WO2019026518A1 - Fire identification device - Google Patents

Fire identification device Download PDF

Info

Publication number
WO2019026518A1
WO2019026518A1 PCT/JP2018/025447 JP2018025447W WO2019026518A1 WO 2019026518 A1 WO2019026518 A1 WO 2019026518A1 JP 2018025447 W JP2018025447 W JP 2018025447W WO 2019026518 A1 WO2019026518 A1 WO 2019026518A1
Authority
WO
WIPO (PCT)
Prior art keywords
storage unit
image
fire
image storage
fire point
Prior art date
Application number
PCT/JP2018/025447
Other languages
French (fr)
Japanese (ja)
Inventor
武藤　佳恭
加流慮 井塚
友也 北
進 栗岡
博明 谷川
山村　太一
Original Assignee
モリタ宮田工業株式会社
武藤　佳恭
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by モリタ宮田工業株式会社, 武藤　佳恭 filed Critical モリタ宮田工業株式会社
Publication of WO2019026518A1 publication Critical patent/WO2019026518A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
    • G08B17/00—Fire alarms; Alarms responsive to explosion
    • 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
    • G—PHYSICS
    • G08—SIGNALLING
    • G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
    • G08B25/00—Alarm systems in which the location of the alarm condition is signalled to a central station, e.g. fire or police telegraphic systems
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N7/00—Television systems
    • H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast

