Fire analysis device and program

The fire analysis device improves fire analysis accuracy by determining flame size and simulating fire spread using image-based data, facilitating precise risk assessment and timely countermeasures.

JP2024166914AActive Publication Date: 2024-11-29KAJIMA CORP
View PDF 7 Cites 0 Cited by

Patent Information

Application Number
JP2023083344
Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
Filing Date
2023-05-19
Publication Date
2024-11-29
Estimated Expiration
2043-05-19

AI Technical Summary

Technical Problem

Existing fire analysis systems rely on text-based judgments from SNS posts, lacking accuracy in determining the physical scale of disasters, particularly fires, and do not effectively utilize image-based information for precise analysis.

Method used

A fire analysis device that acquires fire location and flame size from posted images, performs fire spread simulations using flame dimensions, map information, and weather data to predict fire spread and urgency, and provides timely notifications for countermeasures.

Benefits of technology

Enhances fire analysis accuracy by estimating flame dimensions and simulating fire spread, enabling precise risk assessment and prompt countermeasure notifications.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 2024166914000001_ABST
    Figure 2024166914000001_ABST
Patent Text Reader

Abstract

To provide a fire analysis device, etc. that enables analysis of a fire on a physical scale from contribution information of SNS etc.SOLUTION: A fire analysis device 3 comprises: acquisition means 101 which acquires the occurrence position of a fire based upon information attached to a fire image contributed onto the Internet; size estimation means 102 which finds the size of the fire from the fire image; and simulation means 104 which executes fire spread simulation using the size of the fire, and map information 106 and meteorological information.SELECTED DRAWING: Figure 3
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present invention relates to a fire analysis device and a program for analyzing a fire. [Background technology]

[0002] In recent years, with the spread of mobile devices such as smartphones, there have been an increasing number of cases where ordinary people in disaster areas post information about the disaster in text and images on various SNS (Social Networking Services). This posted information has made it possible for people in other locations to grasp the occurrence of a disaster in real time.

[0003] Patent Document 1 describes a technology that collects posted information related to disasters such as fires from social media and creates a hazard map, which makes it possible to quickly grasp the risk of disasters such as fires. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2019-109730 A Summary of the Invention [Problem to be solved by the invention]

[0005] In Patent Document 1, the impact level of a disaster is judged from text posted on SNS, but the judgment is not based on the physical scale of the disaster, and there is room for improvement in terms of the accuracy of analysis of fires, etc.

[0006] The present invention has been made in consideration of the above problems, and aims to provide a fire analysis device, etc. that enables analysis of fires on a physical scale from information posted on SNS, etc. [Means for solving the problem]

[0007] The first invention for solving the above-mentioned problems is a fire analysis device characterized by comprising an acquisition means for acquiring the location of a fire from a fire image posted on the Internet based on information accompanying the posted fire image, and a dimension estimation means for determining the dimensions of a flame from the fire image.

[0008] In this invention, when analyzing fires from information posted on the Internet, such as on social media, the accuracy of fire analysis can be improved by obtaining the location of the fire's outbreak, as well as estimating a physical measure, namely the flame dimensions, from fire images (still images or videos).

[0009] It is preferable that the fire analysis apparatus of the first invention further comprises simulation means for executing a fire spread simulation using the flame dimensions, map information and meteorological information. In the present invention, by improving the accuracy of fire analysis as described above and performing a fire spread simulation, it becomes possible to make highly accurate estimates of the risk of fire, such as the predicted time it will take for a fire to spread to a specified location and the possibility of fire spreading to a specified location.

[0010] It is preferable that the simulation means obtains a predicted time for a fire to spread to a predetermined position by performing a fire spread simulation. It is also preferable that the fire analysis device of the first invention further includes a notification means for notifying a terminal of a notified person of the predicted time. By calculating the predicted time, it is possible to grasp the urgency of fire prevention measures at a given location. By transmitting this information to the notified party's terminal, it becomes possible to take fire prevention measures in advance, such as prompt evacuation and removal of flammable materials, at the given location.

[0011] It is preferable that the fire analysis apparatus of the first invention further comprises a time series change estimation means for determining a time series change of the fire at the outbreak position using the flame dimensions. This makes it possible to grasp the progression of the fire at the location where it started.

