KR102290667B1 - Building firefighting monitoring system using artificial intelligence - Google Patents

Abstract

A building firefighting monitoring system using artificial intelligence is provided. The building firefighting monitoring system using artificial intelligence includes at least one alarm, at least one CCTV, and a fire monitoring server. An object of the present invention is to accurately detect a fire at an early stage by detecting a fire first by processing CCTV images and detecting a fire secondarily by using deep learning-based artificial intelligence.

Description

Building firefighting monitoring system using artificial intelligence}

The present invention relates to a firefighting and fire monitoring system for buildings using artificial intelligence.

Early fire detection and evacuation route guidance system are very important parts in building management. In order to minimize human casualties in case of fire, a system that detects flame or smoke early is necessary.

Conventional fire detection systems often use a differential detector that detects heat through air expansion due to heat, but when the temperature around the detector increases, the fire has already spread to some extent, so it is vulnerable to initial fire detection. .

In addition, although smoke detectors such as photoelectric detectors or ionization detectors are used, it is difficult to generalize due to economic problems. Since it is sensitive, there is a problem of malfunction due to dew or the like.

In the case of a flame detection sensor, since the UV detection method is used, the sensitivity may be lowered due to absorption of UV light by other floating objects, and there is a disadvantage in that the possibility of false detection is high, such as reacting to welding light.

Accordingly, there is a need for a fire detection system capable of detecting a fire early while reducing malfunctions by solving the problems of the prior art.

The problem to be solved by the present invention is image processing and deep learning-based artificial intelligence that can detect fire at an early stage and accurately detect fire by first detecting fire by processing CCTV images and secondly detecting fire using deep learning-based artificial intelligence. Its purpose is to provide a building firefighting and fire monitoring system using intelligence.

Furthermore, through CCTV including a heat sensor and a location sensor, it monitors the location of the fire and the progress of the fire more precisely, and at the same time, a building firefighting fire using artificial intelligence that can detect the presence or absence of a person in distress and the location of the person in distress. It aims to provide a monitoring system.

In addition, it aims to provide a building firefighting monitoring system using artificial intelligence that can effectively detect blind spots of CCTV.

In addition, an object of the present invention is to provide a building firefighting monitoring system using artificial intelligence that can quickly escape in case of fire through an alarm that can be operated independently while interlocking with CCTV.

The problems of the present invention are not limited to the problems mentioned above, and other technical problems not mentioned will be clearly understood by those skilled in the art from the following description.

In the building fire-fighting monitoring system using artificial intelligence, the building fire-fighting monitoring system using artificial intelligence according to an embodiment for solving the above problems includes at least one alarm, at least one CCTV, and a fire monitoring server, The at least one alarm includes a body part, a speaker part disposed on a lower surface of the body part, and a cover part surrounding the speaker part, wherein the body part includes a first module part and a second module part spaced apart from each other, The first module part and the second module part are embedded in an outer surface of the body part, and each of the first module part and the second module part includes a first sensor part and a first light emitting part sequentially arranged in a first direction. , a second sensor unit and a second light emitting unit, wherein each of the first sensor unit and the second sensor unit includes at least one of a smoke sensor, a heat sensor, and a humidity sensor, and the cover unit includes the first part and the second a plurality of parts, wherein the first part and the second part are alternately arranged in the first direction, and the first part is made of tedradecane, the second part is made of a metal material, and the at least One CCTV includes a fixing part for fixing the CCTV, an extension part connected to the fixing part, an external housing part connected to the extension part, and an inner housing part accommodated in the external housing part and including a photographing part and an auxiliary photographing part The fixing unit includes a plurality of first fixing unit air passages and a second fixing unit air passage, the inside of the first fixing unit air passage is filled with silicon, and the inside of the second fixing unit air passage is Tedradecane. The filling, the first fixing part air passage and the second fixing part air passage are alternately arranged with each other, each of the first fixing part air passage and the second fixing part air passage is made of a columnar shape, the first fixing part The diameter of each of the air passage and the air passage of the second fixing part is 0.05 cm to 0.4 cm, and the extension part connects the fixing part and the external housing part, but the length is variable, and the external housing part is made of a metal material, the inner on the first outer housing It includes a plurality of secondary air passages and a second external housing part air passage, the inside of the first external housing part air passage is filled with silicon, and the inside of the second external housing part air passage is filled with tedradecane, the first external Each of the housing part air passage and the second outer housing part air passage has a cylindrical shape, and the diameter of each of the first external housing part air passage and the second external housing part air passage is 0.1 cm to 0.8 cm, and the external A first sensor module and a second sensor module are disposed on the lower surface of the housing, the first sensor module includes a thermal sensor, and the second sensor module receives a voice recognition sensor for recognizing a human voice frequency and a GPS signal. and a GPS sensor for analyzing and acquiring location information, wherein each of the first sensor module and the second sensor module has a heat conduction shield disposed on the outer surface, wherein the heat conduction shield is 75 wt% of glass fiber, 17 wt% of phenolic resin, carbonic acid Calcium is made by including 8% by weight, the thickness of the heat conduction barrier layer is 0.1 cm to 0.3 cm, the heat conduction barrier layer includes a plurality of holes, and the diameter of the plurality of holes is 0.01 cm to 0.03 cm.