Definitions

  • the present invention relates to a fire identification device used in a fire extinguishing facility.
  • a flame detection apparatus one using a CO 2 resonance radiation band centered on 4.5 ⁇ m which is wavelength light unique to flames is known. Further, in Patent Document 1, a flame candidate region is specified by performing binarization processing of time-series images and comparing successive images before and after, five characteristics such as an elliptical coincidence rate and a change amount of average luminance A flame detection device has been proposed that discriminates flames from quantity.
  • the infrared imaging device for imaging the CO 2 resonance radiation band has a problem that the cooling structure requires a complicated configuration and is large and expensive.
  • the process of specification of a flame candidate field, and a flame judging is complicated.
  • it is possible to extract flames of small danger that would not lead to a self-extinguishing fire if left alone. Detecting flames with such a low risk also places a burden on work such as rushing on the spot and investigation and reporting of fires not only for building users but also for fire brigade personnel who have jurisdiction over the area where the building is located. There is a possibility to increase.
  • an object of the present invention is to provide a fire identification device capable of excluding a fire point having a low risk and identifying a high fire point which may lead to a fire.
  • the fire identification apparatus of the present invention is characterized in that a first determination unit that determines the occurrence of a fire point due to combustion or flame in a caution area, and that the first determination unit determines the occurrence of the fire point.
  • a fire identification apparatus comprising: a second determination unit that determines growth of a fire point; an imaging unit configured to image the inside of the caution area; and an image storage unit configured to store an image captured by the imaging unit together with time series information
  • the second determination unit compares the plurality of images stored in the image storage unit based on the time-series information, and includes the shape and size of the fire point included in the image; It is characterized in that the growth of the fire point is determined from a time-series change of a flame element including at least one of color, brightness, and fluctuation.
  • the time-series change pattern storage for storing the time-series change pattern of the flame element, wherein it is determined that the fire point is growing.
  • the second determination unit compares the time-series change of the flame included in the image with the time-series change pattern stored in the time-series change pattern storage unit. It is characterized by judging.
  • the thermal sensor for detecting a temperature distribution in the warning area is provided, and the first determination unit determines the temperature distribution.
  • a fire point pattern storage unit for storing the fire point pattern of the fire element that determines that the fire point has occurred.
  • the first determination unit determines that the flame element included in the image stored in the image storage unit is compared with the fire point pattern stored in the fire point pattern storage unit. It features.
  • the fire point pattern storage unit storing the fire point pattern of the flame element that determines that the fire point has occurred
  • the fire point pattern stored in the image storage unit, the fire element included in the first image, and the fire element included in the second image captured after the first image are stored in the fire point pattern storage
  • the second determination unit determines at least the third image captured after the second image and stored in the image storage unit. It is characterized in that the time-series change of the flame included in one image is determined by comparing with the time-series change pattern stored in the time-series change pattern storage unit.
  • the imaging unit includes a first imaging unit and a second imaging unit
  • the storage unit includes a first image storage unit and a second image storage unit
  • the first image storage unit stores the image captured by the first imaging unit together with the time-series information
  • the two-image storage unit stores the images captured by the second imaging unit together with the time-series information
  • the second determination unit is included in the plurality of images stored in the first image storage unit.
  • the photographing unit includes a first photographing unit and a second photographing unit.
  • the storage unit includes a first image storage unit and a second image storage unit, and the first image storage unit stores the image captured by the first imaging unit together with the time-series information, and The two-image storage unit stores the images captured by the second imaging unit together with the time-series information, and the second determination unit is included in the plurality of images stored in the first image storage unit.
  • the imaging section includes a first imaging section and a second imaging section, and the image storage section is a first image storage section.
  • a second image storage unit wherein the first image storage unit stores the image captured by the first imaging unit together with the time-series information, and the second image storage unit includes the second image storage unit.
  • the image captured by the imaging unit is stored together with the time-series information
  • the first determination unit compares the flame element included in the image stored in the first image storage unit with the fire point pattern And temporarily determine the flame element contained in the image stored in the second image storage unit by comparing it with the fire point pattern and temporarily determining the image stored in the first image storage unit As a result of temporary determination by the second image storage unit, and the previous stored in the second image storage unit Temporary determination result by the image, and judging said fire point if there is tentative determination of at least one has occurred.
  • the imaging section includes a first imaging section and a second imaging section
  • the image storage section is a first image storage section.
  • a second image storage unit wherein the first image storage unit stores the image captured by the first imaging unit together with the time-series information, and the second image storage unit includes the second image storage unit.
  • the image captured by the imaging unit is stored together with the time-series information, and the first determination unit compares the flame element included in the image stored in the first image storage unit with the fire point pattern And temporarily determine the flame element contained in the image stored in the second image storage unit by comparing it with the fire point pattern and temporarily determining the image stored in the first image storage unit As a result of the temporary determination by the second image stored in the second image storage unit. Provisional judgment result by, and judging with the fire point by the provisional determination of both has occurred.
  • the photographing unit includes a first photographing unit and a second photographing unit
  • the image storage unit includes: The images captured by the first imaging unit and the second imaging unit are stored together with the time-series information, and the second determination unit is configured to store the images captured by the first imaging unit and the second imaging unit. It is characterized by using an originally generated three-dimensional image.
  • the photographing unit includes a first photographing unit and a second photographing unit
  • the image storage unit includes the first photographing. An image captured by the unit and the second imaging unit is stored together with the time-series information, and the first determination unit is generated based on the image captured by the first imaging unit and the second imaging unit It is characterized by using a three-dimensional image.
  • the fire point growth is determined from the time-series change of the fire element, thereby excluding the fire point having a small risk and leading to the fire. It is possible to identify such dangerous fire points.
  • Block diagram showing the fire identification device according to the first embodiment of the present invention as a function realizing means Flow chart showing the operation of the fire identification device
  • the block diagram which represented the fire identification apparatus by 2nd Example of this invention by the function implementation means Flow chart showing the operation of the fire identification device
  • the block diagram which represented the fire identification apparatus by 3rd Example of this invention by the function implementation means Flow chart showing the operation of the fire identification device
  • the fire identification apparatus in the second determination unit, a plurality of images stored in the image storage unit are compared based on time-series information, and the fire point is included in the images.
  • the growth of the fire point is determined from the time series change of the flame element including at least one of the shape, the size, the color, the brightness, and the fluctuation.
  • the fire point having a small risk is excluded, leading to a fire. It can identify dangerous hot spots.
  • a second embodiment of the present invention is a fire identification apparatus according to the first embodiment, wherein a time-series change pattern of storing a time-series change pattern of fire elements, which determines that the fire point is growing, is stored.
  • a storage unit is provided, and the second determination unit determines the time-series change of the flame included in the image in comparison with the time-series change pattern stored in the time-series change pattern storage unit.
  • more accurate discrimination can be instantaneously performed by storing beforehand, for example, a time-series change pattern having a high risk of causing a fire, which is patterned using machine learning. be able to.
  • a third embodiment of the present invention is a fire identification apparatus according to the first or second embodiment, further comprising a thermo sensor for detecting a temperature distribution in the warning area, and the first determination unit If the position of the highest temperature is temporarily determined to be the fire point, and the temperature at the temporarily determined fire point is equal to or higher than the threshold value, it is determined that the fire point occurs, and the first determination unit determines the occurrence of the fire point.
  • the department will start shooting.
  • the use of the thermo sensor makes it possible to generate the fire point at an early stage.
  • the fourth embodiment of the present invention has a fire point pattern storage unit for storing a fire point pattern of a fire element that determines that a fire point has occurred in the fire identification apparatus according to the first or second embodiment.
  • the first determination unit determines the flame element included in the image stored in the image storage unit in comparison with the fire point pattern stored in the fire point pattern storage unit. According to the present embodiment, it is possible to instantaneously determine the occurrence of the fire point with higher accuracy by storing the fire point pattern which is patterned using, for example, machine learning in advance.
  • the fifth embodiment of the present invention is a fire identification apparatus according to the second embodiment, including a fire point pattern storage unit for storing a fire point pattern of a fire element that determines that a fire point has occurred.
  • the fire point pattern storage unit stores the fire element included in the first image and the fire element included in the second image captured after the first image, stored in the image storage unit.
  • the time-series change is determined by comparison with the time-series change pattern stored in the time-series change pattern storage unit.
  • the generation of the fire point and the growth of the fire point are determined separately, and the determination of the growth of the fire point is determined by more images than the determination of the generation of the fire point. For example, using fire point patterns and time-series change patterns patterned using machine learning to determine occurrence and growth of the fire point, a fire point with a high risk of leading to a fire is identified with high accuracy it can.
  • the sixth embodiment of the present invention is the fire identification apparatus according to any one of the first to fifth embodiments, comprising a first imaging unit and a second imaging unit as the imaging unit, and an image
  • the storage unit includes a first image storage unit and a second image storage unit.
  • the first image storage unit stores an image captured by the first imaging unit together with time-series information
  • the second image storage unit And an image captured by the second imaging unit is stored together with the time-series information
  • the second determination unit is configured to generate the fire point from the time-series change of the flame element included in the plurality of images stored in the first image storage unit.
  • the temporary determination of the growth, the temporary determination of the growth of the fire point from the time-series change of the flame element included in the plurality of images stored in the second image storage unit, and the temporary storage of the image stored in the first image storage unit As a result of the determination, and as a result of the temporary determination by the image stored in the second image storage unit, at least one of them If there is a temporary decision of the one in which it is determined that the fire point is growing.
  • the growth of the fire point can be identified at an early stage.
  • a seventh embodiment of the present invention is a fire identification apparatus according to any one of the first to fifth embodiments, which has a first imaging unit and a second imaging unit as an imaging unit, and an image
  • the storage unit includes a first image storage unit and a second image storage unit.
  • the first image storage unit stores an image captured by the first imaging unit together with time-series information
  • the second image storage unit And an image captured by the second imaging unit is stored together with the time-series information
  • the second determination unit is configured to generate the fire point from the time-series change of the flame element included in the plurality of images stored in the first image storage unit.
  • the temporary determination of the growth the temporary determination of the growth of the fire point from the time-series change of the flame element included in the plurality of images stored in the second image storage unit, and the temporary storage of the image stored in the first image storage unit
  • the determination as a result of the temporary determination by the image stored in the second image storage unit, as a result of both temporary determinations One in which it is determined that the point is growing.
  • it is possible to identify with high accuracy the growth of the high-risk fire point.
  • the eighth embodiment of the present invention is a fire identification apparatus according to the fourth embodiment, comprising a first imaging unit and a second imaging unit as an imaging unit, and an image storage unit as a first image storage unit.
  • An image storage unit and a second image storage unit are provided.
  • the first image storage unit stores an image captured by the first imaging unit together with time-series information
  • the second image storage unit includes the second imaging unit.
  • the photographed image is stored together with time-series information
  • the first judgment unit provisionally judges the flame element included in the image stored in the first image storage unit with the fire point pattern, and stores the second image.
  • the flame element contained in the image stored in the unit is temporarily determined by comparison with the fire point pattern, and the result of the temporary determination using the image stored in the first image storage unit is stored in the second image storage unit As a result of the temporary judgment by the image, if there is at least one temporary judgment that the fire point is generated One in which a constant. According to the present embodiment, it is possible to identify the occurrence of the fire point at an early stage.
  • the ninth embodiment of the present invention is a fire identification apparatus according to the fourth embodiment, comprising a first imaging unit and a second imaging unit as an imaging unit, and an image storage unit as a first embodiment.
  • An image storage unit and a second image storage unit are provided.
  • the first image storage unit stores an image captured by the first imaging unit together with time-series information
  • the second image storage unit includes the second imaging unit.
  • the photographed image is stored together with time-series information
  • the first judgment unit provisionally judges the flame element included in the image stored in the first image storage unit with the fire point pattern, and stores the second image.
  • the flame element included in the image stored in the unit is temporarily determined by comparison with the fire point pattern, and the image temporarily stored in the first image storage unit is temporarily stored in the second image storage unit.
  • the present embodiment it is possible to identify the occurrence of a high risk fire point with high accuracy.
  • a tenth embodiment of the present invention is a fire identification apparatus according to the first or second embodiment, including a first imaging unit and a second imaging unit as imaging units, and the image storage unit includes: A three-dimensional image generated based on the images captured by the first imaging unit and the second imaging unit, storing the images captured by the first imaging unit and the second imaging unit together with the time-series information Is used. According to the present embodiment, the determination accuracy can be further enhanced by using a three-dimensional image.
  • An eleventh embodiment of the present invention is the fire identification apparatus according to the fourth embodiment, comprising a first imaging unit and a second imaging unit as the imaging unit, and the image storage unit comprises: (1) The images taken by the imaging unit and the second imaging unit are stored together with time-series information, and the first determination unit generates a three-dimensional image generated based on the images taken by the first imaging unit and the second imaging unit It is used. According to the present embodiment, the determination accuracy can be further enhanced by using a three-dimensional image.
  • FIG. 1 is a block diagram showing a fire identification apparatus according to the present embodiment as function realizing means.
  • the fire identification apparatus according to the present embodiment includes an imaging unit 10 for imaging the inside of the caution area, a thermo sensor 20 for detecting the temperature distribution in the caution area, and a control unit 30.
  • the control unit 30 determines a first determination unit 31 that determines the occurrence of a fire point due to combustion or flame in the caution area, a second determination unit 32 that determines the growth of the fire point, and a threshold of the temperature that is determined to be the fire point.
  • the time-series information may be shooting time, but may be any information that allows determination of the shooting order of the images.
  • the position of the highest temperature in the temperature distribution detected by the thermo sensor 20 is temporarily determined to be the fire point, and the temperature at the temporarily determined fire point is equal to or higher than the threshold stored in the threshold storage unit 33 It is judged that the occurrence of a fire point.
  • the temperatures detected by the thermo sensor 20 are divided in a matrix without averaging.
  • the first determination unit 31 compares all the temperatures detected in a matrix and is detected with a threshold, and when a temperature above the threshold is detected, the position of the detected temperature is identified to generate a fire point. You may judge.
  • the second determination unit 32 compares a plurality of images stored in the image storage unit 34 based on time-series information, and the shape, size, color, brightness, and fluctuation of the fire point included in the image are compared.
  • the fire point growth is determined from the time series change of the flame element including at least one of
  • the second determination unit 32 determines the time-series change of the flame included in the image in comparison with the time-series change pattern stored in the time-series change pattern storage unit 35.
  • the time-series change patterns stored in the time-series change pattern storage unit 35 are preferably those patterned in advance by machine learning.
  • the time-sequential change pattern of the flame element patterned beforehand by machine learning is time-sequential change information, and is data on a time-series growth curve or change curve, or an approximate expression.
  • FIG. 2 is a flowchart showing the operation of the fire identification device.
  • the temperature distribution in the alert area is constantly detected by the thermo sensor 20 (step 1).
  • the position of the highest temperature in the temperature distribution detected by step 1 is specified and provisionally determined as the flash point (step 2).
  • step 2 it is determined whether or not the temperature at the temporarily determined fire point is equal to or higher than the threshold stored in the threshold storage unit 33 for the temporarily determined fire point, and the temperature of the temporarily determined fire point is equal to or higher than the threshold It is determined that a fire point has occurred (step 3). If it is determined in step 3 that a fire point has occurred, the imaging unit 10 starts imaging (step 4). The image captured in step 4 is stored in the image storage unit 34 (step 5).
  • the plurality of images stored in step 5 are compared based on the time-series information in the second determination unit 32, and the growth of the fire point is determined from the time-series change of the flame element (step 6).
  • the determination in step 6 it is preferable to determine the time-series change of the flame included in the image in comparison with the time-series change pattern stored in the time-series change pattern storage unit 35. If it is determined in step 6 that the fire point has grown, it is determined that a fire has occurred (step 7), and an output for warning or extinguishing is performed. If it is determined in step 6 that the fire point has not grown, the imaging by the imaging unit 10 is ended (step 8), and the process returns to step 1 to perform temperature detection by the thermo sensor 20.
  • the generation of the fire point can be performed at an early stage. Further, according to the present embodiment, after the determination of the occurrence of the fire point, by determining the growth of the fire point from the time-series change of the fire element, the fire point having a small risk is excluded, leading to a fire. It can identify dangerous hot spots. Further, according to the present embodiment, it is possible to instantaneously perform more accurate discrimination by previously storing a time-series change pattern having a high risk of leading to a fire, which is patterned using machine learning. Can.
  • FIG. 3 is a block diagram showing the fire identification apparatus according to the present embodiment as function realizing means.
  • the fire identification apparatus according to the present embodiment includes an imaging unit 10 for imaging the inside of a caution area, and a control unit 30.
  • the control unit 30 determines that the first determination unit 31 that determines the occurrence of the fire point due to the combustion or the flame in the caution area, the second determination unit 32 that determines the growth of the fire point, and the flame that determines that the fire point has occurred.
  • the time-series information may be shooting time, but may be any information that allows determination of the shooting order of the images.
  • the first determination unit 31 compares the fire element included in the image stored in the image storage unit 34 with the fire point pattern stored in the fire point pattern storage unit 36 to determine the occurrence of the fire point. It is preferable that the fire point pattern stored in the fire point pattern storage unit 36 be previously patterned by machine learning.
  • the fire point pattern of the flame element patterned beforehand by machine learning is preferably time-series change information by at least two images, and as time-series change information, a time-series occurrence curve or change It is data about a curve or an approximate expression.
  • the second determination unit 32 compares a plurality of images stored in the image storage unit 34 based on time-series information, and the shape, size, color, brightness, and fluctuation of the fire point included in the image are compared.
  • the fire point growth is determined from the time series change of the flame element including at least one of
  • the second determination unit 32 determines the time-series change of the flame included in the image in comparison with the time-series change pattern stored in the time-series change pattern storage unit 35.
  • the time-series change patterns stored in the time-series change pattern storage unit 35 are preferably those patterned in advance by machine learning.
  • the time-sequential change pattern of the flame element patterned beforehand by machine learning is time-sequential change information, and is data on a time-series growth curve or change curve, or an approximate expression.
  • the first determination unit 31 stores the fire point pattern stored in the image storage unit 34, the fire element included in the first image and the fire element included in the second image captured after the first image
  • the second determination unit 32 determines at least three of the second images stored in the image storage unit 34 including the third image captured after the second image. It is further preferable to determine the time-series change of the flame included in the image in comparison with the time-series change pattern stored in the time-series change pattern storage unit 35.
  • FIG. 4 is a flowchart showing the operation of the fire identification device.
  • the inside of the caution area is constantly photographed by the photographing unit 10 (step 4).
  • the image captured in step 4 is stored in the image storage unit 34 (step 5).
  • the image stored in the step 5 is compared with the fire point pattern stored in the fire point pattern storage unit 36 to determine the occurrence of the fire point (see FIG. Step 3).
  • the flame element included in the first image and the flame element included in the second image captured after the first image are compared with the fire point pattern stored in the fire point pattern storage unit 36 It is preferable to make a decision.
  • step 3 If it is determined in step 3 that a fire point has occurred, the plurality of images stored in step 5 are compared based on time-series information, and the growth of the fire point is made from the time-series change of the fire element. Determine (step 6). If it is determined in step 3 that no fire point has occurred, the process returns to step 4 and the imaging unit 10 performs imaging. In the determination in step 6, it is preferable to determine the time-series change of the flame included in the image in comparison with the time-series change pattern stored in the time-series change pattern storage unit 35.
  • step 6 the time-series change pattern of the time-series change of the flame included in at least three images including the third image captured after the second image stored in the image storage unit 34 More preferably, the determination is made in comparison with the time-series change pattern stored in the storage unit 35. If it is determined in step 6 that the fire point has grown, it is determined that a fire has occurred (step 7), and an output for warning or extinguishing is performed. If it is determined in step 6 that the fire point has not grown, the process returns to step 4 and the imaging unit 10 performs imaging.
  • the present embodiment it is possible to instantaneously determine the occurrence of the firepoint with higher accuracy by storing the firepoint pattern patterned by using the machine learning in advance. Further, according to the present embodiment, after the determination of the occurrence of the fire point, by determining the growth of the fire point from the time-series change of the fire element, the fire point having a small risk is excluded, leading to a fire. It can identify dangerous hot spots. Further, according to the present embodiment, it is possible to instantaneously perform more accurate discrimination by previously storing a time-series change pattern having a high risk of leading to a fire, which is patterned using machine learning. Can.
  • FIG. 5 is a block diagram showing the fire identification apparatus according to the present embodiment as function realizing means.
  • the fire identification apparatus according to the present embodiment includes an imaging unit 10 for imaging the inside of a caution area, a control unit 30, and as the imaging unit 10, a first imaging unit 10a and a second imaging unit 10b.
  • the control unit 30 determines that the first determination unit 31 that determines the occurrence of the fire point due to the combustion or the flame in the caution area, the second determination unit 32 that determines the growth of the fire point, and the flame that determines that the fire point has occurred.
  • the time-series information may be shooting time, but may be any information that allows determination of the shooting order of the images.
  • the image storage unit 34 includes a first image storage unit 34a and a second image storage unit 34b, and the first image storage unit 34a stores an image captured by the first imaging unit 10a together with time-series information, The second image storage unit 34 b stores an image captured by the second imaging unit 10 b together with time-series information.
  • the flame element included in the image stored in the first image storage unit 34a is temporarily determined by comparison with the fire point pattern, and included in the image stored in the second image storage unit 34b. At least one of a provisional determination based on an image stored in the first image storage unit 34a and a provisional determination based on an image stored in the second image storage unit 34b by temporarily determining the flame element in comparison with the fire point pattern. If there is a temporary determination, it is determined that a fire point has occurred. In the first determination unit 31, if there is at least one temporary determination, it is possible to identify the occurrence of the fire point at an early stage by determining that the fire point is occurring.
  • the flame element included in the image stored in the first image storage unit 34a is temporarily determined by comparing it with the fire point pattern, and the image stored in the second image storage unit 34b is used. Both the temporary determination based on the image stored in the first image storage unit 34a and the temporary determination based on the image stored in the second image storage unit 34b by temporarily determining the included fire elements by comparing with the fire point pattern It is also possible to determine that a fire point has occurred by the provisional determination of. By determining that the fire point is generated by both the temporary determinations in the first determination unit 31, it is possible to identify the generation of the fire point having a high risk with high accuracy. It is preferable that the fire point pattern stored in the fire point pattern storage unit 36 be previously patterned by machine learning.
  • the fire point pattern of the flame element patterned beforehand by machine learning is preferably time-series change information by at least two images, and as time-series change information, a time-series occurrence curve or change It is data about a curve or an approximate expression.
  • the second determination unit 32 compares a plurality of images stored in the first image storage unit 34 a and the second image storage unit 34 b based on time-series information, and includes the shape and size of the fire point included in the images.
  • the growth of the fire point is determined from the time series change of the flame element including at least one of color, brightness, and fluctuation.
  • the second determination unit 32 temporarily determines the growth of the flash point from the time-series change of the flame element included in the plurality of images stored in the first image storage unit 34a, and is stored in the second image storage unit 34b.
  • the second determination unit 32 temporarily determines the growth of the fire point from the time-series change of the flame element included in the plurality of images stored in the first image storage unit 34a, and stores it in the second image storage unit 34b.
  • Temporary growth of the fire point is temporarily determined from the time-series change of the flame element included in the plurality of images, and the temporary determination based on the image stored in the first image storage unit 34a and the second image storage unit 34b It is also possible to determine that the fire point is growing based on both of the temporary determination with the image and the temporary determination. By determining that the fire point is growing based on both of the temporary determinations in the second determination unit 32, the growth of the highly dangerous fire point can be identified with high accuracy.
  • the second determination unit 32 determines the time-series change of the flame included in the image in comparison with the time-series change pattern stored in the time-series change pattern storage unit 35.
  • the time-series change patterns stored in the time-series change pattern storage unit 35 are preferably those patterned in advance by machine learning.
  • the time-sequential change pattern of the flame element patterned beforehand by machine learning is time-sequential change information, and is data on a time-series growth curve or change curve, or an approximate expression.
  • the flame element included in the first image and the second image captured after the first image, which are stored in the first image storage unit 34a and the second image storage unit 34b, are included.
  • the fire element to be determined is determined by comparison with the fire point pattern stored in the fire point pattern storage unit 36, and the second determination unit 32 stores it in the first image storage unit 34a and the second image storage unit 34b. And the time-series change of the flame included in the at least three images, which includes the third image captured after the second image, and the time-series change pattern stored in the time-series change pattern storage unit 35 More preferably,
  • FIG. 6 is a flowchart showing the operation of the fire identification device.
  • the inside of the caution area is constantly photographed by the first photographing unit 10a (step 4a), and the inside of the caution area is constantly photographed by the second photographing unit 10b (step 4b).
  • the image captured in step 4a is stored in the image storage unit 34a (step 5a), and the image captured in step 4b is stored in the image storage unit 34b (step 5b).
  • the fire element included in the image is compared with the fire point pattern stored in the fire point pattern storage unit 36, and the image stored in step 5b is the image stored in step 5a.
  • the flame element included in the image is compared with the fire point pattern stored in the fire point pattern storage unit 36 to determine the occurrence of the fire point (step 3).
  • the flame element included in the first image and the flame element included in the second image captured after the first image are compared with the fire point pattern stored in the fire point pattern storage unit 36 It is preferable to make a decision.
  • the flame element included in the image stored in the first image storage unit 34a is temporarily determined by comparison with the fire point pattern, and the flame element included in the image stored in the second image storage unit 34b is Temporary determination based on the image stored in the first image storage unit 34a, and temporary determination based on the image stored in the second image storage unit 34b.
  • step 3 the flame element included in the image stored in the first image storage unit 34a is temporarily determined by comparison with the fire point pattern, and the flame element included in the image stored in the second image storage unit 34b Is temporarily compared with the fire point pattern, and both the temporary determination by the image stored in the first image storage unit 34a and the temporary determination by the image stored in the second image storage unit 34b are performed. It can also be determined that a fire point has occurred.
  • step 3 If it is determined in step 3 that a fire point has occurred, the plurality of images stored in step 5 are compared based on time-series information, and the growth of the fire point is made from the time-series change of the fire element. Determine (step 6). If it is determined in step 3 that no fire point has occurred, the process returns to step 4 and the imaging unit 10 performs imaging. In the determination in step 6, it is preferable to determine the time-series change of the flame included in the image in comparison with the time-series change pattern stored in the time-series change pattern storage unit 35.
  • step 6 the time-series change pattern of the time-series change of the flame included in at least three images including the third image captured after the second image stored in the image storage unit 34 More preferably, the determination is made in comparison with the time-series change pattern stored in the storage unit 35.
  • a plurality of images stored in the second image storage unit 34b are tentatively determined on the basis of time-series changes in flame elements included in the plurality of images stored in the first image storage unit 34a. Temporarily determine the growth of the fire point from the time-series change of the flame element included in the first image storage unit 34a, and temporarily determine the image stored in the second image storage unit 34b.
  • the growth of the flash point is temporarily determined from the time-series change of the flame element included in the plurality of images stored in the first image storage unit 34a, and the plurality of images stored in the second image storage unit 34b. Temporarily determines the growth of the fire point from the time-series change of the flame element included in the image, and temporarily determines with the image stored in the first image storage unit 34a and with the image stored in the second image storage unit 34b. It is also possible to determine that the fire point is growing based on both of the preliminary determination and the determination.
  • step 6 If it is determined in step 6 that the fire point has grown, it is determined that a fire has occurred (step 7), and an output for warning or extinguishing is performed. If it is determined in step 6 that the flash point has not grown, the process returns to step 4a and step 4b, and the first imaging unit 10a and the second imaging unit 10b perform imaging.
  • the present embodiment it is possible to instantaneously determine the occurrence of the firepoint with higher accuracy by storing the firepoint pattern patterned by using the machine learning in advance. Further, according to the present embodiment, after the determination of the occurrence of the fire point, by determining the growth of the fire point from the time-series change of the fire element, the fire point having a small risk is excluded, leading to a fire. It can identify dangerous hot spots. Further, according to the present embodiment, it is possible to instantaneously perform more accurate discrimination by previously storing a time-series change pattern having a high risk of leading to a fire, which is patterned using machine learning. Can.
  • the generation of the fire point can be identified at an early stage. Further, according to the present embodiment, the first determination unit 31 can determine that the fire point is generated by both of the temporary determinations, so that the generation of the fire point having a high risk can be identified with high accuracy. Further, according to the present embodiment, if the second determination unit 32 determines that the fire point is growing if there is at least one temporary determination, the growth of the fire point can be identified at an early stage. Further, according to the present embodiment, it is possible to identify the growth of the high risk fire point with high accuracy, by the second determination unit 32 determining that the fire point is growing by both temporary determinations.
  • the first imaging unit 10a and the second imaging unit 10b as the imaging unit 10
  • the first imaging unit 10a and the second determination unit 32 can be used in the first determination unit 31 and the second determination unit 32.
  • a three-dimensional image generated based on an image captured by the second imaging unit 10b can be used.
  • the determination accuracy can be further enhanced by using a three-dimensional image.
  • the present invention is most suitable for a fire identification device in a fire extinguishing equipment using a fire extinguishant filled in a fire extinguishant storage container, but can also be applied to a fire identification device in a fire extinguishing equipment using fire extinguishing water such as a sprinkler.
  • thermo sensor 10 10 imaging unit 10a first imaging unit 10b second imaging unit 20 thermo sensor 30 control unit 31 first determination unit 32 second determination unit 33 threshold storage unit 34 image storage unit 34a first image storage unit 34b second image storage unit 35 Temporal change pattern storage unit 36 Fire point pattern storage unit

Landscapes

  • Business, Economics & Management (AREA)
  • Emergency Management (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Fire-Detection Mechanisms (AREA)

Abstract

The present invention provides a fire identification device which is capable of identifying fire points having a high risk of being led to a fire, excluding fire points having a low risk. The fire identification device of the present invention, which is provided with a first determination unit 31 for determining the occurrence of a fire point due to combustion or flame in a warning zone, and a second determination unit 32 for determining the growth of the fire point when the first determination unit 31 determines that the fire point has occurred, has: an image-capturing unit 10 which captures an image of the inside of the warning zone; and an image storage unit 34 which stores images captured by the image-capturing unit 10 together with time-series information, wherein the second determination unit 32 compares a plurality of images stored in the image storage unit 34 on the basis of the time-series information, and determines the growth of the fire point on the basis of time-series changes in a flame element including at least one among the shape, size, color, luminance, and flickering of the fire point included in the images.