[0012] A second invention is a program for causing a computer to function as a fire analysis device having an acquisition means for acquiring the location of a fire for a fire image posted on the Internet based on information accompanying the posted fire image, and a dimension estimation means for determining the dimensions of a flame from the fire image. Effect of the Invention

[0013] The present invention can provide a fire analysis device or the like that enables analysis of fires on a physical scale from information posted on SNS or the like. [Brief description of the drawings]

[0014] [Figure 1] FIG. 1 shows a fire analysis system 1. [Diagram 2] FIG. 2 is a diagram showing the hardware configuration of the fire analysis device 3. [Diagram 3] Diagram showing the functions of the fire analysis device 3 [Figure 4] 4 is a flowchart showing the process of the fire analysis device 3. [Diagram 5] 100 examples of fire images. [Figure 6] An example of a Fire Hazard Map 200. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0015] Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the drawings.

[0016] (1.Fire analysis device 3) 1 is a diagram showing a fire analysis system 1 including a fire analysis device 3 according to an embodiment of the present invention. The fire analysis system 1 is a system in which the fire analysis device 3 analyzes a fire based on fire-related information posted on the Internet by a poster in the area where the fire occurred, and includes the fire analysis device 3, a poster terminal 4, a server 5, a notified person terminal 6, etc.

[0017] 2 is a diagram showing the hardware configuration of the fire analysis apparatus 3. The fire analysis apparatus 3 can be realized by a computer configured by connecting a control unit 31, a storage unit 32, an input unit 33, a display unit 34, a communication unit 35, etc. via a bus or the like. However, the present invention is not limited to this, and various other configurations can be used as appropriate.

[0018] The control unit 31 is composed of a CPU, a ROM, a RAM, etc. The CPU loads programs related to the processing of the fire analysis device 3 stored in storage media such as the memory unit 32 and the ROM into a work area on the RAM and executes them. The ROM is a non-volatile memory, and permanently stores programs such as the boot program and BIOS, data, etc. The RAM is a volatile memory, and temporarily stores programs and data loaded from the memory unit 32, the ROM, etc., and also has a work area that the control unit 31 uses to perform various processes.

[0019] The storage unit 32 is a hard disk drive, a solid state drive, a flash memory, or the like, and stores programs executed by the fire analysis device 3 during processing described below, data necessary for executing the programs, an OS, and the like.

[0020] The input unit 33 is used for inputting various information to the fire analysis apparatus 3, and includes input devices such as a keyboard, a mouse, and a numeric keypad.

[0021] The display unit 34 is a liquid crystal display or the like, and displays various information related to the processing of the fire analysis apparatus 3.

[0022] The communication unit 35 is a communication interface that mediates communication between the server 5, the notified person terminal 6, etc., and the fire analysis device 3.

[0023] The poster terminal 4 is a terminal used by a poster who posts on the Internet, such as on a social networking site. In this embodiment, in particular, a fire image taken by the poster terminal 4 is posted on the Internet as posted information, and the poster terminal 4 is a mobile terminal with a camera, such as a smartphone.

[0024] The server 5 is a server of a business providing an Internet service such as an SNS, and can be realized by a computer having a control unit, a storage unit, a communication unit, and the like.

[0025] The notified party terminal 6 is a terminal used by a notified party who receives a notification of the analysis result of the fire (such as the predicted time described below), and is a mobile terminal such as a smartphone, but is not limited to this.

[0026] Fig. 3 is a block diagram showing the functions of the fire analysis device 3. As shown in Fig. 3, the fire analysis device 3 has an acquisition means 101, a size estimation means 102, a time series change estimation means 103, a simulation means 104, a notification means 105, etc., and the memory unit 32 stores map information 106 as a database.

[0027] The acquisition means 101 is a means for the control unit 31 of the fire analysis device 3 to acquire the location of a fire in a fire image posted on the Internet based on information accompanying the fire image via the communication unit 35. This accompanying information will be described later.

[0028] The size estimation means 102 is a means by which the control unit 31 of the fire analysis device 3 determines the size of a flame from a fire image.

[0029] The time series change estimation means 103 is a means by which the control unit 31 of the fire analysis device 3 determines the time series change of the fire at the outbreak position using the size of the flame and the like.