The fire monitoring server includes a communication unit that receives the image captured from the at least one CCTV, a fire determination unit that judges whether the photographed image is a fire using deep learning-based artificial intelligence, and the fire determination unit If it is determined that has occurred, a fire notification unit that transmits a fire notification to the manager terminal and the at least one CCTV through the communication unit, and an evacuation route when receiving fire status confirmation information from the manager terminal that has received the fire notification An evacuation route guide unit for guiding the a person location notification unit for transmitting the presence or absence of a person and the location of the person to the manager terminal, the at least one CCTV further includes a microcontroller, and the first sensor module detects heat exceeding a set value, A first control signal for moving the photographing unit in a direction in which heat exceeding the set value is sensed is transmitted to the microcontroller, and the second sensor module is configured to, when a person is detected, in the direction in which the person is detected. A second control signal for moving the photographing unit may be transmitted to the microcontroller.

According to the building fire-fighting fire monitoring system using artificial intelligence according to an embodiment, it is possible to more accurately detect the occurrence of a fire and promptly save lives.

Effects according to the embodiments are not limited by the contents exemplified above, and more various effects are included in the present specification.

1 is a diagram schematically illustrating a building firefighting and fire monitoring system using artificial intelligence according to an embodiment.
2 is a diagram schematically showing an alarm of a building firefighting and fire monitoring system using artificial intelligence according to an embodiment.
3 is a diagram schematically illustrating a CCTV of a building firefighting and fire monitoring system using artificial intelligence according to an embodiment.
4 is a diagram schematically illustrating a fire detection server according to an embodiment.

Advantages and features of the present invention and methods of achieving them will become apparent with reference to the embodiments described below in detail in conjunction with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but will be implemented in a variety of different forms, and only these embodiments allow the disclosure of the present invention to be complete, and common knowledge in the technical field to which the present invention belongs It is provided to fully inform the possessor of the scope of the invention, and the present invention is only defined by the scope of the claims.

It should be understood that all flow diagrams, state transition diagrams, pseudo code, and the like may be tangibly embodied on computer-readable media and represent various processes performed by a computer or processor, whether or not a computer or processor is explicitly shown.

The functions of the various elements shown in the figures including a processor or functional blocks represented by similar concepts may be provided by the use of dedicated hardware as well as hardware having the ability to execute software in association with appropriate software. When provided by a processor, the functionality may be provided by a single dedicated processor, a single shared processor, or a plurality of separate processors, some of which may be shared. Like reference numerals refer to like elements throughout.

Hereinafter, specific embodiments will be described with reference to the accompanying drawings.

1 is a diagram schematically showing a building fire-fighting fire monitoring system using artificial intelligence according to an embodiment, and FIG. 2 is a diagram schematically showing an alarm of a building fire-fighting fire monitoring system using artificial intelligence according to an embodiment, 3 is a diagram schematically showing a CCTV of a building firefighting monitoring system using artificial intelligence according to an embodiment, and FIG. 4 is a diagram schematically showing a fire detection server according to an embodiment.

1 to 4 , the building fire and fire monitoring system 100 using artificial intelligence according to an embodiment of the present invention includes an alarm 10 installed in a space of a building, a CCTV 40, and a communication network 30 ), and a fire monitoring server 50 and a manager terminal 70 may be included.

At least one alarm 10 may be installed in each space of the building. For example, in the case of a building having a plurality of floors, at least one alarm 10 may be installed on each floor, and at least one alarm 10 may be installed at various positions on each floor. At least one alarm 10 may be installed on the ceiling of each floor. However, the present invention is not limited thereto and may be installed on the side of each floor.

In some embodiments, the alarm 10 may include a body part BP, a speaker part SPP disposed on a lower surface of the body part BP, and a cover part CV surrounding the speaker part SPP. .

In some embodiments, the body part BP may include a first module part MP1 and a second module part MP2. The first module part MP1 and the second module part MP2 may be embedded in the outer surface of the body part BP. Also, the first module part MP1 and the second module part MP2 may be disposed to be spaced apart from each other. In some embodiments, an additional module unit may be further included in addition to the first module unit MP1 and the second module unit MP2.

In some embodiments, the first module unit MP1 may include a first sensor unit SP1 , a first light emitting unit LP1 , a second sensor unit SP2 , and a second light emitting unit LP2 . The first sensor unit SP1 , the first light emitting unit LP1 , the second sensor unit SP2 , and the second light emitting unit LP2 may be sequentially disposed in the first direction. The widths of the first sensor unit SP1 , the first light emitting unit LP1 , the second sensor unit SP2 , and the second light emitting unit LP2 in the first direction may all be the same.

Each of the first sensor unit SP and the second sensor unit SP2 may include at least one of a smoke sensor, a heat sensor, and a humidity sensor.

The smoke sensor is a sensor for detecting whether smoke is generated, and it is possible to determine whether or not smoke is generated by detecting a change in transmittance of light generated due to the smoke. For example, the smoke sensor may include a light emitting element and a light receiving element, and detect a change in transmittance of light by detecting light emitted from the light emitting element through the light receiving element. However, the method of the smoke sensor detecting smoke is not limited thereto.

