KR102350729B1 - A method and apparatus for determining fire-prone areas based on road traffic and congested sections and predicting fires using convolutional neural networks for firefighting facilities - Google Patents

Abstract

Disclosed are a method and apparatus for determining fire-prone areas based on road traffic and congestion sections and predicting a fire using convolutional neural networks (CNN) for firefighting facilities. The apparatus includes: at least one processor; and a memory storing instructions causing the at least one processor to perform at least one operation. The at least one operation includes the following operations of: receiving space information for a plurality of areas from a GIS server; generating firefighting space information for a firefighting facility based on the space information; determining a fire-prone area among areas by using the firefighting space information; and combining state information of two or more firefighting facilities, inputting the combined state information into an artificial neural network based on a convolutional neural network (CNN), and predicting the possibility of a fire in a building based on the output of the artificial neural network in advance.

Description

A method and apparatus for determining a fire-prone area based on road traffic volume and congestion section and predicting a fire using a convolutional neural network (CNN) for firefighting facilities TRAFFIC AND CONGESTED SECTIONS AND PREDICTING FIRES USING CONVOLUTIONAL NEURAL NETWORKS FOR FIREFIGHTING FACILITIES}

The present invention relates to firefighting facility management and fire prevention technology, and more particularly, determining a fire-prone area based on road traffic and congestion section, and predicting a fire using a convolutional neural network (CNN) for firefighting facilities It relates to a method and apparatus for doing so.

Geographic information system (GIS) refers to hardware and software on a computer, geographic data, etc. designed to effectively collect, store, update, adjust, analyze, and express all types of geographically referable information.

The geographic information system may support a user by storing, extracting, managing, and analyzing integrated information by associating topographic information representing a spatial location with attribute information representing characteristics corresponding to the topographic information.

The geographic information system is mainly used to handle land information or city information, but the scope of its use is gradually expanding. For example, research is underway to build a risk map related to flooding or flooding based on GIS.

On the other hand, firefighting facilities are machines, devices, and systems that can detect and warn fire to protect or evacuate people, and to immediately extinguish fire in the initial stage of fire. It is necessary to keep the fire alarm effect tense at all times by disposing firefighting facilities close to a high risk of fire and preventing malfunctions due to elapse of a handling period, etc. in advance.

However, it is common for firefighting facilities to be inspected and managed by on-site inspection personnel with substantial labor input, but it is difficult for a small number of on-site inspection personnel to inspect all firefighting facilities because a very large number of firefighting devices are installed in one building. There are many situations like this. In addition, even if the firefighting facility is operating, it is difficult to find an appropriate countermeasure if it is not possible to accurately detect whether a fire has occurred, or if a problem occurs in some of the firefighting facilities, so that the firefighting facilities differently warn whether or not a fire occurs. .

An object of the present invention for solving the above problems is to determine a fire-prone area based on road traffic and congestion section, and predict a fire using a convolutional neural network (CNN) for firefighting facilities and a method and apparatus is to provide

One aspect of the present invention for achieving the above object is a method and apparatus for determining a fire-prone area based on road traffic volume and congestion section, and predicting a fire using a convolutional neural network (CNN) for fire-fighting facilities. to provide.

An apparatus for determining a fire-prone area based on the road traffic volume and congestion section, and predicting a fire using a convolutional neural network (CNN) for firefighting facilities, includes: at least one processor; and a memory that stores instructions instructing the at least one processor to perform at least one operation.

The at least one operation may include: receiving spatial information on a plurality of regions from a GIS server; generating fire-fighting spatial information for a fire-fighting facility based on the spatial information; and determining a fire-prone area among the areas by using the fire-fighting space information.

The spatial information includes topographic information on the plurality of regions and attribute information mapped to a specific location of the topographic information, and the attribute information may include road network information and building information.

The generating of the fire-fighting spatial information may include generating the fire-fighting spatial information by combining the spatial information with the information on the fire-fighting facility.

The information on the firefighting facility may include a type of the firefighting facility, an available period, an inspection time, a location coordinate corresponding to the topographic information, a unique identifier, and state information.

