteristics Image area tracking Through the above steps, it is possible to clearly define whether there is a fire feature and its possible range in the video image. In order to continuously track the range and trend of the fire feature, the viewer can continue to track the video image. In the present embodiment, the position of the fire feature is continuously tracked by the video tracking algorithm by a motion tracking algorithm, and the video image 60 is appropriately marked to enhance the warning effect. Use of this algorithm is a continuous motion 追縱 mean 追縱 adaptive algorithm (CAMS butoxy, Adaptive c_inu〇usly

Mean-S_)'CAMSHIFT mainly achieves the purpose of tracking through the color information of moving objects in the video image. The camsh|ft algorithm uses color probability distribution and statistics.

The alpha mode holds ten and traces the twins in the flame and smoke images. Therefore, the CAMSH|FT and ·^ methods use the color histogram to make a difference. The probability distribution of the color in the shirt image can be CAMSH丨FT algorithm. Handling dynamic color distribution changes, the specific steps include · ^ Step 1. Set the entire image as the search area. =Step 2. Initialize the position and size of the tracking image frame window. Area ΓΓ:· Calculate the color probability distribution in the tracking image frame window. This 〇〇 ' is larger than the tracking image frame window. /Step 4. Operation Run Mean Training (1) The new position and size of the frame window. In order to get the image, step 5 y- -T- 3 to get the value of the initial servant:! Like the video image of the frame 6", using the step °, the position and size of the image frame window f #steps 3'4 and 5, the actual: magic, and repeat the aforementioned flute two solid, feature witness (flame Tracking with smoke. — As an example, after the CAMSH|F Ding algorithm 201140505 video image 60, the fire feature 6 8 2 can be found in the video image, and the video image 60 is marked and tracked. In combination with the foregoing, in this embodiment, the effective region R〇丨(Regi〇n of Interest) divided by the moving history state image algorithm (Μ Η |) is the aforementioned fire feature possible range 681, Then, using the correlation calculation of the color spectral correlation value and the turbulence ratio in the space, the time domain analysis of the smoke and the smoke are used to eliminate the objects that may be confused in the video image 60 (such as wearing red clothes everywhere). The moving person or car, etc.) finally uses the CAMSHIFT algorithm to track the image motion area of the flame and the smoke. Further, in this embodiment, in addition to the position of the flame and the smoke for the video image. In addition to the eye-catching indication, a warning signal can be issued after the fire feature found after step (57) or (59) (by receiving the video image computer to issue a warning signal by voice or wireless signal), so that not only can the need be made Persons who continue to monitor are warned 'to prevent the disaster from continuing to expand. This example develops a video visual algorithm that uses machine vision algorithms to process the input shirt color image, and develops a flame and smoke detection algorithm based on color characteristics and space. Features, time domain characteristics to detect fires in a timely manner, and alarms when flame and smoke signals are detected. Remote monitoring can also be achieved if visual servo technology is implemented with the hardware and software of a single-chip system. Or even remote monitoring and control of the technical effect of the fire extinguishing. Through the testing of different video images 60, this embodiment can quickly and effectively detect and trace the location of fire features in the video image 60. The present invention develops a use Machine vision algorithm to process the video data of the input color image, using algorithm based Color characteristics, space special 18 201140505 Sex, time domain characteristics to timely fire I ', j fire, and prevent, and detect the flame and smoke signal inside the ice _V- MA 5X alarm. At the same time, through the cooperation For example, the soft ^ ^ ^ ^ ^ of the internal immersive early wafer system is more visually responsive to the visual servo technology, so that it can be used as a remote control fire-extinguishing module to capture the device 20 channels 5! The 妲 妲 b wind and fire 杈 group 40 fire extinguishing, the test results show that the calculus, # hair, and versatile can effectively and reliably detect the flame and smoke & and can continue the image of fire and smoke [Brief Description] The figure is a block diagram of a system according to a preferred embodiment of the present invention. The first figure is a flow chart of a preferred embodiment of the present invention. The present invention is a video image representation of a preferred embodiment of the present invention. The fourth figure is a schematic diagram of a fuzzy logic calculus block in accordance with a preferred embodiment of the present invention. [Main component symbol description] 1 0 processing control terminal 20 image capturing device 30 alarm module 40 remote fire extinguishing module 6〇 video image 62 flame 65 red leaf potted plant 681 fire feature possible range 682 fire feature [S] 19