A thermal sensor detects external heat and detects whether the ambient temperature is rising due to a fire.

The humidity sensor detects the current humidity (eg, relative humidity) and the amount of change in humidity.

The first light emitting unit LP1 and the second light emitting unit LP2 output a warning light source when a fire occurs. It may be detected by the sensor unit SP2 , and whether a fire has occurred may be transmitted by the fire monitoring server 50 .

Since the configuration of the second module part MP2 is the same as that of the first module part MP1, a redundant description thereof will be omitted.

The speaker unit SPP is a place to output a warning sound when a fire occurs, and whether a fire occurs may be detected by the first sensor unit SP1 and the second sensor unit SP2 of the alarm 10 , , whether a fire has occurred may be transmitted by the fire monitoring server 50 .

The cover part CV has a rim shape disposed on the lower surface of the body part BP, and may be disposed to surround the outside of the speaker part SPP. The cover part CV may include a first part P1 and a second part P2 , and a plurality of the first parts P1 and the second parts P2 may be alternately disposed in the first direction. . The first portion P1 may be formed of a phase change material such as tedradecane, octadecane, nonadecane, calcium chloride, or the like. The second part P2 may be made of a metal material. An extension length of the second portion P2 in the first direction may be longer than an extension length of the first portion P1 . Also, the area of the second part P2 may be larger than that of the first part P1 .

In some embodiments, the alarm 10 may further include a power module, an extension terminal module, and a communication module control circuit.

The power module may include a built-in battery and an external power adapter.

The extension terminal module may include an external extension terminal and an external output terminal. The alarm 10 may exchange data with other external devices through the extension terminal module by wire or wirelessly.

The external extension terminal is a module for receiving information from other external components (such as a first sensor module and a second sensor module disposed in the CCTV).

The external output terminal is a module for outputting a relay signal for controlling other external devices, and specifically, the relay signal may be transmitted to other devices in a wired manner. Accordingly, even devices without a separate wireless communication module may be controlled by receiving a relay signal from the user through the alarm 10 .

The communication module may be a module for communicating with the fire monitoring server 50 in a wireless manner. The communication module may be a module using a frequency of about 900 MHz, Wi-Fi, or Bluetooth. However, the present invention is not limited thereto, and a module that allows the alarm 10 to communicate with the fire monitoring server 50 wirelessly is sufficient.

At least one CCTV 40 may be installed in each space of the building. For example, in the case of a building composed of a plurality of floors, at least one CCTV 40 may be installed on each floor, and at least one CCTV 40 may be installed at various locations on each floor. At least one CCTV 40 may be installed on the ceiling of each floor. However, the present invention is not limited thereto and may be installed on the side of each floor.

At least one CCTV (40) is connected to a fixed part (FP) for fixing the CCTV (40) in contact with a wall, an extension part (PP) connected to the fixing part (FP), an extension part (PP) and an inner housing part It may include an outer housing part 43 capable of accommodating 44 , an inner housing part 44 accommodated in the outer housing part 43 , and including a photographing unit 41 and an auxiliary photographing unit 42 . .

In some embodiments, the fixing unit FP may be fixed to a wall or the like for fixing the CCTV 40 . The fixing part FP may be made of a metal material and an air passage may be formed therein.

In some embodiments, the buffer material M1 may be filled in the air passage of the fixing part FP. As the buffer material M1, various materials such as silicone, rubber, and sponge may be used, and the present invention is not limited to the material prepared above, and various materials for shock absorption may be provided. In this way, when the buffer material M1 is disposed in the air passage of the fixing part FP, even if an impact is transmitted to a building such as an earthquake, vibration or shock transmitted to the fixing part FP may be attenuated.

In some embodiments, the air passage of the fixing part FP may be filled with the phase change material M2. Such a phase change material M2 can prevent deformation of the fixing part FP according to a change in temperature, and can even obtain an effect of thermal insulation by blocking heat. The phase change material M2 includes tedradecane, octadecane, nonadecane, calcium chloride, and the like, and may be filled in the air passage using other materials according to the user's needs.

In some embodiments, a plurality of air passages of the fixing part FP may be disposed, and the phase change material M2 and the buffer material M1 may be alternately filled in the plurality of air passages. For example, the fixing part FP may include a plurality of first fixing part air passages and second fixing part air passages, the inside of the first fixing part air passages being filled with silicon, and the second fixing part air passages being filled with silicon. The inside of the furnace is filled with tedradecane, and the first fixing part air passage and the second fixing part air passage may be alternately arranged.

In some embodiments, the air passage of the fixing part FP may have a cylindrical shape, and may have a diameter of 0.05 cm to 0.4 cm.