Description

火災識別装置Fire identification device
　本発明は、消火設備に用いる火災識別装置に関する。 The present invention relates to a fire identification device used in a fire extinguishing facility.
　一般に炎の検出装置として、炎特有の波長光である４．５μmを中心としたＣＯ２共鳴放射帯を用いるものが知られている。
　また特許文献１では、時系列の画像を二値化処理し、連続する前後の画像を比較することで炎候補領域を特定し、その後、楕円一致率や平均輝度の変化量等の五つの特徴量から炎判別する炎検出装置が提案されている。 In general, as a flame detection apparatus, one using a CO 2 resonance radiation band centered on 4.5 μm which is wavelength light unique to flames is known.
Further, in Patent Document 1, a flame candidate region is specified by performing binarization processing of time-series images and comparing successive images before and after, five characteristics such as an elliptical coincidence rate and a change amount of average luminance A flame detection device has been proposed that discriminates flames from quantity.
特許第６１４０５９９号公報Patent No. 6140599
　しかし、ＣＯ２共鳴放射帯を撮影する赤外線撮像素子は、冷却構造に複雑な構成を要し、且つ大型で高価であるという課題があった。
　また特許文献１では、上記の課題を解決するが、炎候補領域の特定ならびに炎判定の工程が煩雑である。
　またいずれも、そのまま放っておけば自消し火災には繋がらないような、危険性の小さな火炎も抽出する可能性がある。このような危険性の小さな火炎も検知してしまうことは、建物使用者のみならず、当該建物の存する地域を管轄する消防職員にとっても、現場駆けつけや火災原因調査ならびに報告などの業務の負担を増す可能性がある。 However, the infrared imaging device for imaging the CO 2 resonance radiation band has a problem that the cooling structure requires a complicated configuration and is large and expensive.
Moreover, although the above-mentioned subject is solved in patent documents 1, the process of specification of a flame candidate field, and a flame judging is complicated.
Also, in any case, it is possible to extract flames of small danger that would not lead to a self-extinguishing fire if left alone. Detecting flames with such a low risk also places a burden on work such as rushing on the spot and investigation and reporting of fires not only for building users but also for fire brigade personnel who have jurisdiction over the area where the building is located. There is a possibility to increase.
　そこで本発明は、危険性の小さな火点を除外し、火災に繋がるような危険性の高い火点を識別できる火災識別装置を提供することを目的とする。 Therefore, an object of the present invention is to provide a fire identification device capable of excluding a fire point having a low risk and identifying a high fire point which may lead to a fire.
　請求項１記載の本発明の火災識別装置は、警戒区域内における燃焼又は炎による火点の発生を判定する第１判定部と、前記第１判定部で前記火点の発生を判定すると、前記火点の成長を判定する第２判定部とを備え、前記警戒区域内を撮影する撮影部と、前記撮影部で撮影される画像を時系列情報とともに記憶する画像記憶部とを有する火災識別装置であって、前記第２判定部では、前記画像記憶部に記憶された複数の前記画像を、前記時系列情報を元に比較し、前記画像に含まれる、前記火点の形状、大きさ、色、輝度、及び揺らぎの内の少なくとも一つを含む炎要素の時系列的変化から前記火点の成長を判定することを特徴とする。
　請求項２記載の本発明は、請求項１に記載の火災識別装置において、前記火点が成長していると判定する、前記炎要素の時系列的変化パターンを記憶する時系列的変化パターン記憶部を有し、前記第２判定部では、前記画像に含まれる前記炎の前記時系列的変化を、前記時系列的変化パターン記憶部に記憶している前記時系列的変化パターンと比較して判定することを特徴とする。
　請求項３記載の本発明は、請求項１又は請求項２に記載の火災識別装置において、前記警戒区域内の温度分布を検出するサーモセンサを有し、前記第１判定部では、前記温度分布の中で最高温度の位置を前記火点と仮決定し、仮決定した前記火点での温度が閾値以上の場合に前記火点の前記発生と判定し、前記第１判定部で前記火点の前記発生を判定すると、前記撮影部で撮影を開始することを特徴とする。
　請求項４記載の本発明は、請求項１又は請求項２に記載の火災識別装置において、前記火点が発生したと判定する前記炎要素の火点パターンを記憶する火点パターン記憶部を有し、前記第１判定部では、前記画像記憶部に記憶された前記画像に含まれる前記炎要素を、前記火点パターン記憶部に記憶している前記火点パターンと比較して判定することを特徴とする。
　請求項５記載の本発明は、請求項２に記載の火災識別装置において、前記火点が発生したと判定する前記炎要素の火点パターンを記憶する火点パターン記憶部を有し、前記第１判定部では、前記画像記憶部に記憶された、第１画像に含まれる前記炎要素、及び前記第１画像の後に撮影された第２画像に含まれる前記炎要素を、前記火点パターン記憶部に記憶している前記火点パターンと比較して判定し、前記第２判定部では、前記画像記憶部に記憶された、前記第２画像の後に撮影された第３画像を含む、少なくとも３つの前記画像に含まれる前記炎の前記時系列的変化を、前記時系列的変化パターン記憶部に記憶している前記時系列的変化パターンと比較して判定することを特徴とする。
　請求項６記載の本発明は、請求項１から請求項５のいずれか１項に記載の火災識別装置において、前記撮影部として、第１撮影部と第２撮影部とを有し、前記画像記憶部として、第１画像記憶部と第２画像記憶部とを有し、前記第１画像記憶部では、前記第１撮影部で撮影される前記画像を前記時系列情報とともに記憶し、前記第２画像記憶部では、前記第２撮影部で撮影される前記画像を前記時系列情報とともに記憶し、前記第２判定部では、前記第１画像記憶部に記憶された複数の前記画像に含まれる前記炎要素の時系列的変化から前記火点の成長を仮判定し、前記第２画像記憶部に記憶された複数の前記画像に含まれる前記炎要素の時系列的変化から前記火点の成長を仮判定し、前記第１画像記憶部に記憶された前記画像による仮判定した結果、及び前記第２画像記憶部に記憶された前記画像による仮判定した結果、少なくともいずれかの仮判定があれば前記火点が成長していると判定することを特徴とする。
　請求項７記載の本発明は、請求項１から請求項５のいずれか１項に記載の火災識別装置において、前記撮影部として、第１撮影部と第２撮影部とを有し、前記画像記憶部として、第１画像記憶部と第２画像記憶部とを有し、前記第１画像記憶部では、前記第１撮影部で撮影される前記画像を前記時系列情報とともに記憶し、前記第２画像記憶部では、前記第２撮影部で撮影される前記画像を前記時系列情報とともに記憶し、前記第２判定部では、前記第１画像記憶部に記憶された複数の前記画像に含まれる前記炎要素の時系列的変化から前記火点の成長を仮判定し、前記第２画像記憶部に記憶された複数の前記画像に含まれる前記炎要素の時系列的変化から前記火点の成長を仮判定し、前記第１画像記憶部に記憶された前記画像による仮判定した結果、前記第２画像記憶部に記憶された前記画像による仮判定した結果、双方の仮判定によって前記火点が成長していると判定することを特徴とする。
　請求項８記載の本発明は、請求項４に記載の火災識別装置において、前記撮影部として、第１撮影部と第２撮影部とを有し、前記画像記憶部として、第１画像記憶部と第２画像記憶部とを有し、前記第１画像記憶部では、前記第１撮影部で撮影される前記画像を前記時系列情報とともに記憶し、前記第２画像記憶部では、前記第２撮影部で撮影される前記画像を前記時系列情報とともに記憶し、前記第１判定部では、前記第１画像記憶部に記憶された前記画像に含まれる前記炎要素を、前記火点パターンと比較して仮判定し、前記第２画像記憶部に記憶された前記画像に含まれる前記炎要素を、前記火点パターンと比較して仮判定し、前記第１画像記憶部に記憶された前記画像による仮判定した結果、及び前記第２画像記憶部に記憶された前記画像による仮判定した結果、少なくともいずれかの仮判定があれば前記火点が発生していると判定することを特徴とする。
　請求項９記載の本発明は、請求項４に記載の火災識別装置において、前記撮影部として、第１撮影部と第２撮影部とを有し、前記画像記憶部として、第１画像記憶部と第２画像記憶部とを有し、前記第１画像記憶部では、前記第１撮影部で撮影される前記画像を前記時系列情報とともに記憶し、前記第２画像記憶部では、前記第２撮影部で撮影される前記画像を前記時系列情報とともに記憶し、前記第１判定部では、前記第１画像記憶部に記憶された前記画像に含まれる前記炎要素を、前記火点パターンと比較して仮判定し、前記第２画像記憶部に記憶された前記画像に含まれる前記炎要素を、前記火点パターンと比較して仮判定し、前記第１画像記憶部に記憶された前記画像による仮判定した結果、前記第２画像記憶部に記憶された前記画像による仮判定した結果、双方の仮判定によって前記火点が発生していると判定することを特徴とする。
　請求項１０記載の本発明は、請求項１又は請求項２に記載の火災識別装置において、前記撮影部として、第１撮影部と第２撮影部とを有し、前記画像記憶部には、前記第１撮影部及び前記第２撮影部で撮影される画像を前記時系列情報とともに記憶し、前記第２判定部では、前記第１撮影部及び前記第２撮影部で撮影される前記画像を元に生成した三次元画像を用いることを特徴とする。
　請求項１１記載の本発明は、請求項４に記載の火災識別装置において、前記撮影部として、第１撮影部と第２撮影部とを有し、前記画像記憶部には、前記第１撮影部及び前記第２撮影部で撮影される画像を前記時系列情報とともに記憶し、前記第１判定部では、前記第１撮影部及び前記第２撮影部で撮影される前記画像を元に生成した三次元画像を用いることを特徴とする。 The fire identification apparatus of the present invention according to claim 1 is characterized in that a first determination unit that determines the occurrence of a fire point due to combustion or flame in a caution area, and that the first determination unit determines the occurrence of the fire point. A fire identification apparatus comprising: a second determination unit that determines growth of a fire point; an imaging unit configured to image the inside of the caution area; and an image storage unit configured to store an image captured by the imaging unit together with time series information The second determination unit compares the plurality of images stored in the image storage unit based on the time-series information, and includes the shape and size of the fire point included in the image; It is characterized in that the growth of the fire point is determined from a time-series change of a flame element including at least one of color, brightness, and fluctuation.
According to a second aspect of the present invention, in the fire identification apparatus according to the first aspect, the time-series change pattern storage for storing the time-series change pattern of the flame element, wherein it is determined that the fire point is growing. The second determination unit compares the time-series change of the flame included in the image with the time-series change pattern stored in the time-series change pattern storage unit. It is characterized by judging.
According to a third aspect of the present invention, in the fire identification apparatus according to the first or second aspect, the thermal sensor for detecting a temperature distribution in the warning area is provided, and the first determination unit determines the temperature distribution. The position of the highest temperature among them is temporarily determined as the fire point, and when the temperature at the fire point temporarily determined is equal to or higher than a threshold, it is determined that the fire point is generated, and the first determination unit When the occurrence of the above is determined, the photographing unit starts photographing.
According to a fourth aspect of the present invention, in the fire identification apparatus according to the first or second aspect, there is provided a fire point pattern storage unit for storing the fire point pattern of the fire element that determines that the fire point has occurred. The first determination unit determines that the flame element included in the image stored in the image storage unit is compared with the fire point pattern stored in the fire point pattern storage unit. It features.
According to a fifth aspect of the present invention, in the fire identification apparatus according to the second aspect, the fire point pattern storage unit storing the fire point pattern of the flame element that determines that the fire point has occurred, In the determination unit, the fire point pattern stored in the image storage unit, the fire element included in the first image, and the fire element included in the second image captured after the first image are stored in the fire point pattern storage The second determination unit determines at least the third image captured after the second image and stored in the image storage unit. It is characterized in that the time-series change of the flame included in one image is determined by comparing with the time-series change pattern stored in the time-series change pattern storage unit.
According to a sixth aspect of the present invention, in the fire identification apparatus according to any one of the first to fifth aspects, the imaging unit includes a first imaging unit and a second imaging unit, and the image The storage unit includes a first image storage unit and a second image storage unit, and the first image storage unit stores the image captured by the first imaging unit together with the time-series information, and The two-image storage unit stores the images captured by the second imaging unit together with the time-series information, and the second determination unit is included in the plurality of images stored in the first image storage unit. The growth of the fire point is temporarily determined from the time-series change of the flame element, and the growth of the fire point is generated from the time-sequential change of the flame element included in the plurality of images stored in the second image storage unit Is temporarily determined, and the temporary determination based on the image stored in the first image storage unit As a result, and the second image storage unit provisional decision result according to said stored image, and judging said fire point if there is tentative determination of at least one is growing.
According to a seventh aspect of the present invention, in the fire identification apparatus according to any one of the first to fifth aspects, the photographing unit includes a first photographing unit and a second photographing unit. The storage unit includes a first image storage unit and a second image storage unit, and the first image storage unit stores the image captured by the first imaging unit together with the time-series information, and The two-image storage unit stores the images captured by the second imaging unit together with the time-series information, and the second determination unit is included in the plurality of images stored in the first image storage unit. The growth of the fire point is temporarily determined from the time-series change of the flame element, and the growth of the fire point is generated from the time-sequential change of the flame element included in the plurality of images stored in the second image storage unit Is temporarily determined, and the temporary determination based on the image stored in the first image storage unit As a result, the provisional judgment result by the image stored in the second image storing unit, and judging with the fire point is grown by the provisional determination of both.
According to an eighth aspect of the present invention, in the fire identification apparatus according to the fourth aspect, the imaging section includes a first imaging section and a second imaging section, and the image storage section is a first image storage section. And a second image storage unit, wherein the first image storage unit stores the image captured by the first imaging unit together with the time-series information, and the second image storage unit includes the second image storage unit. The image captured by the imaging unit is stored together with the time-series information, and the first determination unit compares the flame element included in the image stored in the first image storage unit with the fire point pattern And temporarily determine the flame element contained in the image stored in the second image storage unit by comparing it with the fire point pattern and temporarily determining the image stored in the first image storage unit As a result of temporary determination by the second image storage unit, and the previous stored in the second image storage unit Temporary determination result by the image, and judging said fire point if there is tentative determination of at least one has occurred.
According to a ninth aspect of the present invention, in the fire identification apparatus according to the fourth aspect, the imaging section includes a first imaging section and a second imaging section, and the image storage section is a first image storage section. And a second image storage unit, wherein the first image storage unit stores the image captured by the first imaging unit together with the time-series information, and the second image storage unit includes the second image storage unit. The image captured by the imaging unit is stored together with the time-series information, and the first determination unit compares the flame element included in the image stored in the first image storage unit with the fire point pattern And temporarily determine the flame element contained in the image stored in the second image storage unit by comparing it with the fire point pattern and temporarily determining the image stored in the first image storage unit As a result of the temporary determination by the second image stored in the second image storage unit. Provisional judgment result by, and judging with the fire point by the provisional determination of both has occurred.
According to a tenth aspect of the present invention, in the fire identification apparatus according to the first or second aspect, the photographing unit includes a first photographing unit and a second photographing unit, and the image storage unit includes: The images captured by the first imaging unit and the second imaging unit are stored together with the time-series information, and the second determination unit is configured to store the images captured by the first imaging unit and the second imaging unit. It is characterized by using an originally generated three-dimensional image.
According to an eleventh aspect of the present invention, in the fire identification apparatus according to the fourth aspect, the photographing unit includes a first photographing unit and a second photographing unit, and the image storage unit includes the first photographing. An image captured by the unit and the second imaging unit is stored together with the time-series information, and the first determination unit is generated based on the image captured by the first imaging unit and the second imaging unit It is characterized by using a three-dimensional image.
　本発明の火災識別装置によれば、火点の発生の判定の後に、火点の成長を炎要素の時系列的変化から判定することで、危険性の小さな火点を除外し、火災に繋がるような危険性の高い火点を識別できる。 According to the fire identification device of the present invention, after the determination of the occurrence of the fire point, the fire point growth is determined from the time-series change of the fire element, thereby excluding the fire point having a small risk and leading to the fire. It is possible to identify such dangerous fire points.
本発明の第１実施例による火災識別装置を機能実現手段で現したブロック図Block diagram showing the fire identification device according to the first embodiment of the present invention as a function realizing means 同火災識別装置の動作を示すフローチャートFlow chart showing the operation of the fire identification device 本発明の第２実施例による火災識別装置を機能実現手段で現したブロック図The block diagram which represented the fire identification apparatus by 2nd Example of this invention by the function implementation means 同火災識別装置の動作を示すフローチャートFlow chart showing the operation of the fire identification device 本発明の第３実施例による火災識別装置を機能実現手段で現したブロック図The block diagram which represented the fire identification apparatus by 3rd Example of this invention by the function implementation means 同火災識別装置の動作を示すフローチャートFlow chart showing the operation of the fire identification device
　本発明の第１の実施の形態による火災識別装置は、第２判定部では、画像記憶部に記憶された複数の画像を、時系列情報を元に比較し、画像に含まれる、火点の形状、大きさ、色、輝度、及び揺らぎの内の少なくとも一つを含む炎要素の時系列的変化から火点の成長を判定するものである。