The fire-fighting space information may include at least one of a location of a fire station in each of the regions, a real-time movement state of a fire engine operated by the fire station, and a fire-fighting water facility disposed outdoors.

The operation of determining the fire vulnerable area includes determining a first road area into which the fire engine can enter by using the road network information, and disposing the fire station at a position closest to the location of the fire station within the first road area, , the fire station may determine a second road area accessible within a preset time, and determine the fire vulnerable area based on the second road area.

The at least one operation further includes combining state information of two or more firefighting facilities, inputting the combined state information into an artificial neural network, and predicting in advance the possibility of a fire in the building based on the output of the artificial neural network. can do.

In the case of using the GIS-based firefighting facility management and fire prevention method and apparatus using the artificial neural network according to the present invention as described above, the firefighting facility is managed using GIS-based spatial information. You can check the current status and operation status in real time.

In addition, while monitoring the operation state of the firefighting facilities placed in the building through spatial information in real time, the operation of the firefighting facilities is learned using an artificial neural network, and fire occurrence is predicted based on the learned operation state. It is possible to solve the problem of operation errors that occur frequently in

1 is a schematic diagram for explaining a method of managing a GIS-based firefighting facility and preventing a fire using an artificial neural network according to an embodiment of the present invention.
2 is a view for explaining a method of determining a fire vulnerable area according to an embodiment.
3 is a view for explaining a method of monitoring an operating state of a firefighting facility according to an embodiment of the present invention.
4 is a conceptual diagram for explaining a method of predicting the possibility of a fire by using state information of a firefighting facility according to an exemplary embodiment.
5 is a conceptual diagram for explaining the structure of a second artificial neural network according to an embodiment of the present invention.
6 is a diagram exemplarily showing a hardware configuration of a fire management server according to an embodiment.

Since the present invention can have various changes and can have various embodiments, specific embodiments are illustrated in the drawings and described in detail in the detailed description. However, this is not intended to limit the present invention to specific embodiments, and it should be understood to include all modifications, equivalents and substitutes included in the spirit and scope of the present invention. In describing each figure, like reference numerals have been used for like elements.

Terms such as first, second, A, and B may be used to describe various elements, but the elements should not be limited by the terms. The above terms are used only for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, a first component may be referred to as a second component, and similarly, a second component may also be referred to as a first component. and/or includes a combination of a plurality of related listed items or any of a plurality of related listed items.

When an element is referred to as being “connected” or “connected” to another element, it is understood that it may be directly connected or connected to the other element, but other elements may exist in between. it should be On the other hand, when it is said that a certain element is "directly connected" or "directly connected" to another element, it should be understood that the other element does not exist in the middle.

The terms used in the present application are only used to describe specific embodiments, and are not intended to limit the present invention. The singular expression includes the plural expression unless the context clearly dictates otherwise. In the present application, terms such as “comprise” or “have” are intended to designate that a feature, number, step, operation, component, part, or combination thereof described in the specification exists, but one or more other features It should be understood that this does not preclude the existence or addition of numbers, steps, operations, components, parts, or combinations thereof.

Unless defined otherwise, all terms used herein, including technical and scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with the meaning in the context of the related art, and should not be interpreted in an ideal or excessively formal meaning unless explicitly defined in the present application. does not

Hereinafter, preferred embodiments according to the present invention will be described in detail with reference to the accompanying drawings.

1 is a schematic diagram illustrating a method of managing a GIS-based firefighting facility and preventing a fire using an artificial neural network according to an embodiment of the present invention.

Referring to FIG. 1 , a method of managing a GIS (geographic information system)-based firefighting facility and preventing a fire using an artificial neural network generates firefighting space information about the firefighting facility in conjunction with the GIS server 200 and , can be performed by the fire management server 100 that manages firefighting facilities and prevents fires using the generated firefighting space information.

The GIS server 200 may store spatial information on a plurality of regions into a database. For example, the GIS server 200 may store geographic information for a plurality of regions and spatial information including attribute information mapped to a specific location of the geographic information. For example, the attribute information may include road network information and building information. For example, the road network information includes the width of the road, the allowable height of the road, the congested section for each road, the congestion time for each road, the traffic volume for each road, etc., and the building information includes the floor plan of each floor in the building, the height of the building, the building of use (fire department, office use, commercial use, etc.), and the like.