In some embodiments. The extension part PP connects the fixing part FP and the external housing part 43, but may be configured to have a variable length. For example, the length of the extension part PP is increased in order to more precisely capture the timing of a fire, so that the outer housing part 43 may be detachably moved from the fixing part FP. In addition, the length of the extension part PP may be reduced, so that the housing part 43 may be attached to the fixing part FP. The extended length of the extension part PP may be 20 cm to 50 cm for stability of coupling between the external housing part 43 and the fixing part FP. In some embodiments, the extension part PP may have a foldable structure, but is not limited thereto, and may have various configurations such as an elastic member having a variable length, a solenoid, and the like. In addition, a motor may be disposed in the outer housing part 43 to vary the length of the extension part PP.

In some embodiments, the outer housing part 43 may be made of a metal material and an air passage may be formed therein.

In some embodiments, the air passage of the outer housing part 43 may be filled with a buffer material M1. As the buffer material M1, various materials such as silicone, rubber, and sponge may be used, and the present invention is not limited to the material prepared above, and various materials for shock absorption may be provided. In this way, when the buffer material M1 is disposed in the air passage of the external housing part 43 , even when an impact is transmitted to a building such as an earthquake, vibration or shock transmitted to the external housing part 43 can be attenuated.

In some embodiments, the air passage of the outer housing part 43 may be filled with the phase change material M2. Such a phase change material M2 can prevent deformation of the fixing part FP according to a change in temperature, and can even obtain an effect of thermal insulation by blocking heat. The phase change material M2 includes tedradecane, octadecane, nonadecane, calcium chloride, and the like, and may be filled in the air passage using other materials according to the user's needs.

In some embodiments, a plurality of air passages of the outer housing part 43 may be disposed, and the phase change material M2 and the buffer material M1 may be alternately filled in the plurality of air passages. For example, the outer housing part 43 is made of a metal material, and includes a plurality of first external housing part air passages and a second external housing part air passageway therein. is filled, and tedradecane may be filled in the air passage of the second external housing unit.

In some embodiments, the air passage of the outer housing part 43 may have a cylindrical shape, and may have a diameter of 0.1 cm to 0.8 cm.

In some embodiments, the outer housing part 43 may have a circular rim shape, and the inner housing part 44 may be disposed inside the outer housing part 43 .

In some embodiments, a first sensor module SM1 and a second sensor module SM2 may be disposed on a lower surface of the outer housing part 43 . The first sensor module SM1 may include a thermal sensor, and , the second sensor module (SM2) is a sensor for detecting the location of a person in a building, and a voice recognition sensor that recognizes urgent voice frequencies such as screams and shouts by embedding an Arduino voice recognition module, and people in the building It may include a GPS sensor for obtaining current location information by analyzing a GPS signal generated from a smart device of

In some embodiments, an outer surface of each of the first sensor module SM1 and the second sensor module SM2 may be surrounded by a heat conduction blocking layer HBL. Here, the heat conduction blocking film (HBL) may be a composition in which glass fire, phenol resin, and calcium carbonate are mixed. For example, the heat conduction blocking film (HBL) may include 75 wt% of glass fiber, 17 wt% of a phenolic resin, and 8 wt% of calcium carbonate, and the thickness of the heat conduction barrier film (HBL) may be 0.1 cm to 0.3 cm. . The heat conduction blocking layer may include a plurality of holes, and the diameters of the holes may be 0.01 cm to 0.03 cm. Such a hole in the heat conduction shield prevents the functions of the first sensor module SM1 and the second sensor module SM2 from being deteriorated by the heat conduction shield.

In some embodiments, the inner housing 44 may include a photographing unit 41 and an auxiliary photographing unit 42 . Here, each of the photographing unit 41 and the auxiliary photographing unit 42 includes a camera unit, and the angle of the photographing unit 41 may be adjusted in the vertical direction. In addition, the inner housing unit 44 can rotate 360 degrees to the left or right, and accordingly, the photographing unit 41 and the auxiliary photographing unit 42 disposed in the inner housing can also rotate 360 degrees to the left or right. This can effectively reduce blind spots. The auxiliary photographing unit 42 supplements the blind spot of the photographing unit 41 and may be located at 45 degrees to the right of the photographing unit 41 . In addition, the auxiliary photographing unit 42 makes it possible to photograph a wider range of images in case of fire. In some embodiments, a plurality of auxiliary photographing units 42 may be disposed. For example, the auxiliary photographing unit 42 may be located at various positions such as 45 degrees to the right, 90 degrees to the right, and 180 degrees to the right of the photographing unit 41 . This is to prepare for a case where an error occurs in the rotation of the inner housing part 44 due to the occurrence of a fire.

The inner housing portion 44 is protected by the cover portion 45 , and the cover portion 45 may be coupled to the outer housing portion 43 while covering the inner housing portion 44 .

The at least one alarm 10 and the at least one CCTV 40 may communicate with the fire monitoring server 50 through a network such as the communication network 30 .

At least one CCTV 40 may determine whether fire or smoke is detected by image processing the captured image. Such a process may be performed through an image processing unit (not shown) provided in the at least one CCTV 40 . As an example, the at least one CCTV 40 may analyze the captured image in real time by using OpenCV (Open Source Computer Vision), which is a library for real-time image processing.