[0030] The simulation means 104 is a means for the control unit 31 of the fire analysis device 3 to execute a fire spread simulation using the flame dimensions, map information 106, meteorological information, etc. The simulation means 104 executes a fire spread simulation to obtain, as a fire analysis result, a predicted time for a fire to spread to a specified position, etc.

[0031] The notification means 105 is a means by which the control unit 31 of the fire analysis apparatus 3 notifies the notified person terminal 6 of the predicted time and the like via the communication unit 35.

[0032] The map information 106 includes information on buildings and geography of the target area. The information on buildings includes information that can be used to estimate the spread of fire within a building, such as the building's plan shape, height, distance to adjacent buildings, structure (wooden, steel frame, etc.), and fire resistance performance (fireproof building, semi-fireproof building, fireproof wooden building, wooden building, etc.), while the information on geography includes information that can be used to estimate the spread of fire between buildings, such as rivers, parks, green spaces, and roads.

[0033] (2.Fire analysis method) Next, a process (fire analysis method) performed by the fire analysis apparatus 3 of this embodiment will be described. Fig. 4 is a flowchart showing the process of the fire analysis apparatus 3. The control unit 31 of the fire analysis apparatus 3 controls each unit of the fire analysis apparatus 3 to execute each step in Fig. 4.

[0034] In this embodiment, the fire analysis device 3 first acquires fire images posted on the Internet (S1). The fire images are posted information that is taken by a poster at the location of the fire using the poster terminal 4 and posted on SNS or the like, and include both still images and videos.

[0035] In S1, the fire analysis device 3 accesses a website such as an SNS provided by the server 5, and acquires a fire image posted on the Internet by a poster. FIG. 5 is an example of a fire image 100. The fire image 100 is an image that includes a flame, and in S1, for example, an image that includes a flame is acquired as the fire image 100 from images posted on the most recent date and time. Whether or not an image includes a flame can be determined using a known image recognition technique. Instead of the posting date and time, it is also possible to use the shooting date and time that can be acquired from the image's Exif (Exchangeable Image File Format) information, etc.

[0036] Next, the fire analysis device 3 acquires the location of the fire outbreak in the fire image 100 (S2).

[0037] In S2, the fire analysis device 3 acquires the location of the outbreak of the fire based on the information accompanying the fire image 100. For example, if location information such as Global Positioning System (GPS) data of the poster terminal 4 at the time of posting the fire image 100 can be acquired as the information accompanying the fire image 100, the location can be regarded as the location of the outbreak of the fire. Also, if the shooting location of the fire image 100 can be known from metadata such as Exif information as the information accompanying the fire image 100, the shooting location can also be regarded as the location of the outbreak of the fire. Also, in a case where the fire image 100 is posted together with text such as "fire outbreak at XX (location)" as the information accompanying the fire image 100, the location described in the text can sometimes be regarded as the location of the outbreak of the fire.

[0038] Thereafter, the fire analysis device 3 estimates the size (actual size) of the flame by analyzing the fire image 100 (S3).

[0039] In S3, the fire analysis device 3 can determine the flame dimensions using AI (artificial intelligence). For example, a large number of images in which flames of various sizes are captured together with other objects are linked to the flame dimensions in each image to form training data, and known machine learning is performed to determine the relationship between the image feature amount and the flame dimensions. This makes it possible to estimate the flame dimensions from the feature amount of the fire image 100 using the relationship. Examples of the feature amount include the type of object, such as a person, a telephone pole, or a building window, and the size of the flame and object in the image. The type of object can be determined using known image recognition technology.

[0040] In addition, when a subject with an assumed size exists in the fire image 100, the size of the flame can be estimated from the relationship between the size of the subject in the fire image 100 and the size of the flame. For example, when an adult male appears as a subject in the fire image 100, and the size of the flame in the fire image 100 can be calculated to be twice that of the subject by extracting three-dimensional dimensions using a projection method such as three-point perspective projection, if the size of an adult male is assumed to be an average of 1.7 m, the size of the flame can be estimated to be 3.4 m.

[0041] The fire analysis device 3 estimates the time series change of the fire at the fire outbreak position using the flame size and the map information 106 and the like (S4).