For example, the GIS server 200 is deployed in a geographic information system (GIS) control center that is built and operated in advance by the government or public institutions, etc. It may be an operating server of a geographic information system or a database thereof.

The firefighting management server 100 may receive spatial information from the GIS server 200 and generate firefighting spatial information by combining the received spatial information with information on firefighting facilities disposed in each area. The firefighting management server 100 may store firefighting spatial information in an internal storage or a separate distributed database, and update the stored firefighting spatial information using spatial information received from the GIS server 200 in real time.

For example, the fire management server 100 may display topographic information included in spatial information, and the manager of the fire management server 100 may input information about firefighting facilities in specific locations of the topographic information. . In more detail, the manager of the fire management server 100 may input information about the fire fighting facility on the floor plan of each floor inside the building included in the topographic information. As another example, the firefighting management server 100 may receive information on firefighting facilities disposed in each building for each region from a building operation server installed in each building.

The information about the firefighting facility may include a type of firefighting facility, an available period, an inspection time, an arrangement location (eg, location coordinates corresponding to geographic information), a unique identifier, state information, and the like. The status information is information indicating the real-time current status of the firefighting facility, and may be normal, abnormal, alarm, off, and the like. Types of firefighting facilities may include firefighting equipment, alarm equipment, evacuation and rescue equipment, firefighting water equipment, firefighting activity equipment, CCTV, and the like.

Fire extinguishing equipment is a machine, apparatus or equipment that extinguishes fire using water or other fire extinguishing agents, and may include fire extinguishing equipment, automatic fire extinguishing equipment, indoor fire hydrant equipment, sprinkler equipment, outdoor fire hydrant equipment, and the like.

Alarm equipment is a machine, device, or facility that notifies the fact that a fire has occurred. etc.), automatic fire warning facilities, emergency broadcast facilities, temperature sensors, and the like.

Evacuation and rescue equipment is an appliance or equipment used to evacuate in case of fire, and may include an evacuation mechanism, a lifesaving mechanism, a guide lamp, an emergency lighting lamp, and the like.

The fire extinguishing water facility is a facility that supplies or stores water necessary to extinguish a fire, and may include a water supply fire extinguishing water facility, a fire extinguishing tank, a water storage tank, and the like.

Firefighting equipment is equipment used for extinguishing fires or life-saving activities, and may include smoke control equipment, connected water pipe equipment, connected water spray equipment, emergency outlet equipment, wireless communication auxiliary equipment, combustion prevention equipment, and the like.

The fire management server 100 may determine a fire vulnerable area for each area based on the generated fire fighting space information, and provide the determined fire vulnerable area to the fire management terminal 300 . In an embodiment of the present invention, the fire-vulnerable area may be defined as an area in which it is difficult for a fire engine to quickly approach and extinguish a fire or there is a restriction in using surrounding fire-fighting equipment.

In addition, the fire fighting management server 100 may detect the operation state of the fire fighting facilities in each building in real time, and transmit fire space information including the detected operation state of the fire fighting facilities to the fire management terminal 300 . The firefighting management terminal 300 may monitor the operation state of the firefighting facilities disposed in each area and buildings in real time based on the operation state received from the firefighting management server 100 .

The fire management server 100 organically combines the state information of two or more firefighting facilities, inputs the combined state information to an artificial neural network, and whether or not a fire in the building occurs based on the output of the artificial neural network may be predicted in advance, and when it is predicted that a fire will occur in a building, a building fire warning may be transmitted to the firefighting management terminal 300 .

For example, a communication-enabled desktop computer (desktop computer), a laptop computer (laptop computer), a notebook (notebook), a smart phone (smart phone), a tablet PC (tablet PC), mobile phone (mobile) of the fire management terminal 300, phone, smart watch, smart glass, e-book reader, PMP (portable multimedia player), portable game console, navigation device, digital camera, DMB (digital multimedia broadcasting) ) player, digital audio recorder, digital audio player, digital video recorder, digital video player, PDA (Personal Digital Assistant), and the like.