In some embodiments, at least one CCTV 40 transmits the captured image to the fire monitoring server 50 using the communication network 30 when at least one of flame and smoke is detected through real-time image processing in the captured image. can be transmitted

As an example, the at least one CCTV 40 may transmit a real-time image to the fire monitoring server 50 every predetermined frame when at least one of flame and smoke is detected in the captured image.

The fire monitoring server 50 may receive the captured image from at least one CCTV 40 . The fact that at least one CCTV 40 transmits a real-time image to the fire monitoring server 50 means that flame or smoke is primarily detected in the real-time image. The fire monitoring server 50 may re-determine whether the image transmitted from the at least one CCTV 40 is a fire using deep learning-based artificial intelligence.

The fire monitoring server 50 may transmit a fire notification to the manager terminal 70 when it is determined that a fire has occurred as a result of determining whether a fire has occurred using deep learning-based artificial intelligence. The fire monitoring server 50 may guide an evacuation route when receiving fire situation confirmation information from the manager terminal 70 that has received a fire notification. Here, the fire situation confirmation information may be information transmitted to the fire monitoring server 50 after the administrator receives the fire notification through the manager terminal 70 and confirms that a fire occurs as a result of directly checking it. The fire monitoring server 50 may guide the evacuation route according to the location of the fire and the congestion rate analyzed from the images captured by the at least one CCTV 40 .

The fire monitoring server 50 may transmit a fire notification to the microcontroller of the CCTV 40 when it is determined that a fire has occurred as a result of determining whether a fire has occurred using deep learning-based artificial intelligence. The microcontroller of the CCTV 40 that has received the fire notification from the fire monitoring server 50 may operate the first sensor module SM1 and the second sensor module SM2, and the first sensor module SM1 The second sensor module SM2 detects the location of a person in the building while controlling the photographing unit 41 and the auxiliary photographing unit 42 by detecting the direction in which the 42) can be controlled. The location of the fire, the presence of people in the building, and location information obtained through the control of the photographing unit 41 and the auxiliary photographing unit 42 are transmitted to the fire monitoring server 50, and the fire monitoring server 50 is an administrator It is possible to transmit it to the terminal 70 .

When the fire detection server 50 receives the fire false detection information from the manager terminal 70 that has received the fire notification, the fire monitoring server 50 may recognize that the fire determination is incorrect and feed it back to the deep learning-based artificial intelligence. Here, the false fire detection information may be information that is transmitted to the fire monitoring server 50 after the manager receives the fire notification through the manager terminal 700 and confirms that it is not a fire as a result of directly checking it.

The fire monitoring server 50 may automatically guide an evacuation route when it does not receive fire situation confirmation information or fire false detection information within a predetermined time from the manager terminal 70 that has received the fire notification. In other words, if the manager is absent or fails to check the fire notification, the evacuation route guidance cannot be delayed. can guide you.

When it is determined that a fire has occurred as a result of determining whether a fire has occurred using the deep learning-based artificial intelligence of the fire monitoring server 50, a fire notification can be transmitted to the alarm 10, and the alarm 10 transmits a light emitting control signal. The first light emitting unit LP1 and the second light emitting unit LP2 may be operated to output a warning light source. In addition, the alarm 10 may output a warning sound by operating the speaker unit SPP through the speaker control signal.

The communication network 30 is a communication network in which at least one alarm 10, at least one CCTV 40, the fire monitoring server 50 and the manager terminal 70 can communicate with each other. It can be configured regardless of whether it is a short-range network (PAN), a local area network (LAN), a metropolitan area network (MAN), and a wide area network (WAN). can be configured. In addition, the communication network 30 may be a known World Wide Web (WWW), and may use a wireless transmission technology used for short-range communication such as infrared (IrDA) or Bluetooth (Bluetooth). have. Since the communication network 30 is widely known to those skilled in the art, a detailed description thereof will be omitted.

The manager terminal 70 is a terminal capable of receiving a fire notification from the fire monitoring server 50, and may be various types of terminals or electronic equipment capable of data communication, such as a user computer or a mobile terminal, in a wired or wireless manner. It may be variously changed within a range known to those skilled in the art. The manager terminal 70 may be a user computer installed in a central control room or the like, or may be a mobile terminal directly owned by the manager. The manager may receive a fire notification through the manager terminal 70 , and when it is determined as an actual fire situation, it may transmit fire situation confirmation information to the fire monitoring server 50 . When the manager receives a fire notification through the manager terminal 70 but determines that it is not a fire, the manager may transmit false fire detection information to the fire monitoring server 50 .

Referring to FIG. 4 , the fire monitoring server 50 using AI based on image processing and deep learning according to an embodiment of the present invention includes a communication unit 51 , a fire determination unit 52 and a fire notification unit 53 . , an evacuation route guide unit 54 , a person presence determination unit 55 , a person location notification unit 56 , and a database unit 57 .

The communication unit 51 may receive a captured image from at least one alarm 10 and at least one CCTV 40 installed in each space of a building. In addition, the communication unit 51 may transmit the fire notification received from the fire notification unit 53 to the manager terminal 70 , and transmit the fire situation confirmation information and fire false detection information transmitted from the manager terminal 70 to the fire determination unit. (52) and can be transmitted to the evacuation route guide (54). The server communication unit 51 can communicate with at least one alarm 10, at least one CCTV 40, and the manager terminal 70 through the communication network 30, and is implemented as a communication module supporting various wired and wireless communication methods. and may be variously changed within a range known to those skilled in the art.