　本実施の形態によれば、火点の発生の判定の後に、火点の成長を炎要素の時系列的変化から判定することで、危険性の小さな火点を除外し、火災に繋がるような危険性の高い火点を識別できる。 In the fire identification apparatus according to the first embodiment of the present invention, in the second determination unit, a plurality of images stored in the image storage unit are compared based on time-series information, and the fire point is included in the images. The growth of the fire point is determined from the time series change of the flame element including at least one of the shape, the size, the color, the brightness, and the fluctuation.
According to the present embodiment, after the determination of the occurrence of the fire point, by determining the growth of the fire point from the time-series change of the fire element, the fire point having a small risk is excluded, leading to a fire. It can identify dangerous hot spots.
　本発明の第２の実施の形態は、第１の実施の形態による火災識別装置において、火点が成長していると判定する、炎要素の時系列的変化パターンを記憶する時系列的変化パターン記憶部を有し、第２判定部では、画像に含まれる炎の時系列的変化を、時系列的変化パターン記憶部に記憶している時系列的変化パターンと比較して判定するものである。
　本実施の形態によれば、例えば機械学習を用いてパターン化した、火災に繋がるような危険性の高い時系列的変化パターンをあらかじめ記憶しておくことで、より精度の高い判別を瞬時に行うことができる。 A second embodiment of the present invention is a fire identification apparatus according to the first embodiment, wherein a time-series change pattern of storing a time-series change pattern of fire elements, which determines that the fire point is growing, is stored. A storage unit is provided, and the second determination unit determines the time-series change of the flame included in the image in comparison with the time-series change pattern stored in the time-series change pattern storage unit. .
According to the present embodiment, more accurate discrimination can be instantaneously performed by storing beforehand, for example, a time-series change pattern having a high risk of causing a fire, which is patterned using machine learning. be able to.
　本発明の第３の実施の形態は、第１又は第２の実施の形態による火災識別装置において、警戒区域内の温度分布を検出するサーモセンサを有し、第１判定部では、温度分布の中で最高温度の位置を火点と仮決定し、仮決定した火点での温度が閾値以上の場合に火点の発生と判定し、第１判定部で火点の発生を判定すると、撮影部で撮影を開始するものである。
　本実施の形態によれば、サーモセンサを用いることで、火点の発生を早期に行える。 A third embodiment of the present invention is a fire identification apparatus according to the first or second embodiment, further comprising a thermo sensor for detecting a temperature distribution in the warning area, and the first determination unit If the position of the highest temperature is temporarily determined to be the fire point, and the temperature at the temporarily determined fire point is equal to or higher than the threshold value, it is determined that the fire point occurs, and the first determination unit determines the occurrence of the fire point. The department will start shooting.
According to the present embodiment, the use of the thermo sensor makes it possible to generate the fire point at an early stage.
　本発明の第４の実施の形態は、第１又は第２の実施の形態による火災識別装置において、火点が発生したと判定する炎要素の火点パターンを記憶する火点パターン記憶部を有し、第１判定部では、画像記憶部に記憶された画像に含まれる炎要素を、火点パターン記憶部に記憶している火点パターンと比較して判定するものである。
　本実施の形態によれば、例えば機械学習を用いてパターン化した火点パターンをあらかじめ記憶しておくことで、より精度の高い火点の発生の判別を瞬時に行うことができる。 The fourth embodiment of the present invention has a fire point pattern storage unit for storing a fire point pattern of a fire element that determines that a fire point has occurred in the fire identification apparatus according to the first or second embodiment. The first determination unit determines the flame element included in the image stored in the image storage unit in comparison with the fire point pattern stored in the fire point pattern storage unit.
According to the present embodiment, it is possible to instantaneously determine the occurrence of the fire point with higher accuracy by storing the fire point pattern which is patterned using, for example, machine learning in advance.
　本発明の第５の実施の形態は、第２の実施の形態による火災識別装置であって、火点が発生したと判定する炎要素の火点パターンを記憶する火点パターン記憶部を有し、第１判定部では、画像記憶部に記憶された、第１画像に含まれる炎要素、及び第１画像の後に撮影された第２画像に含まれる炎要素を、火点パターン記憶部に記憶している火点パターンと比較して判定し、第２判定部では、画像記憶部に記憶された、第２画像の後に撮影された第３画像を含む、少なくとも３つの画像に含まれる炎の時系列的変化を、時系列的変化パターン記憶部に記憶している時系列的変化パターンと比較して判定するものである。
　本実施の形態によれば、火点の発生と火点の成長を分けて判定し、また火点の発生の判定より火点の成長の判定を多くの画像によって判定し、更には火点の発生と火点の成長との判定に、例えば機械学習を用いてパターン化した火点パターン及び時系列的変化パターンを用いることで、火災に繋がるような危険性の高い火点を高い精度で識別できる。 The fifth embodiment of the present invention is a fire identification apparatus according to the second embodiment, including a fire point pattern storage unit for storing a fire point pattern of a fire element that determines that a fire point has occurred. In the first determination unit, the fire point pattern storage unit stores the fire element included in the first image and the fire element included in the second image captured after the first image, stored in the image storage unit. Of the flame included in at least three images, including the third image captured after the second image, stored in the image storage unit, in the second determination unit. The time-series change is determined by comparison with the time-series change pattern stored in the time-series change pattern storage unit.
According to the present embodiment, the generation of the fire point and the growth of the fire point are determined separately, and the determination of the growth of the fire point is determined by more images than the determination of the generation of the fire point. For example, using fire point patterns and time-series change patterns patterned using machine learning to determine occurrence and growth of the fire point, a fire point with a high risk of leading to a fire is identified with high accuracy it can.
　本発明の第６の実施の形態は、第１から第５のいずれかの実施の形態による火災識別装置であって、撮影部として、第１撮影部と第２撮影部とを有し、画像記憶部として、第１画像記憶部と第２画像記憶部とを有し、第１画像記憶部では、第１撮影部で撮影される画像を時系列情報とともに記憶し、第２画像記憶部では、第２撮影部で撮影される画像を時系列情報とともに記憶し、第２判定部では、第１画像記憶部に記憶された複数の画像に含まれる炎要素の時系列的変化から火点の成長を仮判定し、第２画像記憶部に記憶された複数の画像に含まれる炎要素の時系列的変化から火点の成長を仮判定し、第１画像記憶部に記憶された画像による仮判定した結果、及び第２画像記憶部に記憶された画像による仮判定した結果、少なくともいずれかの仮判定があれば火点が成長していると判定するものである。
　本実施の形態によれば、火点の成長を早い段階で識別できる。 The sixth embodiment of the present invention is the fire identification apparatus according to any one of the first to fifth embodiments, comprising a first imaging unit and a second imaging unit as the imaging unit, and an image The storage unit includes a first image storage unit and a second image storage unit. The first image storage unit stores an image captured by the first imaging unit together with time-series information, and the second image storage unit And an image captured by the second imaging unit is stored together with the time-series information, and the second determination unit is configured to generate the fire point from the time-series change of the flame element included in the plurality of images stored in the first image storage unit. The temporary determination of the growth, the temporary determination of the growth of the fire point from the time-series change of the flame element included in the plurality of images stored in the second image storage unit, and the temporary storage of the image stored in the first image storage unit As a result of the determination, and as a result of the temporary determination by the image stored in the second image storage unit, at least one of them If there is a temporary decision of the one in which it is determined that the fire point is growing.
According to this embodiment, the growth of the fire point can be identified at an early stage.
　本発明の第７の実施の形態は、第１から第５のいずれかの実施の形態による火災識別装置であって、撮影部として、第１撮影部と第２撮影部とを有し、画像記憶部として、第１画像記憶部と第２画像記憶部とを有し、第１画像記憶部では、第１撮影部で撮影される画像を時系列情報とともに記憶し、第２画像記憶部では、第２撮影部で撮影される画像を時系列情報とともに記憶し、第２判定部では、第１画像記憶部に記憶された複数の画像に含まれる炎要素の時系列的変化から火点の成長を仮判定し、第２画像記憶部に記憶された複数の画像に含まれる炎要素の時系列的変化から火点の成長を仮判定し、第１画像記憶部に記憶された画像による仮判定した結果、第２画像記憶部に記憶された画像による仮判定した結果、双方の仮判定によって火点が成長していると判定するものである。
　本実施の形態によれば、危険性の高い火点の成長を高い精度で識別できる。 A seventh embodiment of the present invention is a fire identification apparatus according to any one of the first to fifth embodiments, which has a first imaging unit and a second imaging unit as an imaging unit, and an image The storage unit includes a first image storage unit and a second image storage unit. The first image storage unit stores an image captured by the first imaging unit together with time-series information, and the second image storage unit And an image captured by the second imaging unit is stored together with the time-series information, and the second determination unit is configured to generate the fire point from the time-series change of the flame element included in the plurality of images stored in the first image storage unit. The temporary determination of the growth, the temporary determination of the growth of the fire point from the time-series change of the flame element included in the plurality of images stored in the second image storage unit, and the temporary storage of the image stored in the first image storage unit As a result of the determination, as a result of the temporary determination by the image stored in the second image storage unit, as a result of both temporary determinations One in which it is determined that the point is growing.
According to the present embodiment, it is possible to identify with high accuracy the growth of the high-risk fire point.
　本発明の第８の実施の形態は、第４の実施の形態による火災識別装置であって、撮影部として、第１撮影部と第２撮影部とを有し、画像記憶部として、第１画像記憶部と第２画像記憶部とを有し、第１画像記憶部では、第１撮影部で撮影される画像を時系列情報とともに記憶し、第２画像記憶部では、第２撮影部で撮影される画像を時系列情報とともに記憶し、第１判定部では、第１画像記憶部に記憶された画像に含まれる炎要素を、火点パターンと比較して仮判定し、第２画像記憶部に記憶された画像に含まれる炎要素を、火点パターンと比較して仮判定し、第１画像記憶部に記憶された画像による仮判定した結果、及び第２画像記憶部に記憶された画像による仮判定した結果、少なくともいずれかの仮判定があれば火点が発生していると判定するものである。
　本実施の形態によれば、火点の発生を早い段階で識別できる。 The eighth embodiment of the present invention is a fire identification apparatus according to the fourth embodiment, comprising a first imaging unit and a second imaging unit as an imaging unit, and an image storage unit as a first image storage unit. An image storage unit and a second image storage unit are provided. The first image storage unit stores an image captured by the first imaging unit together with time-series information, and the second image storage unit includes the second imaging unit. The photographed image is stored together with time-series information, and the first judgment unit provisionally judges the flame element included in the image stored in the first image storage unit with the fire point pattern, and stores the second image. The flame element contained in the image stored in the unit is temporarily determined by comparison with the fire point pattern, and the result of the temporary determination using the image stored in the first image storage unit is stored in the second image storage unit As a result of the temporary judgment by the image, if there is at least one temporary judgment that the fire point is generated One in which a constant.
According to the present embodiment, it is possible to identify the occurrence of the fire point at an early stage.
　本発明の第９の実施の形態は、第４の実施の形態による火災識別装置であって、撮影部として、第１撮影部と第２撮影部とを有し、画像記憶部として、第１画像記憶部と第２画像記憶部とを有し、第１画像記憶部では、第１撮影部で撮影される画像を時系列情報とともに記憶し、第２画像記憶部では、第２撮影部で撮影される画像を時系列情報とともに記憶し、第１判定部では、第１画像記憶部に記憶された画像に含まれる炎要素を、火点パターンと比較して仮判定し、第２画像記憶部に記憶された画像に含まれる炎要素を、火点パターンと比較して仮判定し、第１画像記憶部に記憶された画像による仮判定した結果、第２画像記憶部に記憶された画像による仮判定した結果、双方の仮判定によって火点が発生していると判定するものである。
　本実施の形態によれば、危険性の高い火点の発生を高い精度で識別できる。 The ninth embodiment of the present invention is a fire identification apparatus according to the fourth embodiment, comprising a first imaging unit and a second imaging unit as an imaging unit, and an image storage unit as a first embodiment. An image storage unit and a second image storage unit are provided. The first image storage unit stores an image captured by the first imaging unit together with time-series information, and the second image storage unit includes the second imaging unit. The photographed image is stored together with time-series information, and the first judgment unit provisionally judges the flame element included in the image stored in the first image storage unit with the fire point pattern, and stores the second image. The flame element included in the image stored in the unit is temporarily determined by comparison with the fire point pattern, and the image temporarily stored in the first image storage unit is temporarily stored in the second image storage unit. As a result of the temporary judgment by this, it is judged that the fire point has occurred by both temporary judgments
According to the present embodiment, it is possible to identify the occurrence of a high risk fire point with high accuracy.
　本発明の第１０の実施の形態は、第１又は第２の実施の形態による火災識別装置において、撮影部として、第１撮影部と第２撮影部とを有し、画像記憶部には、第１撮影部及び第２撮影部で撮影される画像を時系列情報とともに記憶し、第２判定部では、第１撮影部及び第２撮影部で撮影される画像を元に生成した三次元画像を用いるものである。
　本実施の形態によれば、三次元画像を用いることで判定精度を更に高めることができる。 A tenth embodiment of the present invention is a fire identification apparatus according to the first or second embodiment, including a first imaging unit and a second imaging unit as imaging units, and the image storage unit includes: A three-dimensional image generated based on the images captured by the first imaging unit and the second imaging unit, storing the images captured by the first imaging unit and the second imaging unit together with the time-series information Is used.
According to the present embodiment, the determination accuracy can be further enhanced by using a three-dimensional image.
　本発明の第１１の実施の形態は、第４の実施の形態による火災識別装置であって、撮影部として、第１撮影部と第２撮影部とを有し、画像記憶部には、第１撮影部及び第２撮影部で撮影される画像を時系列情報とともに記憶し、第１判定部では、第１撮影部及び第２撮影部で撮影される画像を元に生成した三次元画像を用いるものである。
　本実施の形態によれば、三次元画像を用いることで判定精度を更に高めることができる。 An eleventh embodiment of the present invention is the fire identification apparatus according to the fourth embodiment, comprising a first imaging unit and a second imaging unit as the imaging unit, and the image storage unit comprises: (1) The images taken by the imaging unit and the second imaging unit are stored together with time-series information, and the first determination unit generates a three-dimensional image generated based on the images taken by the first imaging unit and the second imaging unit It is used.
According to the present embodiment, the determination accuracy can be further enhanced by using a three-dimensional image.
　以下本発明の第１実施例による火災識別装置について説明する。
　図１は本実施例による火災識別装置を機能実現手段で現したブロック図である。
　本実施例による火災識別装置は、警戒区域内を撮影する撮影部１０と、警戒区域内の温度分布を検出するサーモセンサ２０と、制御部３０とを備えている。
　制御部３０は、警戒区域内における燃焼又は炎による火点の発生を判定する第１判定部３１と、火点の成長を判定する第２判定部３２と、火点と判断する温度の閾値を記憶する閾値記憶部３３と、撮影部１０で撮影される画像を時系列情報とともに記憶する画像記憶部３４と、火点が成長していると判定する、炎要素の時系列的変化パターンを記憶する時系列的変化パターン記憶部３５とを備えている。時系列情報は撮影時刻でもよいが、画像の撮影順が判別できる情報であればよい。 Hereinafter, a fire identification apparatus according to a first embodiment of the present invention will be described.
FIG. 1 is a block diagram showing a fire identification apparatus according to the present embodiment as function realizing means.
The fire identification apparatus according to the present embodiment includes an imaging unit 10 for imaging the inside of the caution area, a thermo sensor 20 for detecting the temperature distribution in the caution area, and a control unit 30.
The control unit 30 determines a first determination unit 31 that determines the occurrence of a fire point due to combustion or flame in the caution area, a second determination unit 32 that determines the growth of the fire point, and a threshold of the temperature that is determined to be the fire point. The threshold storage unit 33 to be stored, the image storage unit 34 to store the image captured by the imaging unit 10 together with time-series information, and the time-series change pattern of the flame element which determines that the fire point is growing And a time-series change pattern storage unit 35. The time-series information may be shooting time, but may be any information that allows determination of the shooting order of the images.
　第１判定部３１では、サーモセンサ２０で検出する温度分布の中で最高温度の位置を火点と仮決定し、仮決定した火点での温度が閾値記憶部３３に記憶した閾値以上の場合に火点の発生と判定する。なお、サーモセンサ２０で検出される温度は、平均化することなくマトリックス状に区分されている。第１判定部３１では、マトリックス状に区分されて検出される全ての温度を閾値と比較し、閾値以上の温度を検出した場合に、検出した温度の位置を特定することで火点の発生と判定してもよい。
　第２判定部３２では、画像記憶部３４に記憶された複数の画像を、時系列情報を元に比較し、画像に含まれる、火点の形状、大きさ、色、輝度、及び揺らぎの内の少なくとも一つを含む炎要素の時系列的変化から火点の成長を判定する。
　なお、第２判定部３２では、画像に含まれる炎の時系列的変化を、時系列的変化パターン記憶部３５に記憶している時系列的変化パターンと比較して判定することが好ましい。時系列的変化パターン記憶部３５に記憶する時系列的変化パターンは、あらかじめ機械学習によってパターン化されたものが好ましい。あらかじめ機械学習によってパターン化された炎要素の時系列的変化パターンは、時系列的な変化情報であり、時系列的な成長曲線若しくは変化曲線、又は近似式に関するデータである。 In the first determination unit 31, the position of the highest temperature in the temperature distribution detected by the thermo sensor 20 is temporarily determined to be the fire point, and the temperature at the temporarily determined fire point is equal to or higher than the threshold stored in the threshold storage unit 33 It is judged that the occurrence of a fire point. The temperatures detected by the thermo sensor 20 are divided in a matrix without averaging. The first determination unit 31 compares all the temperatures detected in a matrix and is detected with a threshold, and when a temperature above the threshold is detected, the position of the detected temperature is identified to generate a fire point. You may judge.