Accordingly, the firefighting management server 100 according to an embodiment of the present invention combines the spatial information received from the GIS server 200 with information about the firefighting facility to generate and build firefighting spatial information, thereby managing the firefighting facility. It has the advantage of being able to quickly respond to a fire by predicting the possibility of a fire in advance using an artificial neural network and alerting the firefighting management terminal 300 before a fire occurs.

2 is a view for explaining a method of determining a fire vulnerable area according to an embodiment.

Referring to FIG. 2 , the fire space information generated by the fire management server 100 includes the location of the fire station 119 in each area, the real-time movement status of the fire engine 118 operated by the fire station 119, and the location of the fire station 119 in each area. It may include a fire extinguishing water facility 117 and the like.

Accordingly, the fire management terminal 300 uses the fire space information received from the fire management server 100 to determine the location of the fire station 119, the real-time movement state of the fire engine 118, and the fire extinguishing water facility 117 disposed outdoors. ) can be monitored in real time.

In addition, the fire management server 100 may determine a fire vulnerable area (DAREA) by using the fire space information, combine the determined fire vulnerable area (DAREA) with the fire space information, and transmit it to the fire management terminal 300 . . The fire management terminal 300 may check the fire-prone area (DAREA) in real time through fire-fighting management information, and may support easy additional measures to supplement the fire-prone area (DAREA).

In order to determine the fire vulnerable area (DAREA), the fire management server 100 may determine a first road area into which a fire engine can enter by using road network information included in the fire space information. For example, the fire management server 100 may determine the first road area by comparing the width and the allowable height of the road included in the road network information with the width and height of a currently operated fire engine.

The fire management server 100 arranges the fire station 119 in the position closest to the fire station 119 within the determined first road area, and is accessible within a preset time from the fire station 119 disposed in the first road area. A second road area may be determined. For example, if 5 to 10 minutes have elapsed from the time of the fire, the preset time may be determined as a value selected from 5 to 10 minutes in consideration of the high tendency of building fires to escalate into large-scale fires. have. In this case, the fire management server 100 may obtain the traffic volume and congestion section of the road located in the first road area from the road information, and determine the second road area according to the obtained traffic volume and congestion section.

For example, the fire management server 100 may determine the second road area based on Equation 1 below.

That is, the fire management server 100 is a congested section among sections within the first road area starting with the preset time t and the fire station 119 as a starting point, and in some sections, congested sections and some sections of the remaining sections excluding the congested section. When the relationship between the amount of traffic in Equation 1 above is satisfied, the second road area may be determined as a corresponding congestion section and a partial section. In Equation 1 above, k is a proportionality constant, v1 is an average speed during a congestion section, v0 is an average speed during a normal section, v1 is smaller than v0, and v1 and v0 may be preset.

The fire management server 100 uses the determined straight-line distance (l), population density (p), building density (b), and fire rate (f) from the second road area to determine the fire risk at each location in the area. It can be calculated as in Equation 2 below.

Referring to Equation 2, w1 to w4 are preset weight values for each factor, and the fire management server 100, the straight line distance (l), population density (p), building density (b), fire rate The weighted average of (f) can be calculated as the fire risk.

The fire management server 100 may determine locations where the calculated fire risk exceeds a preset threshold as fire vulnerable areas (DAREA). In addition to the fire risk according to Equation 2 above, the fire management server 100 additionally adds a fire vulnerable area (DRAEA) according to the material of the building included in the spatial information, the use of the building (for example, whether to handle dangerous substances), etc. may decide

The fire management server 100 may add the determined fire vulnerable area (DAREA) to the fire space information and transmit it to the fire management terminal 300 , and the fire management terminal 300 may transmit the fire vulnerable area (DAREA) to the topographic information It can be displayed to the manager as spatial information superimposed with At this time, the fire management terminal 300 may display the color of the fire vulnerable area (DAREA) differently according to the previously calculated size of the fire risk so that the manager can easily understand the fire risk level.

3 is a view for explaining a method of monitoring an operating state of a firefighting facility according to an embodiment of the present invention.

Referring to FIG. 3 , the firefighting management server 100 may receive state information of firefighting facilities installed on each floor in the building from an input of a manager or an external operation server within the building.