In addition, the communication unit 51 may receive the location of the fire, the presence of a person in the building, and location information from at least one CCTV 40 . The communication unit 51 may transmit the information received from the person presence determination unit 55 and the person location notification unit 56 in the building to the manager terminal 70 .

The fire determination unit 52 may receive a real-time image in which flame or smoke is primarily detected from at least one CCTV 40, and use the image transmitted from at least one CCTV 40 to perform deep learning-based artificial intelligence. It can be used to judge whether there is a fire or not.

On the other hand, according to an embodiment, the fire determination unit 52 first determines whether fire or smoke is detected by image processing the photographed image, and when it is determined that fire or smoke is detected as a result of the first determination, the It is possible to secondarily determine whether a fire exists using the captured video using deep learning-based artificial intelligence. In this case, the at least one CCTV 40 may transmit the captured image at regular frame intervals to the fire monitoring server 50 without performing real-time image processing for detecting flame or smoke in the captured image. That is, according to the embodiment, the primary determination of detecting flame or smoke in the captured image may be performed by at least one CCTV 40, or the fire determination unit 52 of the fire monitoring server 50 may perform the first determination. may be

The deep learning-based artificial intelligence performed by the fire determination unit 52 may operate in the following way.

Image information captured by at least one CCTV 40 installed in a building may be transmitted to the fire monitoring server 50, and the fire monitoring server 50 may store the received CCTV image in the database unit 57. .

In this way, the normal CCTV image accumulated in the database unit 57 may be a standard data value. The normal CCTV images may be classified and stored according to each of at least one CCTV 40 .

When fire or smoke is detected in the image captured by at least one CCTV 40 according to the result of the first real-time image processing, the fire determination unit 52 performs the deep learning-based artificial intelligence through the previously learned data. It can be determined once more whether the object of the captured image is definitely fire or smoke. As an example, TensorFlow may be used as the deep learning-based artificial intelligence engine, and an Inception v3 model may be used. To achieve the goal of early fire detection, supervised learning, which is a method of learning a machine learning model using training data that knows the correct answer, can be used, and the structure, weight, parameter, and data set of the artificial intelligence model can be set in various ways. can The image transmitted from at least one CCTV 40 may be stored in the database unit 57 for each predetermined frame, and learning may be performed by various source codes.

The fire determination unit 52 compares the CCTV image, which is primarily determined to have detected fire or smoke, with the ordinary CCTV image accumulated in the database unit 57, so that the similarity with the ordinary CCTV image is significantly different. In this case, it can be considered a fire. As an example, if the similarity between the CCTV image, which is primarily determined to have detected fire or smoke, and the usual CCTV image, is more than 90%, the primary determination result is judged to be malfunctioning, and feedback is given and a fire notification may not be generated. have. In addition, when the similarity between the CCTV image, which is primarily determined to have detected fire or smoke, and the ordinary CCTV image, is less than 90%, it is determined that the fire is a fire and the fire evacuation system can be operated. That is, the fire determination unit 52 may set a predetermined similarity reference value, and compare the similarity between the CCTV image determined primarily that fire or smoke is detected and the usual CCTV image with the similarity reference value to determine the primary fire. You can judge whether the results are correct or not. The predetermined similarity reference value may be variously changed according to each building, each zone, and each CCTV 40 .

Since the fire determination unit 52 uses deep learning-based artificial intelligence, the usual CCTV images of at least one CCTV 40 are continuously accumulated in the database unit 57 so that a lot of training data is provided and the fire determination is made continuously. It has the effect of increasing the accuracy of fire judgment. Therefore, the more the deep learning-based artificial intelligence of the fire determination unit 52 learns, the higher the accuracy and reliability of the recognition of the fire situation in the building, so in the end, an unmanned early fire detection without a separate manager and evacuation route guidance may be possible.

Meanwhile, according to an embodiment, the deep learning-based artificial intelligence performed by the fire determination unit 52 may operate as a method of learning a flame or smoke image. That is, the database unit 57 can store a large amount of flame images and smoke images, and the deep learning-based artificial intelligence of the fire determination unit 52 uses the flame images and smoke images stored in the database unit 57 . It can improve the ability to recognize flames and smoke. The fire determination unit 52 may secondarily determine whether flames or smoke exist in the CCTV image by analyzing the CCTV image, which is primarily determined to have detected fire or smoke, according to deep learning-based artificial intelligence. have.

When the fire determination unit 52 finally determines that a fire has occurred, the fire notification unit 53 may transmit a fire notification to the manager terminal 70 through the communication unit 51 . The fire notification may be transmitted to the manager terminal 70 in the form of a text message or push message, and the location of the fire suspicious area, the video of the CCTV in which the fire is detected, etc. may be transmitted to the manager terminal 70 together. . In addition, the fire notification unit 53 may transmit a fire notification to the microcontroller of the CCTV 40 through the communication unit 51 when the fire determination unit 52 finally determines that a fire has occurred, and the CCTV 40 of the microcontroller may operate the first sensor module SM1 and the second sensor module SM2. The first sensor module SM1 detects the direction in which the fire occurred and controls the photographing unit 41 and the auxiliary photographing unit 42 , while the second sensor module SM2 detects the location of a person in the building and shoots it It is possible to control the unit 41 and the auxiliary photographing unit 42 .