[0042] In S4, the fire analysis device 3 judges whether the fire will spread to combustible objects using the size of the flame, the amount of radiant heat received from the flame, the distance from the fire start position to combustible objects (for example, adjacent buildings, etc.), etc., and if it is judged that the fire will spread, it similarly judges whether the fire will spread to the next combustible object using the size of the combustible object, etc. By repeating this calculation while taking into account the passage of time, it becomes possible to obtain the scale of the fire and its changes over time.

[0043] In addition to the amount of radiant heat received from the flames and the distance to the combustibles, the color of the smoke (black, white, yellow, etc.) and the presence of the fire brigade (whether there is a fire brigade and the number of fire brigade members) can also be obtained from the fire image 100 and used to estimate the time series change. For example, if there are many fire brigade members, it is considered that the fire at the outbreak location will be contained early. In addition, the fire analysis device 3 can also use other surrounding conditions obtained from the fire image 100 to estimate the time series change, and if there is audio data accompanying the fire image 100, such as in the case where the fire image 100 is a video, the audio data can be analyzed and analyzed to use the feature amount of the audio data (for example, the volume of the audio, etc.) to estimate the time series change.

[0044] Next, the fire analysis device 3 performs a fire spread simulation around the fire outbreak position (S4) to obtain the fire spread process. In S4, the fire analysis device 3 obtains meteorological information such as wind direction, wind speed, and weather via the website of the Japan Meteorological Agency or a private weather forecasting company. Then, the fire spread simulation is performed using the scale of the fire obtained from the flame dimensions, map information 106, meteorological information, etc.

[0045] The fire spread simulation is a simulation of fire spread calculated by a predetermined formula using map information 106 such as buildings and geography of the target area and meteorological information such as wind direction and wind speed. One example is a fire spread simulation program by the Ministry of Land, Infrastructure, Transport and Tourism's comprehensive technology development project "Development of disaster prevention evaluation and countermeasure technology in urban development" (http: / / www.nilim.go.jp / lab / jdg / result.htm, searched on May 8, 2023). In particular, in this embodiment, in order to determine whether a building adjacent to the fire outbreak position or a building adjacent to a building that has spread to the fire will spread, a simulation is performed using map information 106 such as the fire outbreak position, the scale of the fire, buildings in the target area, geography, meteorological information, etc. as input values. The simulation can obtain the time until the fire spreads to the adjacent building. This time is composed of, for example, the time for the fire to spread inside the building and the time for the fire to spread from the building to the building adjacent to the building.

[0046] As described above, the building information in the map information 106 includes the structure and fire resistance of the building, but these may be assumed from other information, such as whether the house is a new build or an existing one, the number of floors of the building, etc. For example, since 60% of low-rise buildings (1-2 story buildings) among new builds and existing houses are made of wood, and 55% of residential fires are caused by wooden buildings, it may be tentatively assumed that low-rise buildings are flammable buildings.

[0047] The results of the fire spread simulation can be created as a fire hazard map. Figures 6(a)-(d) are examples of fire hazard maps 200, which show buildings that are predicted to spread to a fire in color on a map of a target area around the location where the fire started. The fire hazard map 200 makes it possible to understand the urgency of fire countermeasures at each location within the area. Figures 6(a)-(d) show examples of fire hazard maps 200 at 10 minutes, 1 hour, 2 hours, 3 hours, etc., after the fire started, respectively.

[0048] The fire analysis device 3 performs a fire spread simulation and creates a fire hazard map, thereby calculating the predicted time for a fire to spread to a target point at a predetermined position within the above-mentioned area (S5). For example, in the example of Fig. 6, the predicted time for a fire to spread to target point A is calculated as 3 hours after the fire breaks out.

[0049] The fire analysis device 3 then transmits a notification of the predicted time and the like to the notified party terminal 6, providing the notified party with information about the time at which the target point A will become dangerous (S6). The notified party may be, for example, a construction manager when the target point A is a construction site, but may also be a building manager, maintenance manager, or other person in charge of a building, or other general public.

