Fire alarm system with false alarm prevention

Abstract

The present invention relates to a fire monitoring system equipped with a false alarm prevention function, and more specifically, to a fire monitoring system equipped with a false alarm prevention function that can detect the occurrence of a fire by analyzing images captured by a camera to determine whether smoke or flames are generated, and can prevent false alarms by allowing an administrator to directly check the images.

Description

Technology Field

[0001] The present invention relates to a fire monitoring system equipped with a false alarm prevention function, and more specifically, to a fire monitoring system equipped with a false alarm prevention function that can detect the occurrence of a fire by analyzing images captured by a camera to determine whether smoke or flames are generated, and can prevent false alarms by allowing an administrator to directly check the images. Background Technology

[0003] With industrial development, the types of fires are diversifying and the scale of damage is increasing, making fire prediction and monitoring increasingly difficult.

[0004] Accordingly, although various types of fire detection devices are being installed, false alarms occur frequently due to the aging of the devices themselves or a lack of adaptability to the surrounding environment, and as a result, social and economic losses, such as unnecessary fire dispatches, are steadily increasing.

[0005] In addition, as false alarms occur frequently, people do not evacuate because they assume that a fire alarm is a common false alarm and end up suffering an accident, or managers mistakenly turn off the power to the fire control system because they think it is a trivial false alarm.

[0006] In 2022, when an LP gas fire occurred in the underground parking lot of an apartment in Buldang-dong, Cheonan, South Chungcheong Province, caused by an employee of a mobile car wash company, an apartment management office employee made the mistake of cutting off the power to the entire fire safety system after judging the fire alarm to be a false alarm, which had occurred frequently in the past, resulting in massive property damage and a situation where human lives were nearly at risk.

[0007] According to a report by the Jeonbuk Fire Department, there were 8,405 non-fire alarms, known as alarm malfunctions, over the past three years (2020–2022). Broken down by year, there were 2,043 cases in 2020, 3,078 in 2021, and 3,284 in 2022. In particular, the most significant issue was that 2,078 cases (63.3%) of the malfunctions occurred at 425 locations (31.9%) out of a total of 1,331, where the alarms occurred three or more times. It was found that administrative resources are being wasted due to dispatching more than 10 times a day because of alarm malfunctions, raising concerns about potential delays in responding to actual fires.

[0008] Therefore, it became necessary to develop a fire monitoring system equipped with a function to prevent false alarms. Prior art literature

[0010] Korean Registered Patent No. 10-2487481 (January 6, 2023) The problem to be solved

[0011] The present invention was devised to solve the aforementioned problem and aims to provide a fire monitoring system equipped with a false alarm prevention function that can prevent false alarms by analyzing images captured by a camera to detect the occurrence of a fire based on whether smoke or flames are generated, and by allowing an administrator to directly verify the images.

[0012] In addition, the present invention aims to provide a fire monitoring system equipped with a non-fire alarm prevention function that can protect privacy by applying mosaic processing to a person's face or entire body when monitoring camera images.

[0013] In addition, the present invention aims to provide a fire monitoring system equipped with a false alarm prevention function that allows for more accurate verification of the fire situation by not applying mosaic processing to the monitored video when it is determined that a fire has occurred.

[0014] In addition, the present invention aims to provide a fire monitoring system equipped with a false alarm prevention function that can prevent false alarms by more accurately determining whether an actual fire has occurred and thereby preventing false alarms, as the system is equipped with an integrated sensor module capable of detecting heat, gas, smoke, etc. and can detect the occurrence of a fire from the measurement values ​​of the integrated sensor module.

[0015] In addition, the present invention aims to provide a fire monitoring system equipped with a false alarm prevention function that prevents a manager from arbitrarily deactivating a fire alarm due to human error when fire detection by camera imagery and fire detection by an integrated sensor module overlap. means of solving the problem

[0017] The present invention for achieving the above objective comprises: a camera that photographs within a surveillance area; an image analysis unit that receives images captured by the camera in real time and determines whether smoke or flames have occurred within the images; an alarm issuing unit that issues a fire alarm when it is determined by the image analysis unit that smoke or flames have occurred; and a monitoring unit that receives the images from the image analysis unit and displays them visibly, and receives and displays the smoke or flame occurrence determination result from the image analysis unit, or displays the result when a fire alarm is issued by the alarm issuing unit. The image analysis unit is characterized by acquiring a pixel change pattern from the images and determining whether smoke or flames have occurred based on whether the acquired pixel change pattern matches a previously learned movement pattern of smoke or flames.

[0018] In addition, the image analysis unit of the present invention is characterized by learning the change pattern of pixels in an image in which actual smoke or flames have occurred using an AI model.

[0019] In addition, the image analysis unit of the present invention is characterized by learning the change pattern of pixels in an image where actual smoke or flames occur for each item of environmental variables including two or more of the width, length, height, wind direction, and wind speed of the space where smoke or flames occur.