Claims (1)

201140505 VII, Shenming patent scope: Yan Li: 2: machine vision fire system and automatic fire extinguishing system, including - processing terminal, _ image silk fire extinguishing module, wherein: H group and - remote control = control terminal by the image operation The device continuously reads a video image—the visual fire test method determines whether there is a fire feature in the captured video image; 4 the processing control terminal determines that the video image is stored==: the group is sent, and the processing control terminal The second and second detection methods continue to track the location and status of the fire image in the continuous reading of the Shih Hing video; and the ^ ^ to the apricot terminal controls the remote fire extinguishing module through the wireless signal: The video image generates a fire-frozen feature position corresponding to the proximity "1 pair of positions: the judgment control or remote control mode performs a fire-fighting means on the corresponding positions of the physical features of the fire feature. Self-service ^, 2 towel 4 patent model 111 item 1 The machine 11 visual fire accumulation test and 0, the system of the remote control fire extinguishing module is - remote control load 3 cut material (four) line (four) fire means, material (four) fire mode = raw 7 signal number to the treatment The module is configured to calculate the positional relationship of the fire module to calculate the remote control, the vector and the rotation vector, and approximate the fire feature. 3_ The machine is automatically fire extinguished as described in claim 2 The system, the feedback signal is - the light source signal. U measurement and (the application of the scope of the device as described in the scope of the first or second or third item of the device fire detection and automatic fire extinguishing system 'the machine visual fire detection method Step 20 201140505 includes: dividing a possible range of the fire feature in the video image: capturing, by the moving object determining means, a possible range of the fire feature in the video image, wherein the moving object determining means Determine the view. The target object with movement and jitter in the shirt image as the possible range of the fire feature; color conversion and fire feature correlation comparison: the fire feature may The range is first and second color conversion, and the result of the color conversion is compared with a comparison template, and a color spectrum correlation relationship is generated; the dynamic behavior of the flame characteristic is determined: whether the possible range of the fire feature is determined :Preparation - geometric topology shape irregularity and a sudden movement characteristic, and ° ten, 4 flame characteristics may be the norm mt*, the color spectral phase, the sexual comparison value and the turbulence ratio obtained through the simulation - Possibility: After the private will have the highest probability index - fire feature possible range:: domain: analysis and spatial position analysis, find out the presence and location of the fire feature in the video '5V image _; and the fire rule feature image area Tracking: Tracking the change and position of the fire feature in the video image with a motion ^, image to the video. For example, the automatic fire extinguishing system described in the fourth application of the patent scope, wherein: the eight-foot 1 Belehan video segmentation video In the fire, a special cow w — 铕 铕 厂 古 古 古 J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J J Fire characteristic in the video image directly after the labeling of the fire characteristic [Sl identified to complete the analysis of the dynamic behavior of the fire characteristic step. 21 201140505 Automatic: The machine vision fire detection and automatic fire extinguishing system described in the fifth item, which makes the background segmentation algorithm of the moving object state image. Slave is a history of movement 7. If you apply for a patented automatic fire extinguishing system, the machine vision fire detection and tracking algorithm described in this article. , the law is - the average value of continuous adaptation 8_ such as the patent application automatic bonfire & μ brother 6 said Ma, the violation, ~ system, the motion tracking, 机器 machine vision fire detection and algorithm . The algorithm is a * immediate ~ continuous adaptive mean, schema: such as the next page m 22