[0050] The information to be notified includes, in addition to the predicted time, the fire image 100, an alarm or text message to warn about the fire, the straight-line distance from the fire outbreak position to the target point A, a video showing the progress of the fire spread range over time as shown in Fig. 6(a) to (d), a probability of the fire spreading to the target point A, the necessity of evacuation or moving flammable materials, suggestions of evacuation destinations, etc. The content of the information to be transmitted may be different depending on whether the person is a construction manager or not, or may be the same.

[0051] If the fire spread worsens due to a change in wind direction, etc., it is possible to send the latest predicted time based on a fire spread simulation to the notified party terminal 6 as needed to warn the user. Conversely, if the fire spread improves (when information is received that the fire has been put out early, or when information is received that the fire brigade has put out the fire and is leaving), notifications to the notified party terminal 6 can be stopped.

[0052] As described above, in this embodiment, when analyzing a fire from information posted on the Internet such as SNS, in addition to acquiring the fire outbreak position, the accuracy of the fire analysis can be improved by estimating a physical measure, namely the flame dimensions, from the fire image 100. Furthermore, in this embodiment, by improving the accuracy of the fire analysis as described above and performing a fire spread simulation in S5 (see FIG. 4) described above, it becomes possible to make highly accurate estimates of the risk of fire, such as the predicted time it will take for the target point A to spread to fire and the possibility of the fire spreading to the target point A.

[0053] In addition, in this embodiment, by calculating the predicted time it will take for the fire to spread to the target point A in S6, the urgency of fire prevention measures at the target point A can be grasped, and by sending the predicted time, etc. to the notified party terminal 6 and notifying it in S7, it becomes possible to take fire prevention measures in advance at the target point A, such as rapid evacuation and removal of flammable materials.

[0054] In this embodiment, in S4, the progress of the fire at the start position can be grasped by determining the time series change at the start position of the fire.

[0055] However, the present invention is not limited to the above embodiment. For example, in this embodiment, a fire spread simulation is performed from the fire image 100 posted by the poster, but if a fire image 100 related to a fire at the same location is posted later, the input values ​​may be updated using the fire image 100, and a new fire spread simulation may be performed.

[0056] In this embodiment, the map information 106 is stored in the database of the storage unit 32, but the map information 106 may be stored in advance in another server. If the results of a field survey of the target area or other public data are available, the data can be added to the map information 106 and used for the fire spread simulation in S5.

[0057] Although the preferred embodiment of the present invention has been described above with reference to the accompanying drawings, the present invention is not limited to the above examples. It is clear that a person skilled in the art can come up with various modified or altered examples within the scope of the technical ideas disclosed in this application, and it is understood that these also naturally belong to the technical scope of the present invention. [Explanation of symbols]

[0058] 1: Fire Analysis System 3:Fire analyzer 4: Poster's terminal 5: Server 6: Notified party terminal 100: Fire image 101: Acquisition method 102: Dimension estimation means 103: Time series change estimation means 104: Simulation means 105: Notification means 200: Fire Hazard Map

Claims

1. An acquisition means for acquiring a location of a fire from a fire image posted on the Internet based on information accompanying the posted fire image; A size estimation means for calculating a size of a flame from the fire image; A fire analysis apparatus comprising:

2. 2. The fire analysis device according to claim 1, further comprising a simulation means for executing a fire spread simulation using the size of the flame, map information and meteorological information.

3. 3. The fire analysis device according to claim 2, wherein said simulation means performs a fire spread simulation to obtain a predicted time until a fire spreads to a predetermined position.

4. 4. The fire analysis apparatus according to claim 3, further comprising a notification unit for notifying a terminal of a person to be notified of the predicted time.

5. 2. The fire analysis apparatus according to claim 1, further comprising a time series change estimation means for determining a time series change of the fire at the outbreak position using the flame size.

6. Computer, An acquisition means for acquiring a location of a fire from a fire image posted on the Internet based on information accompanying the posted fire image; A size estimation means for calculating a size of a flame from the fire image; A program for causing the device to function as a fire analysis device.

Citation Information

Patent Citations

  • Fire ceneration position detection device

    JP2000331262A

  • Fire simulation device, fire simulation method and fire simulation program

    JP2018018297A

  • Fire-fighting command assistance device, method and program

    JP2019219852A

  • Control system, control method, and program

    JP2021144445A

  • Information processing device, program, and information processing method

    JP7215815B1