The manager terminal 70 may notify the manager of the fire notification reception through a notification sound or the like. When the manager receives a fire notification, the manager can directly check the fire situation once again to see if the fire situation is correct. can make it happen

The evacuation route guide unit 54 may guide the evacuation route according to the location of the fire and the congestion rate analyzed from the image captured by the at least one CCTV 40 . The congestion rate can be calculated according to how densely people are and the ease of movement by analyzing CCTV images.

The evacuation route guide unit 54 may guide the evacuation route when receiving fire status confirmation information from the manager terminal 70 that has received the fire notification.

The evacuation route guide unit 54 may automatically guide the evacuation route when the fire situation confirmation information or fire false detection information is not received within a predetermined time from the manager terminal 70 that has received the fire notification. That is, if the notification cannot be confirmed due to the absence of an administrator, etc., the evacuation route guide unit 54 may automatically guide the evacuation route after a specific time.

The fire determination unit 52, when receiving the fire false detection information from the manager terminal 70 that has received the fire notification, recognizes that the fire determination by the fire determination unit 52 is wrong, and is applied to the deep learning-based artificial intelligence. can give feedback The deep learning-based artificial intelligence of the fire determination unit 52 may recognize that the CCTV image from which the false fire detection information was received was not a fire, learn anew, and reflect it in the next fire determination.

According to an embodiment, in a state in which there is no response from the manager terminal 70 that has received the fire notification, the fire determination unit 52 ) may determine that a fire has occurred and guide the evacuation route through the evacuation route guide unit 54 . That is, the fire determination unit 52 analyzes the video taken from the CCTV in the suspected fire area, and when it detects a phenomenon in which congestion increases or the movement speed of people increases due to a large number of people in the area, a fire has occurred and is urgent. It can judge the situation and guide the evacuation route. That is, the fire determination unit 52 according to an embodiment of the present invention may guide the evacuation route by analyzing the congestion level and the movement speed of people in the suspected fire area analyzed through CCTV.

The presence/absence determining unit 55 in the building may receive a real-time image detected from at least one CCTV 40 , and use the image transmitted from at least one CCTV 40 to construct a building using deep learning-based artificial intelligence. I can judge whether I am a person or not. For example, it is estimated that there is a person through a voice recognition sensor that recognizes urgent voice frequencies such as screams and shouts, and a GPS sensor that acquires current location information by analyzing GPS signals generated from people's smart devices in buildings. It is possible to receive the CCTV (40) image of the place, and judge the presence or absence of a person in the building using the image based on deep learning artificial intelligence.

On the other hand, according to the embodiment, the person presence determination unit 55 in the building first determines whether there is a person by image processing the captured image, and when it is determined that a person in the building is detected as a result of the first determination, The position of a person may be secondarily determined by using the artificial intelligence based on deep learning on the captured image. In this case, the at least one CCTV 40 may transmit the captured image at regular frame intervals to the fire monitoring server 50 without performing real-time image processing for detecting a person in the captured image. That is, according to the embodiment, the first determination of detecting a person in the photographed image may be performed by at least one CCTV 40 , and performed by the person presence determination unit 55 in the building of the fire monitoring server 50 . You may.

The deep learning-based artificial intelligence performed by the person presence determination unit 55 in the building may operate in the following way.

The CCTV 40, which has received the fire notification from the fire monitoring server 50, operates the first sensor module SM1 and the second sensor module SM2 to detect the location of the fire and the presence or absence of a person in the building, and the photographing unit (41) and the auxiliary photographing unit 42 may be controlled to photograph the detected position. Image information captured by at least one CCTV 40 installed in a building may be transmitted to the fire monitoring server 50, and the fire monitoring server 50 may store the received CCTV image in the database unit 57. .

In this way, the normal CCTV image accumulated in the database unit 57 may be a standard data value. The normal CCTV images may be classified and stored according to each of at least one CCTV 40 .

When it is detected as a person in the image captured by at least one CCTV 40 according to the result of real-time image processing, the fire determination unit 52 performs deep learning-based artificial intelligence (AI) of the captured image through pre-learned data. It can be determined once more that the object is definitely a person. As an example, TensorFlow may be used as the deep learning-based artificial intelligence engine, and an Inception v3 model may be used. To achieve the goal of early fire detection, supervised learning, which is a method of learning a machine learning model using training data that knows the correct answer, can be used, and the structure, weight, parameter, and data set of the artificial intelligence model can be set in various ways. can The image transmitted from at least one CCTV 40 may be stored in the database unit 57 for each predetermined frame, and learning may be performed by various source codes.

The person presence determination unit 55 in the building compares the CCTV image determined to be detected as a person and the ordinary CCTV image accumulated in the database unit 57, and when the similarity with the ordinary CCTV image is significantly different It can be determined that the object is a person.