The second determination unit 32 compares a plurality of images stored in the image storage unit 34 based on time-series information, and the shape, size, color, brightness, and fluctuation of the fire point included in the image are compared. The fire point growth is determined from the time series change of the flame element including at least one of
Preferably, the second determination unit 32 determines the time-series change of the flame included in the image in comparison with the time-series change pattern stored in the time-series change pattern storage unit 35. The time-series change patterns stored in the time-series change pattern storage unit 35 are preferably those patterned in advance by machine learning. The time-sequential change pattern of the flame element patterned beforehand by machine learning is time-sequential change information, and is data on a time-series growth curve or change curve, or an approximate expression.
　図２は同火災識別装置の動作を示すフローチャートである。
　サーモセンサ２０によって常時警戒区域内の温度分布を検出する（ステップ１）。
　ステップ１によって検出される温度分布の中で最高温度の位置を特定して火点と仮決定する（ステップ２）。
　ステップ２で、仮決定された火点について、仮決定した火点での温度が閾値記憶部３３に記憶した閾値以上か否かを判定し、仮決定された火点の温度が閾値以上であると火点が発生したと判定する（ステップ３）。
　ステップ３において、火点が発生していると判定されると、撮影部１０での撮影を開始する（ステップ４）。
　ステップ４で撮影される画像は画像記憶部３４に記憶される（ステップ５）。
　ステップ５で記憶された複数の画像は、第２判定部３２において、時系列情報を元に比較し、炎要素の時系列的変化から火点の成長を判定する（ステップ６）。
　ステップ６における判定では、画像に含まれる炎の時系列的変化を、時系列的変化パターン記憶部３５に記憶している時系列的変化パターンと比較して判定することが好ましい。
　ステップ６において、火点が成長していると判定すると、火災として判定し（ステップ７）、警報や消火のための出力を行う。
　ステップ６において、火点が成長していないと判定すると、撮影部１０での撮影を終了し（ステップ８）、ステップ１に戻ってサーモセンサ２０による温度検出を行う。 FIG. 2 is a flowchart showing the operation of the fire identification device.
The temperature distribution in the alert area is constantly detected by the thermo sensor 20 (step 1).
The position of the highest temperature in the temperature distribution detected by step 1 is specified and provisionally determined as the flash point (step 2).
In step 2, it is determined whether or not the temperature at the temporarily determined fire point is equal to or higher than the threshold stored in the threshold storage unit 33 for the temporarily determined fire point, and the temperature of the temporarily determined fire point is equal to or higher than the threshold It is determined that a fire point has occurred (step 3).
If it is determined in step 3 that a fire point has occurred, the imaging unit 10 starts imaging (step 4).
The image captured in step 4 is stored in the image storage unit 34 (step 5).
The plurality of images stored in step 5 are compared based on the time-series information in the second determination unit 32, and the growth of the fire point is determined from the time-series change of the flame element (step 6).
In the determination in step 6, it is preferable to determine the time-series change of the flame included in the image in comparison with the time-series change pattern stored in the time-series change pattern storage unit 35.
If it is determined in step 6 that the fire point has grown, it is determined that a fire has occurred (step 7), and an output for warning or extinguishing is performed.
If it is determined in step 6 that the fire point has not grown, the imaging by the imaging unit 10 is ended (step 8), and the process returns to step 1 to perform temperature detection by the thermo sensor 20.
　本実施例によれば、サーモセンサ２０を用いることで、火点の発生を早期に行える。
　また本実施例によれば、火点の発生の判定の後に、火点の成長を炎要素の時系列的変化から判定することで、危険性の小さな火点を除外し、火災に繋がるような危険性の高い火点を識別できる。
　また本実施例によれば、機械学習を用いてパターン化した、火災に繋がるような危険性の高い時系列的変化パターンをあらかじめ記憶しておくことで、より精度の高い判別を瞬時に行うことができる。 According to the present embodiment, by using the thermo sensor 20, the generation of the fire point can be performed at an early stage.
Further, according to the present embodiment, after the determination of the occurrence of the fire point, by determining the growth of the fire point from the time-series change of the fire element, the fire point having a small risk is excluded, leading to a fire. It can identify dangerous hot spots.
Further, according to the present embodiment, it is possible to instantaneously perform more accurate discrimination by previously storing a time-series change pattern having a high risk of leading to a fire, which is patterned using machine learning. Can.
　以下本発明の第２実施例による火災識別装置について説明する。
　図３は本実施例による火災識別装置を機能実現手段で現したブロック図である。
　本実施例による火災識別装置は、警戒区域内を撮影する撮影部１０と、制御部３０とを備えている。
　制御部３０は、警戒区域内における燃焼又は炎による火点の発生を判定する第１判定部３１と、火点の成長を判定する第２判定部３２と、火点が発生したと判定する炎要素の火点パターンを記憶する火点パターン記憶部３６と、撮影部１０で撮影される画像を時系列情報とともに記憶する画像記憶部３４と、火点が成長していると判定する、炎要素の時系列的変化パターンを記憶する時系列的変化パターン記憶部３５とを備えている。時系列情報は撮影時刻でもよいが、画像の撮影順が判別できる情報であればよい。 Hereinafter, a fire identification apparatus according to a second embodiment of the present invention will be described.
FIG. 3 is a block diagram showing the fire identification apparatus according to the present embodiment as function realizing means.
The fire identification apparatus according to the present embodiment includes an imaging unit 10 for imaging the inside of a caution area, and a control unit 30.
The control unit 30 determines that the first determination unit 31 that determines the occurrence of the fire point due to the combustion or the flame in the caution area, the second determination unit 32 that determines the growth of the fire point, and the flame that determines that the fire point has occurred. A fire point pattern storage unit 36 for storing a fire point pattern of an element, an image storage unit 34 for storing an image captured by the imaging unit 10 together with time-series information, and a fire element which determines that a fire point is growing And a time-series change pattern storage unit 35 for storing time-series change patterns. The time-series information may be shooting time, but may be any information that allows determination of the shooting order of the images.
　第１判定部３１では、画像記憶部３４に記憶された画像に含まれる炎要素を、火点パターン記憶部３６に記憶している火点パターンと比較して火点の発生を判定する。火点パターン記憶部３６に記憶する火点パターンは、あらかじめ機械学習によってパターン化されたものが好ましい。あらかじめ機械学習によってパターン化された炎要素の火点パターンは、少なくとも２つの画像による時系列的な変化情報であることが好ましく、時系列的な変化情報としては、時系列的な発生曲線若しくは変化曲線、又は近似式に関するデータである。
　第２判定部３２では、画像記憶部３４に記憶された複数の画像を、時系列情報を元に比較し、画像に含まれる、火点の形状、大きさ、色、輝度、及び揺らぎの内の少なくとも一つを含む炎要素の時系列的変化から火点の成長を判定する。
　なお、第２判定部３２では、画像に含まれる炎の時系列的変化を、時系列的変化パターン記憶部３５に記憶している時系列的変化パターンと比較して判定することが好ましい。時系列的変化パターン記憶部３５に記憶する時系列的変化パターンは、あらかじめ機械学習によってパターン化されたものが好ましい。あらかじめ機械学習によってパターン化された炎要素の時系列的変化パターンは、時系列的な変化情報であり、時系列的な成長曲線若しくは変化曲線、又は近似式に関するデータである。
　また、第１判定部３１では、画像記憶部３４に記憶された、第１画像に含まれる炎要素、及び第１画像の後に撮影された第２画像に含まれる炎要素を、火点パターン記憶部３６に記憶している火点パターンと比較して判定し、第２判定部３２では、画像記憶部３４に記憶された、第２画像の後に撮影された第３画像を含む、少なくとも３つの画像に含まれる炎の時系列的変化を、時系列的変化パターン記憶部３５に記憶している時系列的変化パターンと比較して判定することが更に好ましい。 The first determination unit 31 compares the fire element included in the image stored in the image storage unit 34 with the fire point pattern stored in the fire point pattern storage unit 36 to determine the occurrence of the fire point. It is preferable that the fire point pattern stored in the fire point pattern storage unit 36 be previously patterned by machine learning. The fire point pattern of the flame element patterned beforehand by machine learning is preferably time-series change information by at least two images, and as time-series change information, a time-series occurrence curve or change It is data about a curve or an approximate expression.
The second determination unit 32 compares a plurality of images stored in the image storage unit 34 based on time-series information, and the shape, size, color, brightness, and fluctuation of the fire point included in the image are compared. The fire point growth is determined from the time series change of the flame element including at least one of
Preferably, the second determination unit 32 determines the time-series change of the flame included in the image in comparison with the time-series change pattern stored in the time-series change pattern storage unit 35. The time-series change patterns stored in the time-series change pattern storage unit 35 are preferably those patterned in advance by machine learning. The time-sequential change pattern of the flame element patterned beforehand by machine learning is time-sequential change information, and is data on a time-series growth curve or change curve, or an approximate expression.
Further, the first determination unit 31 stores the fire point pattern stored in the image storage unit 34, the fire element included in the first image and the fire element included in the second image captured after the first image The second determination unit 32 determines at least three of the second images stored in the image storage unit 34 including the third image captured after the second image. It is further preferable to determine the time-series change of the flame included in the image in comparison with the time-series change pattern stored in the time-series change pattern storage unit 35.
　図４は同火災識別装置の動作を示すフローチャートである。
　撮影部１０によって常時警戒区域内を撮影する（ステップ４）。
　ステップ４で撮影される画像は画像記憶部３４に記憶される（ステップ５）。
　ステップ５で記憶された画像は、第１判定部３１において、画像に含まれる炎要素が、火点パターン記憶部３６に記憶している火点パターンと比較され、火点の発生を判定する（ステップ３）。
　ステップ３における判定では、第１画像に含まれる炎要素、及び第１画像の後に撮影された第２画像に含まれる炎要素を、火点パターン記憶部３６に記憶している火点パターンと比較して判定することが好ましい。
　ステップ３において、火点が発生していると判定されると、ステップ５で記憶された複数の画像は、時系列情報を元に比較し、炎要素の時系列的変化から火点の成長を判定する（ステップ６）。
　ステップ３において、火点が発生していないと判定されると、ステップ４に戻って撮影部１０による撮影を行う。
　ステップ６における判定では、画像に含まれる炎の時系列的変化を、時系列的変化パターン記憶部３５に記憶している時系列的変化パターンと比較して判定することが好ましい。
　またステップ６における判定では、画像記憶部３４に記憶された、第２画像の後に撮影された第３画像を含む、少なくとも３つの画像に含まれる炎の時系列的変化を、時系列的変化パターン記憶部３５に記憶している時系列的変化パターンと比較して判定することが更に好ましい。
　ステップ６において、火点が成長していると判定すると、火災として判定し（ステップ７）、警報や消火のための出力を行う。
　ステップ６において、火点が成長していないと判定すると、ステップ４に戻って撮影部１０による撮影を行う。 FIG. 4 is a flowchart showing the operation of the fire identification device.
The inside of the caution area is constantly photographed by the photographing unit 10 (step 4).
The image captured in step 4 is stored in the image storage unit 34 (step 5).
In the first determination unit 31, the image stored in the step 5 is compared with the fire point pattern stored in the fire point pattern storage unit 36 to determine the occurrence of the fire point (see FIG. Step 3).
In the determination in step 3, the flame element included in the first image and the flame element included in the second image captured after the first image are compared with the fire point pattern stored in the fire point pattern storage unit 36 It is preferable to make a decision.
If it is determined in step 3 that a fire point has occurred, the plurality of images stored in step 5 are compared based on time-series information, and the growth of the fire point is made from the time-series change of the fire element. Determine (step 6).
If it is determined in step 3 that no fire point has occurred, the process returns to step 4 and the imaging unit 10 performs imaging.
In the determination in step 6, it is preferable to determine the time-series change of the flame included in the image in comparison with the time-series change pattern stored in the time-series change pattern storage unit 35.
Further, in the determination in step 6, the time-series change pattern of the time-series change of the flame included in at least three images including the third image captured after the second image stored in the image storage unit 34 More preferably, the determination is made in comparison with the time-series change pattern stored in the storage unit 35.
If it is determined in step 6 that the fire point has grown, it is determined that a fire has occurred (step 7), and an output for warning or extinguishing is performed.
If it is determined in step 6 that the fire point has not grown, the process returns to step 4 and the imaging unit 10 performs imaging.
　本実施例によれば、機械学習を用いてパターン化した火点パターンをあらかじめ記憶しておくことで、より精度の高い火点の発生の判別を瞬時に行うことができる。
　また本実施例によれば、火点の発生の判定の後に、火点の成長を炎要素の時系列的変化から判定することで、危険性の小さな火点を除外し、火災に繋がるような危険性の高い火点を識別できる。
　また本実施例によれば、機械学習を用いてパターン化した、火災に繋がるような危険性の高い時系列的変化パターンをあらかじめ記憶しておくことで、より精度の高い判別を瞬時に行うことができる。 According to the present embodiment, it is possible to instantaneously determine the occurrence of the firepoint with higher accuracy by storing the firepoint pattern patterned by using the machine learning in advance.
Further, according to the present embodiment, after the determination of the occurrence of the fire point, by determining the growth of the fire point from the time-series change of the fire element, the fire point having a small risk is excluded, leading to a fire. It can identify dangerous hot spots.
Further, according to the present embodiment, it is possible to instantaneously perform more accurate discrimination by previously storing a time-series change pattern having a high risk of leading to a fire, which is patterned using machine learning. Can.
　以下本発明の第３実施例による火災識別装置について説明する。
　図５は本実施例による火災識別装置を機能実現手段で現したブロック図である。
　本実施例による火災識別装置は、警戒区域内を撮影する撮影部１０と、制御部３０とを備え、撮影部１０として、第１撮影部１０ａと第２撮影部１０ｂとを有している。
　制御部３０は、警戒区域内における燃焼又は炎による火点の発生を判定する第１判定部３１と、火点の成長を判定する第２判定部３２と、火点が発生したと判定する炎要素の火点パターンを記憶する火点パターン記憶部３６と、撮影部１０で撮影される画像を時系列情報とともに記憶する画像記憶部３４と、火点が成長していると判定する、炎要素の時系列的変化パターンを記憶する時系列的変化パターン記憶部３５とを備えている。時系列情報は撮影時刻でもよいが、画像の撮影順が判別できる情報であればよい。
　画像記憶部３４として、第１画像記憶部３４ａと第２画像記憶部３４ｂとを有し、第１画像記憶部３４ａでは、第１撮影部１０ａで撮影される画像を時系列情報とともに記憶し、第２画像記憶部３４ｂでは、第２撮影部１０ｂで撮影される画像を時系列情報とともに記憶する。 Hereinafter, a fire identification system according to a third embodiment of the present invention will be described.
FIG. 5 is a block diagram showing the fire identification apparatus according to the present embodiment as function realizing means.
The fire identification apparatus according to the present embodiment includes an imaging unit 10 for imaging the inside of a caution area, a control unit 30, and as the imaging unit 10, a first imaging unit 10a and a second imaging unit 10b.
The control unit 30 determines that the first determination unit 31 that determines the occurrence of the fire point due to the combustion or the flame in the caution area, the second determination unit 32 that determines the growth of the fire point, and the flame that determines that the fire point has occurred. A fire point pattern storage unit 36 for storing a fire point pattern of an element, an image storage unit 34 for storing an image captured by the imaging unit 10 together with time-series information, and a fire element which determines that a fire point is growing And a time-series change pattern storage unit 35 for storing time-series change patterns. The time-series information may be shooting time, but may be any information that allows determination of the shooting order of the images.
The image storage unit 34 includes a first image storage unit 34a and a second image storage unit 34b, and the first image storage unit 34a stores an image captured by the first imaging unit 10a together with time-series information, The second image storage unit 34 b stores an image captured by the second imaging unit 10 b together with time-series information.
　第１判定部３１では、第１画像記憶部３４ａに記憶された画像に含まれる炎要素を、火点パターンと比較して仮判定し、第２画像記憶部３４ｂに記憶された画像に含まれる炎要素を、火点パターンと比較して仮判定し、第１画像記憶部３４ａに記憶された画像による仮判定、及び第２画像記憶部３４ｂに記憶された画像による仮判定の少なくともいずれかの仮判定があれば火点が発生していると判定する。
　第１判定部３１において、少なくともいずれかの仮判定があれば火点が発生していると判定することで、火点の発生を早い段階で識別できる。
　なお、第１判定部３１では、第１画像記憶部３４ａに記憶された画像に含まれる炎要素を、火点パターンと比較して仮判定し、第２画像記憶部３４ｂに記憶された画像に含まれる炎要素を、火点パターンと比較して仮判定し、第１画像記憶部３４ａに記憶された画像による仮判定と、第２画像記憶部３４ｂに記憶された画像による仮判定との双方の仮判定によって火点が発生していると判定することもできる。
　第１判定部３１において、双方の仮判定によって火点が発生していると判定することで、危険性の高い火点の発生を高い精度で識別できる。
　火点パターン記憶部３６に記憶する火点パターンは、あらかじめ機械学習によってパターン化されたものが好ましい。あらかじめ機械学習によってパターン化された炎要素の火点パターンは、少なくとも２つの画像による時系列的な変化情報であることが好ましく、時系列的な変化情報としては、時系列的な発生曲線若しくは変化曲線、又は近似式に関するデータである。 In the first determination unit 31, the flame element included in the image stored in the first image storage unit 34a is temporarily determined by comparison with the fire point pattern, and included in the image stored in the second image storage unit 34b. At least one of a provisional determination based on an image stored in the first image storage unit 34a and a provisional determination based on an image stored in the second image storage unit 34b by temporarily determining the flame element in comparison with the fire point pattern. If there is a temporary determination, it is determined that a fire point has occurred.