For example, as shown in FIG. 3 , the fire management server 100 arranges fire fighting facilities installed on each floor in the building in the fire space information, and adds status information for each of the placed fire facilities to the fire space information. can

Here, the status information may be at least one of normal, abnormal, alarm, and off, but is not limited thereto. For example, when the firefighting facility is a temperature sensor, it may include a measured temperature sensing value, and when the firefighting facility is a CCTV installed on each floor, it may include a real-time image captured through the CCTV.

Accordingly, the fire management server 100 may receive and monitor real-time images of each floor in the building through CCTV, and may monitor status information of other firefighting facilities to immediately determine whether a fire alarm is present.

Meanwhile, the fire management server 100 may obtain an output value quantifying the possibility of a fire by analyzing the real-time image captured by the CCTV through the first artificial neural network 10 ′.

For example, the first artificial neural network 10 ′ includes only the convolutional layer 11 , the activation layer 12 , and the pooling layer 13 in the second artificial neural network 10 described with reference to FIG. 5 . may be a convolutional neural network, and an output of the pooling layer 13 may be output as an output value of the first artificial neural network 10'. That is, since the pooling layer 13 digitizes and outputs the input value as a value in a specific range (for example, 0 and 1) using an activation function, when using the output of the pooling layer 13, the possibility of fire A numerical output can be obtained.

The first artificial neural network 10 ′ may be supervised learning various fire scene images in advance, and may receive a real-time image captured by CCTV, digitize whether a fire has occurred, and output it as an output value.

That is, the first artificial neural network 10 'predicts the occurrence of a fire by synthesizing the state information of two or more firefighting facilities to be described later by outputting a numerical value of the possibility of a fire as an output value, rather than immediately determining whether or not a fire occurs. is used for

4 is a conceptual diagram for explaining a method of predicting the possibility of a fire by using state information of a firefighting facility according to an exemplary embodiment.

Referring to FIG. 4 , the fire management server 100 may divide fire space information corresponding to each floor in a building into grid-type zones having preset sizes. In this case, the firefighting management server 100 may partition the firefighting space information so that one alarm facility (or CCTV) is disposed in one gridzone.

In addition, the fire management server 100 may generate at least two matrices mat1 and mat2 having component values corresponding to each of the partitioned grid zones. That is, the fire management server 100 transmits state information of an alarm facility or CCTV that can be an indicator for determining the possibility of a fire among firefighting facilities arranged in at least some of the grid-type zones. It is possible to create at least two matrices (mat1, mat2) with component values corresponding to .

In this case, at least two matrices mat1 and mat2 may be generated such that they correspond 1:1 with an alarm facility or CCTV, and component values are determined by state information. For example, the fire management server 100 may generate a first matrix mat1 by matching the state information of CCTV to a grid-type zone (GridZone), and grid state information of a heat detector, which is one of the alarm facilities. A second matrix mat2 may be generated in correspondence with the type zone GridZone.

Referring to Figure 4, the fire management server 100, the first matrix (mat1) having an output value obtained by inputting a real-time image, which is state information of CCTV, into the first artificial neural network 10' as a component value and a heat detector The second matrix ( mat2) can be created.

In particular, in the case of the first matrix mat1 corresponding to the CCTV, the real-time image, which is state information, is not as a component value, but an output value obtained by inputting the real-time image to the first artificial neural network 10' as a component value. By doing so, it is possible to quantify the presence or absence of fire in CCTV and form a matrix.

Fire management server 100, the remaining component values that do not correspond to the state information of the alarm equipment or CCTV among the component values of the generated at least two matrices (mat1, mat2) using the padding value to process the padding (padding) can That is, the fire management server 100 may generate at least two matrices mat1 and mat2 having all component values by filling the remaining component values with padding values. In this case, the padding value may be 1, but is not limited thereto.

Fire management server 100 may standardize each component value of at least two matrices (mat1, mat2). For example, the fire management server 100 calculates the average and standard deviation of the component values of the first matrix mat1, divides the calculated average from the component values, and then divides the first matrix mat1 by the standard deviation. ) can be standardized.