When it is determined that the object is a person, the person presence determination unit 55 in the building may calculate the location of the person. For example, the location information is calculated through the location of the CCTV 40 that has transmitted the corresponding image, and the location of the person is more precise using the GPS signal detected through the second sensor module SM2 of the CCTV 40 . can be calculated.

The person location notification unit 56 may transmit the location of the person to the manager terminal 70 through the communication unit 51 when the person presence determination unit 55 detects the person and determines the location. The person location notification may be transmitted to the manager terminal 70 in the form of a text message or a push message, etc., and the location of the person as well as the CCTV image of the fire area may be transmitted to the manager terminal 70 together. have.

The database unit 57 may store the captured images received from at least one CCTV 40 , and may also store information on user terminals owned by people in the building for evacuation route guidance.

Although the embodiments of the present invention have been described above with reference to the accompanying drawings, those of ordinary skill in the art to which the present invention pertains may be embodied in other specific forms without changing the technical spirit or essential features of the present invention. you will be able to understand Therefore, it should be understood that the embodiments described above are illustrative in all respects and not restrictive.

10: alarm
40: CCTV
30: communication network
50: fire monitoring server
70: manager terminal
100: Building firefighting and fire monitoring system using artificial intelligence

Claims (2)

In the building firefighting monitoring system using artificial intelligence including at least one alarm, at least one CCTV and a fire monitoring server,
The at least one alarm includes a body part, a speaker part disposed on a lower surface of the body part, and a cover part surrounding the speaker part,
The body part includes a first module part and a second module part spaced apart from each other, and the first module part and the second module part are embedded in an outer surface of the body part,
Each of the first module unit and the second module unit includes a first sensor unit, a first light emitting unit, a second sensor unit, and a second light emitting unit sequentially arranged in a first direction,
Each of the first sensor unit and the second sensor unit includes at least one of a smoke sensor, a heat sensor, and a humidity sensor,
The cover part includes a first part and a second part, the first part and the second part are alternately arranged in a plurality in the first direction, and the first part is made of tedradecane, the second part is made of metal,
The at least one CCTV includes a fixing part for fixing the CCTV, an extension part connected to the fixing part, an external housing part connected to the extension part, and an internal part accommodated in the external housing part and including a shooting part and an auxiliary shooting part including a housing,
The fixing part includes a plurality of first fixing part air passages and second fixing part air passages, the inside of the first fixing part air passage is filled with silicon, and the inside of the second fixing part air passage is filled with tedradecane. , the first fixing part air passage and the second fixing part air passage are alternately arranged with each other, and each of the first fixing part air passage and the second fixing part air passage has a cylindrical shape, and the first The diameter of each of the fixing part air passage and the second fixing part air passage is 0.05 cm to 0.4 cm,
The extension part connects the fixing part and the external housing part, the length of which is variable,
The outer housing part is made of a metal material, and includes a plurality of first external housing part air passages and second external housing part air passages therein, the inside of the first external housing part air passage is filled with silicon, and the second external housing part air passage is filled with silicon. The air passage of the housing part is filled with tedradecane, and each of the first external housing part air passage and the second external housing part air passage has a cylindrical shape, and the first external housing part air passage and the second external housing part Each air passage has a diameter of 0.1 cm to 0.8 cm,
A first sensor module and a second sensor module are disposed on a lower surface of the outer housing part, the first sensor module includes a heat sensor, and the second sensor module includes a voice recognition sensor and GPS for recognizing a human voice frequency. It includes a GPS sensor that analyzes the signal to obtain location information,
Each of the first sensor module and the second sensor module has a heat conduction blocking film disposed on an outer surface thereof, and the heat conduction barrier film comprises 75% by weight of glass fiber, 17% by weight of phenolic resin, and 8% by weight of calcium carbonate, and the heat conduction The thickness of the barrier film is 0.1 cm to 0.3 cm, the heat conduction barrier film includes a plurality of holes, and the diameter of the plurality of holes is 0.01 cm to 0.03 cm.
The method of claim 1,
The fire monitoring server,
a communication unit receiving the captured image from the at least one CCTV;
a fire determination unit that judges whether the photographed image is a fire using deep learning-based artificial intelligence;
a fire notification unit configured to transmit a fire notification to the manager terminal and the at least one CCTV through the communication unit when the fire determination unit determines that a fire has occurred;
an evacuation route guide for guiding an evacuation route when receiving fire status confirmation information from the manager terminal that has received the fire notification;
a person presence determination unit in the building that analyzes the CCTV image of a place where a person is detected through the second sensor module to determine the presence or absence of a person in the building and the location of the person; and
A person location notification unit for transmitting the presence or absence of a person in the building and the location of the person to the manager terminal,
The at least one CCTV further comprises a microcontroller,
When the first sensor module detects heat exceeding a set value, the first sensor module transmits a first control signal for moving the photographing unit in a direction in which heat exceeding the set value is detected to the microcontroller,
When a person is detected, the second sensor module transmits a second control signal for moving the photographing unit in a direction in which the person is detected to the microcontroller.

Images