In the first determination unit 31, if there is at least one temporary determination, it is possible to identify the occurrence of the fire point at an early stage by determining that the fire point is occurring.
In the first determination unit 31, the flame element included in the image stored in the first image storage unit 34a is temporarily determined by comparing it with the fire point pattern, and the image stored in the second image storage unit 34b is used. Both the temporary determination based on the image stored in the first image storage unit 34a and the temporary determination based on the image stored in the second image storage unit 34b by temporarily determining the included fire elements by comparing with the fire point pattern It is also possible to determine that a fire point has occurred by the provisional determination of.
By determining that the fire point is generated by both the temporary determinations in the first determination unit 31, it is possible to identify the generation of the fire point having a high risk with high accuracy.
It is preferable that the fire point pattern stored in the fire point pattern storage unit 36 be previously patterned by machine learning. The fire point pattern of the flame element patterned beforehand by machine learning is preferably time-series change information by at least two images, and as time-series change information, a time-series occurrence curve or change It is data about a curve or an approximate expression.
　第２判定部３２では、第１画像記憶部３４ａ及び第２画像記憶部３４ｂに記憶された複数の画像を、時系列情報を元に比較し、画像に含まれる、火点の形状、大きさ、色、輝度、及び揺らぎの内の少なくとも一つを含む炎要素の時系列的変化から火点の成長を判定する。
　第２判定部３２では、第１画像記憶部３４ａに記憶された複数の画像に含まれる炎要素の時系列的変化から火点の成長を仮判定し、第２画像記憶部３４ｂに記憶された複数の画像に含まれる炎要素の時系列的変化から火点の成長を仮判定し、第１画像記憶部３４ａに記憶された画像による仮判定、及び第２画像記憶部３４ｂに記憶された画像による仮判定の少なくともいずれかの仮判定があれば火点が成長していると判定する。
　第２判定部３２において、少なくともいずれかの仮判定があれば火点が成長していると判定することで、火点の成長を早い段階で識別できる。
　なお、第２判定部３２では、第１画像記憶部３４ａに記憶された複数の画像に含まれる炎要素の時系列的変化から火点の成長を仮判定し、第２画像記憶部３４ｂに記憶された複数の画像に含まれる炎要素の時系列的変化から火点の成長を仮判定し、第１画像記憶部３４ａに記憶された画像による仮判定と、第２画像記憶部３４ｂに記憶された画像による仮判定との双方の仮判定によって火点が成長していると判定することもできる。
　第２判定部３２において、双方の仮判定によって火点が成長していると判定することで、危険性の高い火点の成長を高い精度で識別できる。 The second determination unit 32 compares a plurality of images stored in the first image storage unit 34 a and the second image storage unit 34 b based on time-series information, and includes the shape and size of the fire point included in the images. The growth of the fire point is determined from the time series change of the flame element including at least one of color, brightness, and fluctuation.
The second determination unit 32 temporarily determines the growth of the flash point from the time-series change of the flame element included in the plurality of images stored in the first image storage unit 34a, and is stored in the second image storage unit 34b. The temporary determination of the growth of the fire point from the time-series change of the flame element included in the plurality of images, the temporary determination based on the image stored in the first image storage unit 34a, and the image stored in the second image storage unit 34b If there is at least one of the provisional determinations by the above, it is determined that the fire point is growing.
By determining that the fire point is growing if there is at least one temporary determination in the second determination unit 32, the growth of the fire point can be identified at an early stage.
The second determination unit 32 temporarily determines the growth of the fire point from the time-series change of the flame element included in the plurality of images stored in the first image storage unit 34a, and stores it in the second image storage unit 34b. Temporary growth of the fire point is temporarily determined from the time-series change of the flame element included in the plurality of images, and the temporary determination based on the image stored in the first image storage unit 34a and the second image storage unit 34b It is also possible to determine that the fire point is growing based on both of the temporary determination with the image and the temporary determination.
By determining that the fire point is growing based on both of the temporary determinations in the second determination unit 32, the growth of the highly dangerous fire point can be identified with high accuracy.
　本実施例においても、第２判定部３２では、画像に含まれる炎の時系列的変化を、時系列的変化パターン記憶部３５に記憶している時系列的変化パターンと比較して判定することが好ましい。時系列的変化パターン記憶部３５に記憶する時系列的変化パターンは、あらかじめ機械学習によってパターン化されたものが好ましい。あらかじめ機械学習によってパターン化された炎要素の時系列的変化パターンは、時系列的な変化情報であり、時系列的な成長曲線若しくは変化曲線、又は近似式に関するデータである。
　また、第１判定部３１では、第１画像記憶部３４ａ及び第２画像記憶部３４ｂに記憶された、第１画像に含まれる炎要素、及び第１画像の後に撮影された第２画像に含まれる炎要素を、火点パターン記憶部３６に記憶している火点パターンと比較して判定し、第２判定部３２では、第１画像記憶部３４ａ及び第２画像記憶部３４ｂに記憶された、第２画像の後に撮影された第３画像を含む、少なくとも３つの画像に含まれる炎の時系列的変化を、時系列的変化パターン記憶部３５に記憶している時系列的変化パターンと比較して判定することが更に好ましい。 Also in the present embodiment, the second determination unit 32 determines the time-series change of the flame included in the image in comparison with the time-series change pattern stored in the time-series change pattern storage unit 35. Is preferred. The time-series change patterns stored in the time-series change pattern storage unit 35 are preferably those patterned in advance by machine learning. The time-sequential change pattern of the flame element patterned beforehand by machine learning is time-sequential change information, and is data on a time-series growth curve or change curve, or an approximate expression.
In the first determination unit 31, the flame element included in the first image and the second image captured after the first image, which are stored in the first image storage unit 34a and the second image storage unit 34b, are included. The fire element to be determined is determined by comparison with the fire point pattern stored in the fire point pattern storage unit 36, and the second determination unit 32 stores it in the first image storage unit 34a and the second image storage unit 34b. And the time-series change of the flame included in the at least three images, which includes the third image captured after the second image, and the time-series change pattern stored in the time-series change pattern storage unit 35 More preferably,
　図６は同火災識別装置の動作を示すフローチャートである。
　第１撮影部１０ａによって常時警戒区域内を撮影し（ステップ４ａ）、第２撮影部１０ｂによって常時警戒区域内を撮影する（ステップ４ｂ）。
　ステップ４ａで撮影される画像は画像記憶部３４ａに記憶され（ステップ５ａ）、ステップ４ｂで撮影される画像は画像記憶部３４ｂに記憶される（ステップ５ｂ）。
　ステップ５ａで記憶された画像は、第１判定部３１において、画像に含まれる炎要素が、火点パターン記憶部３６に記憶している火点パターンと比較され、ステップ５ｂで記憶された画像は、第１判定部３１において、画像に含まれる炎要素が、火点パターン記憶部３６に記憶している火点パターンと比較され、火点の発生を判定する（ステップ３）。
　ステップ３における判定では、第１画像に含まれる炎要素、及び第１画像の後に撮影された第２画像に含まれる炎要素を、火点パターン記憶部３６に記憶している火点パターンと比較して判定することが好ましい。
　ステップ３では、第１画像記憶部３４ａに記憶された画像に含まれる炎要素を、火点パターンと比較して仮判定し、第２画像記憶部３４ｂに記憶された画像に含まれる炎要素を、火点パターンと比較して仮判定し、第１画像記憶部３４ａに記憶された画像による仮判定、及び第２画像記憶部３４ｂに記憶された画像による仮判定の少なくともいずれかの仮判定があれば火点が発生していると判定する。
　またステップ３では、第１画像記憶部３４ａに記憶された画像に含まれる炎要素を、火点パターンと比較して仮判定し、第２画像記憶部３４ｂに記憶された画像に含まれる炎要素を、火点パターンと比較して仮判定し、第１画像記憶部３４ａに記憶された画像による仮判定と、第２画像記憶部３４ｂに記憶された画像による仮判定との双方の仮判定によって火点が発生していると判定することもできる。 FIG. 6 is a flowchart showing the operation of the fire identification device.
The inside of the caution area is constantly photographed by the first photographing unit 10a (step 4a), and the inside of the caution area is constantly photographed by the second photographing unit 10b (step 4b).
The image captured in step 4a is stored in the image storage unit 34a (step 5a), and the image captured in step 4b is stored in the image storage unit 34b (step 5b).
In the first determination unit 31, the fire element included in the image is compared with the fire point pattern stored in the fire point pattern storage unit 36, and the image stored in step 5b is the image stored in step 5a. In the first determination unit 31, the flame element included in the image is compared with the fire point pattern stored in the fire point pattern storage unit 36 to determine the occurrence of the fire point (step 3).
In the determination in step 3, the flame element included in the first image and the flame element included in the second image captured after the first image are compared with the fire point pattern stored in the fire point pattern storage unit 36 It is preferable to make a decision.
In step 3, the flame element included in the image stored in the first image storage unit 34a is temporarily determined by comparison with the fire point pattern, and the flame element included in the image stored in the second image storage unit 34b is Temporary determination based on the image stored in the first image storage unit 34a, and temporary determination based on the image stored in the second image storage unit 34b. If there is, it is determined that a fire point has occurred.
In step 3, the flame element included in the image stored in the first image storage unit 34a is temporarily determined by comparison with the fire point pattern, and the flame element included in the image stored in the second image storage unit 34b Is temporarily compared with the fire point pattern, and both the temporary determination by the image stored in the first image storage unit 34a and the temporary determination by the image stored in the second image storage unit 34b are performed. It can also be determined that a fire point has occurred.
　ステップ３において、火点が発生していると判定されると、ステップ５で記憶された複数の画像は、時系列情報を元に比較し、炎要素の時系列的変化から火点の成長を判定する（ステップ６）。
　ステップ３において、火点が発生していないと判定されると、ステップ４に戻って撮影部１０による撮影を行う。
　ステップ６における判定では、画像に含まれる炎の時系列的変化を、時系列的変化パターン記憶部３５に記憶している時系列的変化パターンと比較して判定することが好ましい。
　またステップ６における判定では、画像記憶部３４に記憶された、第２画像の後に撮影された第３画像を含む、少なくとも３つの画像に含まれる炎の時系列的変化を、時系列的変化パターン記憶部３５に記憶している時系列的変化パターンと比較して判定することが更に好ましい。
　ステップ６では、第１画像記憶部３４ａに記憶された複数の画像に含まれる炎要素の時系列的変化から火点の成長を仮判定し、第２画像記憶部３４ｂに記憶された複数の画像に含まれる炎要素の時系列的変化から火点の成長を仮判定し、第１画像記憶部３４ａに記憶された画像による仮判定、及び第２画像記憶部３４ｂに記憶された画像による仮判定の少なくともいずれかの仮判定があれば火点が成長していると判定する。
　またステップ６では、第１画像記憶部３４ａに記憶された複数の画像に含まれる炎要素の時系列的変化から火点の成長を仮判定し、第２画像記憶部３４ｂに記憶された複数の画像に含まれる炎要素の時系列的変化から火点の成長を仮判定し、第１画像記憶部３４ａに記憶された画像による仮判定と、第２画像記憶部３４ｂに記憶された画像による仮判定との双方の仮判定によって火点が成長していると判定することもできる。 If it is determined in step 3 that a fire point has occurred, the plurality of images stored in step 5 are compared based on time-series information, and the growth of the fire point is made from the time-series change of the fire element. Determine (step 6).
If it is determined in step 3 that no fire point has occurred, the process returns to step 4 and the imaging unit 10 performs imaging.
In the determination in step 6, it is preferable to determine the time-series change of the flame included in the image in comparison with the time-series change pattern stored in the time-series change pattern storage unit 35.
Further, in the determination in step 6, the time-series change pattern of the time-series change of the flame included in at least three images including the third image captured after the second image stored in the image storage unit 34 More preferably, the determination is made in comparison with the time-series change pattern stored in the storage unit 35.
In step 6, a plurality of images stored in the second image storage unit 34b are tentatively determined on the basis of time-series changes in flame elements included in the plurality of images stored in the first image storage unit 34a. Temporarily determine the growth of the fire point from the time-series change of the flame element included in the first image storage unit 34a, and temporarily determine the image stored in the second image storage unit 34b. It is determined that the fire point has grown if at least one of the tentative determinations of
In step 6, the growth of the flash point is temporarily determined from the time-series change of the flame element included in the plurality of images stored in the first image storage unit 34a, and the plurality of images stored in the second image storage unit 34b. Temporarily determines the growth of the fire point from the time-series change of the flame element included in the image, and temporarily determines with the image stored in the first image storage unit 34a and with the image stored in the second image storage unit 34b. It is also possible to determine that the fire point is growing based on both of the preliminary determination and the determination.
　ステップ６において、火点が成長していると判定すると、火災として判定し（ステップ７）、警報や消火のための出力を行う。
　ステップ６において、火点が成長していないと判定すると、ステップ４ａ及びステップ４ｂに戻って第１撮影部１０ａ及び第２撮影部１０ｂによる撮影を行う。 If it is determined in step 6 that the fire point has grown, it is determined that a fire has occurred (step 7), and an output for warning or extinguishing is performed.
If it is determined in step 6 that the flash point has not grown, the process returns to step 4a and step 4b, and the first imaging unit 10a and the second imaging unit 10b perform imaging.
　本実施例によれば、機械学習を用いてパターン化した火点パターンをあらかじめ記憶しておくことで、より精度の高い火点の発生の判別を瞬時に行うことができる。
　また本実施例によれば、火点の発生の判定の後に、火点の成長を炎要素の時系列的変化から判定することで、危険性の小さな火点を除外し、火災に繋がるような危険性の高い火点を識別できる。
　また本実施例によれば、機械学習を用いてパターン化した、火災に繋がるような危険性の高い時系列的変化パターンをあらかじめ記憶しておくことで、より精度の高い判別を瞬時に行うことができる。
　また本実施例によれば、第１判定部３１において、少なくともいずれかの仮判定があれば火点が発生していると判定することで、火点の発生を早い段階で識別できる。
　また本実施例によれば、第１判定部３１において、双方の仮判定によって火点が発生していると判定することで、危険性の高い火点の発生を高い精度で識別できる。
　また本実施例によれば、第２判定部３２において、少なくともいずれかの仮判定があれば火点が成長していると判定することで、火点の成長を早い段階で識別できる。
　また本実施例によれば、第２判定部３２において、双方の仮判定によって火点が成長していると判定することで、危険性の高い火点の成長を高い精度で識別できる。 According to the present embodiment, it is possible to instantaneously determine the occurrence of the firepoint with higher accuracy by storing the firepoint pattern patterned by using the machine learning in advance.
Further, according to the present embodiment, after the determination of the occurrence of the fire point, by determining the growth of the fire point from the time-series change of the fire element, the fire point having a small risk is excluded, leading to a fire. It can identify dangerous hot spots.
Further, according to the present embodiment, it is possible to instantaneously perform more accurate discrimination by previously storing a time-series change pattern having a high risk of leading to a fire, which is patterned using machine learning. Can.
Further, according to the present embodiment, when the first determination unit 31 determines that the fire point is generated if there is at least one temporary determination, the generation of the fire point can be identified at an early stage.
Further, according to the present embodiment, the first determination unit 31 can determine that the fire point is generated by both of the temporary determinations, so that the generation of the fire point having a high risk can be identified with high accuracy.
Further, according to the present embodiment, if the second determination unit 32 determines that the fire point is growing if there is at least one temporary determination, the growth of the fire point can be identified at an early stage.
Further, according to the present embodiment, it is possible to identify the growth of the high risk fire point with high accuracy, by the second determination unit 32 determining that the fire point is growing by both temporary determinations.
　なお、本実施例のように、撮影部１０として、第１撮影部１０ａと第２撮影部１０ｂとを有することで、第１判定部３１及び第２判定部３２では、第１撮影部１０ａ及び第２撮影部１０ｂで撮影される画像を元に生成した三次元画像を用いることができる。このように、三次元画像を用いることで判定精度を更に高めることができる。 As in the present embodiment, by including the first imaging unit 10a and the second imaging unit 10b as the imaging unit 10, in the first determination unit 31 and the second determination unit 32, the first imaging unit 10a and the second determination unit 32 can be used. A three-dimensional image generated based on an image captured by the second imaging unit 10b can be used. Thus, the determination accuracy can be further enhanced by using a three-dimensional image.
　本発明は、消火剤貯蔵容器に充填された消火剤を用いる消火設備における火災識別装置に最適であるが、スプリンクラーなどの消火水を用いる消火設備における火災識別装置にも適用できる。 The present invention is most suitable for a fire identification device in a fire extinguishing equipment using a fire extinguishant filled in a fire extinguishant storage container, but can also be applied to a fire identification device in a fire extinguishing equipment using fire extinguishing water such as a sprinkler.
　１０　撮影部
　１０ａ　第１撮影部
　１０ｂ　第２撮影部
　２０　サーモセンサ
　３０　制御部
　３１　第１判定部
　３２　第２判定部
　３３　閾値記憶部
　３４　画像記憶部
　３４ａ　第１画像記憶部
　３４ｂ　第２画像記憶部
　３５　時系列的変化パターン記憶部
　３６　火点パターン記憶部 10 imaging unit 10a first imaging unit 10b second imaging unit 20 thermo sensor 30 control unit 31 first determination unit 32 second determination unit 33 threshold storage unit 34 image storage unit 34a first image storage unit 34b second image storage unit 35 Temporal change pattern storage unit 36 Fire point pattern storage unit