The fire management server 100 may generate a final matrix by dot product of at least two matrices mat1 and mat2 having standardized component values with each other.

In this case, the second artificial neural network 10 used by the fire management server 100 to predict the possibility of a fire may be a convolutional neural network (CNN). CNN has the most research in the field of AI using artificial neural networks, so it has high performance, receives an image as an input value, classifies the received image into two or more classes, and sets the class with the highest value as the output value. can be output as In an embodiment of the present invention, a class may mean a plurality of levels (eg, very high, high, normal, low, very low) indicating the possibility of a fire.

The fire management server 100 converts the final matrix into an image, inputs the converted image to the second artificial neural network 10 , and obtains an output value indicating the possibility of fire as an output of the second artificial neural network 10 . can do.

In this case, in order to convert the final matrix into an image, the fire management server 100 may convert each of the component values constituting the final matrix to a pixel value to convert the final matrix into a grid-shaped image. At this time, in order to determine the pixel value corresponding to the component value, the fire management server 100 sets the standardized component value range (eg, between 0 and 1) to the pixel value range (eg, 8-bit pixels). case between 0 and 255) and can be matched at equal intervals.

The second artificial neural network 10 generates a final matrix by using the state information of the firefighting equipment or CCTV for each step so as to classify the image input in the plurality of steps described above, and converts the generated final matrix into an image for each A step-by-step image may be supervised as a learning image.

As described above, the fire management server 100 according to an embodiment of the present invention predicts the possibility of a fire by using the second artificial neural network 10 based on the state information of the CCTV and alarm facilities, so a specific alarm facility It is possible to predict and respond to the possibility of a fire early by breaking away from the conventional fire alarm level that relies on the alarm of In particular, since the final matrix is formed by synthesizing the state information of a plurality of alarm facilities, even if there is a condition such as malfunction or off of a specific alarm facility, the possibility of fire is predicted by considering all the overall operations of such alarm facilities, so the prediction reliability is low. has a high advantage.

5 is a conceptual diagram for explaining the structure of a second artificial neural network according to an embodiment of the present invention.

Referring to FIG. 5 , a second artificial neural network 10 according to an embodiment of the present invention receives an image of a preset size (an image converted from a final matrix) as an input image, and extracts a feature map The convolutional layer 11 that performs the following functions, the activation layer 12 that determines whether to activate the output by using the activation function for the extracted features, and the pooling layer 13 that performs sampling on the output according to the activation layer 12 ), a fully connected layer 14 that performs classification according to a class, and an output layer 15 that finally outputs an output according to the fully connected layer 14 .

The convolutional layer 11 may be a layer that extracts features of input data by convolutional product of an input image and a filter. Here, the filter is a function that detects a characteristic part of the input image, is generally expressed as a matrix, and may be a function that is determined as it is continuously learned by the training data. The feature extracted by the convolutional layer 11 may be referred to as a feature map. Also, an interval value for which convolution is performed may be referred to as a stride, and a feature map having a different size may be extracted according to the stride value. In this case, when the size of the filter is smaller than that of the input image, the feature map has a size smaller than that of the existing input image, and a padding process may be additionally performed to prevent the feature from being lost through several steps. In this case, the padding process may be a process of maintaining the same size of the input image and the size of the feature map by adding a preset value (eg, 0 or 1) to the periphery of the generated feature map.

Here, the convolutional layer 11 according to an embodiment of the present invention may use a structure in which a 1×1 convolutional layer and a 3×3 convolutional layer are sequentially and repeatedly connected.

The activation layer 12 is a layer that determines whether to activate or not by changing a feature extracted with a certain value (or matrix) to a non-linear value according to the activation function. (softmax) function or the like may be used. For example, the softmax function may be a function having a characteristic that all input values are normalized to values between 0 and 1 and the sum of output values is always 1.

The pooling layer 130 performs subsampling or pooling on the output of the activation layer 12 to select a feature representing the feature map, and has the largest value for a certain region of the feature map. Max pooling for extracting , average pooling for extracting an average value, etc. may be performed. In this case, the pooling layer is not necessarily performed after the activation function, but may be selectively performed.