Claims (11)

  1. 　警戒区域内における燃焼又は炎による火点の発生を判定する第１判定部と、
    前記第１判定部で前記火点の発生を判定すると、前記火点の成長を判定する第２判定部と
    を備え、
    前記警戒区域内を撮影する撮影部と、
    前記撮影部で撮影される画像を時系列情報とともに記憶する画像記憶部と
    を有する火災識別装置であって、
    前記第２判定部では、
    前記画像記憶部に記憶された複数の前記画像を、前記時系列情報を元に比較し、
    前記画像に含まれる、前記火点の形状、大きさ、色、輝度、及び揺らぎの内の少なくとも一つを含む炎要素の時系列的変化から前記火点の成長を判定する
    ことを特徴とする火災識別装置。     A first determination unit that determines the occurrence of a fire point due to combustion or flame in the caution area;
    And a second determination unit that determines growth of the fire point when the first determination unit determines the occurrence of the fire point.
    An imaging unit for imaging the inside of the caution area;
    A fire identification apparatus comprising: an image storage unit for storing an image captured by the imaging unit together with time-series information;
    In the second determination unit,
    Comparing the plurality of images stored in the image storage unit based on the time series information;
    It is characterized in that the growth of the fire point is determined from a time-series change of a fire element including at least one of the shape, size, color, brightness and fluctuation of the fire point contained in the image. Fire identification device.
  2. 　前記火点が成長していると判定する、前記炎要素の時系列的変化パターンを記憶する時系列的変化パターン記憶部を有し、
    前記第２判定部では、前記画像に含まれる前記炎の前記時系列的変化を、前記時系列的変化パターン記憶部に記憶している前記時系列的変化パターンと比較して判定する
    ことを特徴とする請求項１に記載の火災識別装置。     It has a time-series change pattern storage unit for storing time-series change patterns of the flame element, which determines that the fire point is growing,
    The second determination unit is characterized by comparing the time-series change of the flame included in the image with the time-series change pattern stored in the time-series change pattern storage unit. The fire identification device according to claim 1.
  3. 　前記警戒区域内の温度分布を検出するサーモセンサを有し、
    前記第１判定部では、
    前記温度分布の中で最高温度の位置を前記火点と仮決定し、
    仮決定した前記火点での温度が閾値以上の場合に前記火点の前記発生と判定し、
    前記第１判定部で前記火点の前記発生を判定すると、前記撮影部で撮影を開始する
    ことを特徴とする請求項１又は請求項２に記載の火災識別装置。     It has a thermo sensor which detects the temperature distribution in the said caution area,
    In the first determination unit,
    Temporarily determine the position of the highest temperature in the temperature distribution as the fire point,
    When the temperature at the temporarily determined fire point is equal to or higher than a threshold value, it is determined that the fire point is generated.
    The fire identification device according to claim 1 or 2, wherein the shooting unit starts shooting when the first determination unit determines the occurrence of the flash point.
  4. 　前記火点が発生したと判定する前記炎要素の火点パターンを記憶する火点パターン記憶部を有し、
    前記第１判定部では、
    前記画像記憶部に記憶された前記画像に含まれる前記炎要素を、前記火点パターン記憶部に記憶している前記火点パターンと比較して判定する
    ことを特徴とする請求項１又は請求項２に記載の火災識別装置。     It has a fire point pattern storage unit that stores the fire point pattern of the flame element that determines that the fire point has occurred,
    In the first determination unit,
    2. The fire point pattern according to claim 1, wherein the fire element included in the image stored in the image storage unit is determined by comparing with the fire point pattern stored in the fire point pattern storage unit. The fire identification device according to 2.
  5. 　前記火点が発生したと判定する前記炎要素の火点パターンを記憶する火点パターン記憶部を有し、
    前記第１判定部では、
    前記画像記憶部に記憶された、第１画像に含まれる前記炎要素、及び前記第１画像の後に撮影された第２画像に含まれる前記炎要素を、前記火点パターン記憶部に記憶している前記火点パターンと比較して判定し、
    前記第２判定部では、
    前記画像記憶部に記憶された、前記第２画像の後に撮影された第３画像を含む、少なくとも３つの前記画像に含まれる前記炎の前記時系列的変化を、前記時系列的変化パターン記憶部に記憶している前記時系列的変化パターンと比較して判定する
    ことを特徴とする請求項２に記載の火災識別装置。     It has a fire point pattern storage unit that stores the fire point pattern of the flame element that determines that the fire point has occurred,
    In the first determination unit,
    The flame element included in the first image and the flame element included in the second image captured after the first image, stored in the image storage unit, are stored in the fire point pattern storage unit Judged in comparison with the above-mentioned fire point pattern,
    In the second determination unit,
    The chronological change pattern storage unit, wherein the temporal change of the flame included in at least three of the images, including the third image captured after the second image, stored in the image storage unit 3. The fire identification device according to claim 2, wherein the fire identification device is determined in comparison with the time-series change pattern stored in.
  6. 　前記撮影部として、第１撮影部と第２撮影部とを有し、
    前記画像記憶部として、第１画像記憶部と第２画像記憶部とを有し、
    前記第１画像記憶部では、前記第１撮影部で撮影される前記画像を前記時系列情報とともに記憶し、
    前記第２画像記憶部では、前記第２撮影部で撮影される前記画像を前記時系列情報とともに記憶し、
    前記第２判定部では、
    前記第１画像記憶部に記憶された複数の前記画像に含まれる前記炎要素の時系列的変化から前記火点の成長を仮判定し、
    前記第２画像記憶部に記憶された複数の前記画像に含まれる前記炎要素の時系列的変化から前記火点の成長を仮判定し、
    前記第１画像記憶部に記憶された前記画像による仮判定した結果、及び前記第２画像記憶部に記憶された前記画像による仮判定した結果、少なくともいずれかの仮判定があれば前記火点が成長していると判定する
    ことを特徴とする請求項１から請求項５のいずれか１項に記載の火災識別装置。     The imaging unit includes a first imaging unit and a second imaging unit.
    The image storage unit includes a first image storage unit and a second image storage unit.
    The first image storage unit stores the image captured by the first imaging unit together with the time-series information,
    The second image storage unit stores the image captured by the second imaging unit together with the time-series information,
    In the second determination unit,
    The temporary determination of the growth of the fire point from the time series change of the flame element included in the plurality of images stored in the first image storage unit,
    The temporary determination of the growth of the fire point from the time series change of the flame element included in the plurality of images stored in the second image storage unit,
    As a result of the temporary determination by the image stored in the first image storage unit, and as a result of the temporary determination by the image stored in the second image storage unit, if there is at least one temporary determination, the fire point is The fire identification device according to any one of claims 1 to 5, which is determined to be growing.
  7. 　前記撮影部として、第１撮影部と第２撮影部とを有し、
    前記画像記憶部として、第１画像記憶部と第２画像記憶部とを有し、
    前記第１画像記憶部では、前記第１撮影部で撮影される前記画像を前記時系列情報とともに記憶し、
    前記第２画像記憶部では、前記第２撮影部で撮影される前記画像を前記時系列情報とともに記憶し、
    前記第２判定部では、
    前記第１画像記憶部に記憶された複数の前記画像に含まれる前記炎要素の時系列的変化から前記火点の成長を仮判定し、
    前記第２画像記憶部に記憶された複数の前記画像に含まれる前記炎要素の時系列的変化から前記火点の成長を仮判定し、
    前記第１画像記憶部に記憶された前記画像による仮判定した結果、前記第２画像記憶部に記憶された前記画像による仮判定した結果、双方の仮判定によって前記火点が成長していると判定する
    ことを特徴とする請求項１から請求項５のいずれか１項に記載の火災識別装置。     The imaging unit includes a first imaging unit and a second imaging unit.
    The image storage unit includes a first image storage unit and a second image storage unit.
    The first image storage unit stores the image captured by the first imaging unit together with the time-series information,
    The second image storage unit stores the image captured by the second imaging unit together with the time-series information,
    In the second determination unit,
    The temporary determination of the growth of the fire point from the time series change of the flame element included in the plurality of images stored in the first image storage unit,
    The temporary determination of the growth of the fire point from the time series change of the flame element included in the plurality of images stored in the second image storage unit,
    As a result of the temporary determination by the image stored in the first image storage unit, as a result of the temporary determination by the image stored in the second image storage unit, the fire point is grown by both temporary determinations The fire identification device according to any one of claims 1 to 5, characterized in that it is determined.
  8. 　前記撮影部として、第１撮影部と第２撮影部とを有し、
    前記画像記憶部として、第１画像記憶部と第２画像記憶部とを有し、
    前記第１画像記憶部では、前記第１撮影部で撮影される前記画像を前記時系列情報とともに記憶し、
    前記第２画像記憶部では、前記第２撮影部で撮影される前記画像を前記時系列情報とともに記憶し、
    前記第１判定部では、
    前記第１画像記憶部に記憶された前記画像に含まれる前記炎要素を、前記火点パターンと比較して仮判定し、
    前記第２画像記憶部に記憶された前記画像に含まれる前記炎要素を、前記火点パターンと比較して仮判定し、
    前記第１画像記憶部に記憶された前記画像による仮判定した結果、及び前記第２画像記憶部に記憶された前記画像による仮判定した結果、少なくともいずれかの仮判定があれば前記火点が発生していると判定する
    ことを特徴とする請求項４に記載の火災識別装置。     The imaging unit includes a first imaging unit and a second imaging unit.
    The image storage unit includes a first image storage unit and a second image storage unit.
    The first image storage unit stores the image captured by the first imaging unit together with the time-series information,
    The second image storage unit stores the image captured by the second imaging unit together with the time-series information,
    In the first determination unit,
    The flame element included in the image stored in the first image storage unit is temporarily determined by comparison with the fire point pattern,
    The flame element included in the image stored in the second image storage unit is temporarily determined by comparison with the fire point pattern,
    As a result of the temporary determination by the image stored in the first image storage unit, and as a result of the temporary determination by the image stored in the second image storage unit, if there is at least one temporary determination, the fire point is The fire identification device according to claim 4, wherein it is determined that the fire has occurred.
  9. 　前記撮影部として、第１撮影部と第２撮影部とを有し、
    前記画像記憶部として、第１画像記憶部と第２画像記憶部とを有し、
    前記第１画像記憶部では、前記第１撮影部で撮影される前記画像を前記時系列情報とともに記憶し、
    前記第２画像記憶部では、前記第２撮影部で撮影される前記画像を前記時系列情報とともに記憶し、
    前記第１判定部では、
    前記第１画像記憶部に記憶された前記画像に含まれる前記炎要素を、前記火点パターンと比較して仮判定し、
    前記第２画像記憶部に記憶された前記画像に含まれる前記炎要素を、前記火点パターンと比較して仮判定し、
    前記第１画像記憶部に記憶された前記画像による仮判定した結果、前記第２画像記憶部に記憶された前記画像による仮判定した結果、双方の仮判定によって前記火点が発生していると判定する
    ことを特徴とする請求項４に記載の火災識別装置。     The imaging unit includes a first imaging unit and a second imaging unit.
    The image storage unit includes a first image storage unit and a second image storage unit.
    The first image storage unit stores the image captured by the first imaging unit together with the time-series information,
    The second image storage unit stores the image captured by the second imaging unit together with the time-series information,
    In the first determination unit,
    The flame element included in the image stored in the first image storage unit is temporarily determined by comparison with the fire point pattern,
    The flame element included in the image stored in the second image storage unit is temporarily determined by comparison with the fire point pattern,
    As a result of the temporary determination based on the image stored in the first image storage unit, as a result of the temporary determination based on the image stored in the second image storage unit, the fire point is generated by both temporary determinations The fire identification device according to claim 4, characterized in that it is determined.
  10. 　前記撮影部として、第１撮影部と第２撮影部とを有し、
    前記画像記憶部には、前記第１撮影部及び前記第２撮影部で撮影される画像を前記時系列情報とともに記憶し、
    前記第２判定部では、前記第１撮影部及び前記第２撮影部で撮影される前記画像を元に生成した三次元画像を用いる
    ことを特徴とする請求項１又は請求項２に記載の火災識別装置。     The imaging unit includes a first imaging unit and a second imaging unit.
    The image storage unit stores images taken by the first imaging unit and the second imaging unit together with the time-series information,
    The fire according to claim 1 or 2, wherein the second determination unit uses a three-dimensional image generated based on the images captured by the first imaging unit and the second imaging unit. Identification device.
  11. 　前記撮影部として、第１撮影部と第２撮影部とを有し、
    前記画像記憶部には、前記第１撮影部及び前記第２撮影部で撮影される画像を前記時系列情報とともに記憶し、
    前記第１判定部では、前記第１撮影部及び前記第２撮影部で撮影される前記画像を元に生成した三次元画像を用いる
    ことを特徴とする請求項４に記載の火災識別装置。     The imaging unit includes a first imaging unit and a second imaging unit.
    The image storage unit stores images taken by the first imaging unit and the second imaging unit together with the time-series information,
    The fire identification apparatus according to claim 4, wherein the first determination unit uses a three-dimensional image generated based on the images captured by the first imaging unit and the second imaging unit.
PCT/JP2018/025447 2017-08-04 2018-07-05 Fire identification device WO2019026518A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2017151237A JP6980188B2 (en) 2017-08-04 2017-08-04 Fire identification device
JP2017-151237 2017-08-04

Publications (1)

Publication Number Publication Date
WO2019026518A1 true WO2019026518A1 (en) 2019-02-07

Family

ID=65233877

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2018/025447 WO2019026518A1 (en) 2017-08-04 2018-07-05 Fire identification device

Country Status (2)

Country Link
JP (1) JP6980188B2 (en)
WO (1) WO2019026518A1 (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110415480A (en) * 2019-06-24 2019-11-05 安徽和润智能工程有限公司 A kind of camera shooting security system of all-around intelligent regulation
CN110428579A (en) * 2019-08-08 2019-11-08 冯仙武 Indoor Monitoring System, method and device based on image recognition
CN113763664A (en) * 2021-09-18 2021-12-07 安徽建筑大学 Intelligent building fire control system

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111986436B (en) * 2020-09-02 2022-12-13 成都视道信息技术有限公司 Comprehensive flame detection method based on ultraviolet and deep neural networks

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH07245757A (en) * 1994-03-07 1995-09-19 Mitsubishi Heavy Ind Ltd Fault detector having visible ray camera and infrared ray camera
JPH10188168A (en) * 1996-12-20 1998-07-21 Nohmi Bosai Ltd Fire detector
JP2000076568A (en) * 1998-08-31 2000-03-14 Hochiki Corp Fire detecting device
US20040061777A1 (en) * 2002-05-20 2004-04-01 Mokhtar Sadok Detecting fire using cameras
US20100117839A1 (en) * 2008-11-11 2010-05-13 Lee Yeu Yong System and method for detecting fire
WO2015190744A1 (en) * 2014-06-09 2015-12-17 박상래 Image heat ray device and intrusion detection system using same

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPS6082388U (en) * 1983-11-09 1985-06-07 データイースト株式会社 flame detection device
JP6166650B2 (en) * 2013-12-04 2017-07-19 能美防災株式会社 Flame detection apparatus and flame detection method

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH07245757A (en) * 1994-03-07 1995-09-19 Mitsubishi Heavy Ind Ltd Fault detector having visible ray camera and infrared ray camera
JPH10188168A (en) * 1996-12-20 1998-07-21 Nohmi Bosai Ltd Fire detector
JP2000076568A (en) * 1998-08-31 2000-03-14 Hochiki Corp Fire detecting device
US20040061777A1 (en) * 2002-05-20 2004-04-01 Mokhtar Sadok Detecting fire using cameras
US20100117839A1 (en) * 2008-11-11 2010-05-13 Lee Yeu Yong System and method for detecting fire
WO2015190744A1 (en) * 2014-06-09 2015-12-17 박상래 Image heat ray device and intrusion detection system using same

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110415480A (en) * 2019-06-24 2019-11-05 安徽和润智能工程有限公司 A kind of camera shooting security system of all-around intelligent regulation
CN110428579A (en) * 2019-08-08 2019-11-08 冯仙武 Indoor Monitoring System, method and device based on image recognition
CN110428579B (en) * 2019-08-08 2022-01-18 刘宝鑫 Indoor monitoring system, method and device based on image recognition
CN113763664A (en) * 2021-09-18 2021-12-07 安徽建筑大学 Intelligent building fire control system

Also Published As

Publication number Publication date
JP2019028957A (en) 2019-02-21
JP6980188B2 (en) 2021-12-15

Similar Documents

Publication Publication Date Title
WO2019026518A1 (en) Fire identification device
CN112560657B (en) Method, device, computer device and storage medium for identifying smoke and fire
KR100901784B1 (en) System for fire warning and the method thereof
CN105512667B (en) Infrared and visible light video image fusion recognition fire method
KR102045871B1 (en) System For Detecting Fire Based on Artificial Intelligence And Method For Detecting Fire Based on Artificial Intelligence
KR100922784B1 (en) Image base fire sensing method and system of crime prevention and disaster prevention applying method thereof
KR102216961B1 (en) Fire surveillance monitoring device and method
KR20190063670A (en) Intelligent system for ignition point surveillance using composite image of thermal camera and color camera
KR102585066B1 (en) Combined fire alarm system using stand-alone fire alarm and visible light camera
KR20180021521A (en) Fire detection system using thermovision camera
KR101420130B1 (en) Identification of friend or foe system based on image processing, and control method of the same
WO2015087163A2 (en) Flame detection system and method
KR101224548B1 (en) Fire imaging detection system and method
KR101476764B1 (en) Flame dete ction method based on gray imaging signal of a cameras
KR102113527B1 (en) Thermal ip cctv system and method for detecting fire by using the same
CN113205075A (en) Method and device for detecting smoking behavior and readable storage medium
KR20070115412A (en) Fire alarm system for recording image and method of fire alarm thereof
KR101524922B1 (en) Apparatus, method, and recording medium for emergency alert
JP5302926B2 (en) Smoke detector
KR102494247B1 (en) Fire monitoring control device using blockchain network, fire monitoring control system and method including same
KR101029936B1 (en) Apparatus for fire surveillance
KR20070028812A (en) Method and apparatus for monitoring forest fire using images photographed by camera device
KR20230064388A (en) Composite image based fire detection system
Ramli et al. Aggressive Movement Feature Detection Using Color-Based Approach On Thermal Images
JP4954459B2 (en) Suspicious person detection device

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18842157

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 18842157

Country of ref document: EP

Kind code of ref document: A1