Also, here, the second artificial neural network 10 may include a plurality of connection structures of the convolutional layer 11 , the activation layer 12 , and the pooling layer 13 .

6 is a diagram exemplarily showing a hardware configuration of a fire management server according to an embodiment.

Referring to FIG. 6 , the fire management server 100 includes at least one processor 110 ; and a memory 120 that stores instructions instructing the at least one processor 110 to perform at least one operation.

The fire management server 100 may also be referred to as a device for managing a GIS-based fire fighting facility and preventing a fire.

The at least one operation may include at least some of the operations of the fire management server 100 described with reference to FIGS. 1 to 5 .

Here, the at least one processor 110 may mean a central processing unit (CPU), a graphics processing unit (GPU), or a dedicated processor on which methods according to embodiments of the present invention are performed. can

The memory 120 may be configured as at least one of a volatile storage medium and a non-volatile storage medium. For example, the memory 120 may be configured as at least one of a read only memory (ROM) and a random access memory (RAM). For example, the storage device 160 may be a hard disk drive (HDD), a solid state drive (SSD), or the like.

In addition, the fire management server 100 may include a transceiver 130 for performing communication through a wireless network. In addition, the fire management server 100 may further include an input interface device 140 , an output interface device 150 , a storage device 160 , and the like. Each of the components included in the fire management server 100 may be connected by a bus 170 to communicate with each other.

The methods according to the present invention may be implemented in the form of program instructions that can be executed by various computer means and recorded in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc. alone or in combination. The program instructions recorded on the computer-readable medium may be specially designed and configured for the present invention, or may be known and available to those skilled in the art of computer software.

Examples of computer-readable media may include hardware devices specially configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions may include not only machine language codes such as those generated by a compiler, but also high-level language codes that can be executed by a computer using an interpreter or the like. The hardware device described above may be configured to operate as at least one software module to perform the operations of the present invention, and vice versa.

In addition, the above-described method or apparatus may be implemented by combining all or part of its configuration or function, or may be implemented separately.

Although the above has been described with reference to the preferred embodiment of the present invention, those skilled in the art can variously modify and change the present invention within the scope without departing from the spirit and scope of the present invention as described in the claims below. You will understand that it can be done.

100: fire management server 200: GIS server
300: fire management terminal 10: first artificial neural network
10': second artificial neural network 11: convolutional layer
12: activation layer 13: pooling layer
14: fully connected layer 15: output layer

Claims (1)

A device that determines a fire-prone area based on road traffic volume and congestion section, and predicts a fire using a convolutional neural network for firefighting facilities,
at least one processor; and
and a memory for storing instructions instructing the at least one processor to perform at least one operation,
The at least one operation is
receiving spatial information for a plurality of regions from the GIS server;
generating fire-fighting spatial information for a fire-fighting facility based on the spatial information;
determining a first road area into which a fire engine can enter by using road network information included in the fire fighting space information;
determining a second road area accessible within a preset time based on a first location closest to a location of a fire station within the first road area, and determining a fire vulnerable area based on the second road area; and
Combining the state information of two or more firefighting facilities, inputting the combined state information into an artificial neural network based on a convolutional neural network (CNN), and predicting the possibility of a fire in a building based on the output of the artificial neural network in advance do,
The spatial information includes topographic information for the plurality of regions and attribute information mapped with a specific location of the topographic information, wherein the attribute information includes road network information and building information,
The operation of predicting the possibility of a fire in the building in advance,
dividing the fire space information corresponding to each floor in the building into grid zones having a preset size;
generating at least two matrices corresponding to each of the firefighting facilities disposed in at least some of the grid-like zones;
generating a final matrix by dot producting the at least two matrices with each other;
converting each of the component values constituting the final matrix to a pixel value to convert the final matrix into a grid-shaped image; and
and inputting the lattice-shaped image into the artificial neural network to predict the possibility of a fire in the building in advance,
At least some of the component values of each of the at least two matrices correspond to state information of each of the firefighting facilities disposed in at least some of the grid-type zones,
The operation of generating the at least two matrices comprises:
and padding the remaining component values that do not correspond to the state information among component values of each of the at least two matrices by using a preset padding value.

Images