[0020] In addition, the image analysis unit of the present invention is characterized by being able to infer the pixel change pattern in an unlearned environmental variable using the pixel change pattern learned for each environmental variable.

[0021] In addition, the image analysis unit of the present invention is characterized by, when providing the image to the monitoring unit, acquiring a pixel change pattern from the image, determining the presence or absence of a person by whether the acquired pixel change pattern matches a previously learned human movement pattern, and mosaic processing the pixel area having the pixel change pattern determined to be where a person is present.

[0022] In addition, the image analysis unit of the present invention is characterized by not performing mosaic processing on the image provided to the monitoring unit when it is determined that smoke or flames have occurred as a result of analyzing the image.

[0023] In addition, the alarm issuing unit of the present invention is characterized by issuing a first fire alarm when information is received from the image analysis unit that smoke or flame has been generated, and releasing the issuance of the first fire alarm when an alarm release signal is received from the monitoring unit.

[0024] In addition, the present invention includes an integrated sensor module that provides a measurement value to the alarm issuing unit by including one or more of the heat detection sensor, gas detection sensor, and smoke detection sensor; wherein the alarm issuing unit issues a second fire alarm if a measurement value greater than a set value is generated from the integrated sensor module within a set time after the issuance of a first fire alarm, and the second fire alarm cannot be deactivated by the operation of the monitoring unit.

[0025] In addition, the present invention is characterized in that an alarm release means is provided on one side of the housing of the integrated sensor module, and a secondary fire alarm can be released by operating the alarm release means.

[0026] In addition, the camera of the present invention is rotatably mounted on one side of the integrated sensor module, and the alarm issuing unit is characterized by rotating the camera when a result determining that smoke or flame has occurred is not received from the image analysis unit even though a measurement value greater than a set value has been generated from the integrated sensor module, and issuing a secondary fire alarm when a result determining that smoke or flame has occurred is received from the image analysis unit after the camera has been rotated. Effects of the invention

[0028] The present invention can detect the occurrence of a fire by analyzing images captured by a camera to determine whether smoke or flames are present, and since the manager can directly verify the images, it has the effect of preventing false alarms.

[0029] In addition, the present invention has the effect of protecting privacy by applying mosaic processing to a person's face or entire body when monitoring camera images.

[0030] In addition, the present invention has the effect of allowing for more accurate verification of the fire situation by not applying mosaic processing to the monitored video when it is determined that a fire has occurred.

[0031] In addition, the present invention is equipped with an integrated sensor module capable of detecting heat, gas, smoke, etc., and since the occurrence of a fire can be detected from the measurement values ​​of the integrated sensor module, it is possible to determine whether there is an actual fire more accurately, and thereby has the effect of preventing false alarms.

[0032] In addition, the present invention has the effect of preventing the deactivation of a fire alarm due to human error by ensuring that an administrator cannot arbitrarily deactivate the fire alarm when the detection of a fire by camera image and the detection of a fire by an integrated sensor module overlap. Brief explanation of the drawing

[0034] FIG. 1 is a configuration diagram of a fire monitoring system equipped with a non-fire alarm prevention function according to an embodiment of the present invention. FIG. 2 is a perspective view of an integrated sensor module of a fire monitoring system equipped with a non-fire alarm prevention function according to an embodiment of the present invention. FIG. 3 is a diagram showing the usage state of a fire monitoring system equipped with a non-fire alarm prevention function according to an embodiment of the present invention. FIG. 4 is a diagram showing the usage state of a fire monitoring system equipped with a non-fire alarm prevention function according to another embodiment of the present invention. FIG. 5 is a usage state diagram illustrating the change in the camera monitoring area of ​​FIG. 4. Specific details for implementing the invention

[0035] Hereinafter, embodiments of the present invention will be described in detail with reference to the attached drawings.

[0037] The present invention comprises, as illustrated in FIGS. 1 to 5, a camera (100) that photographs within a surveillance area; an image analysis unit (200) that receives images captured by the camera (100) in real time and determines whether smoke or flames have occurred in the images; an alarm issuing unit (300) that issues a fire alarm when it is determined by the image analysis unit (200) that smoke or flames have occurred; and a monitoring unit (400) that receives images from the image analysis unit (200) and displays them visually, receives and displays the smoke or flame occurrence determination results from the image analysis unit (200), or displays the fire alarm when it is issued by the alarm issuing unit (300).

[0039] The camera (100) serves to photograph the area within the surveillance zone and provide the captured image to the image analysis unit (200). This camera (100) is installed in the housing of the integrated sensor module (500) and can be mounted on a wall or ceiling, etc., together with the integrated sensor module (500). However, it is also possible to configure the camera (100) separately from the integrated sensor module (500).

[0040] When the camera (100) is installed in the housing of the integrated sensor module (500), the camera (100) can be installed in a rotatable form. At this time, the camera (100) is connected to an electric rotation device and configured to rotate by a control signal from the alarm issuing unit (300). The meaning of the camera (100) rotating is that the surveillance area being filmed is changed, and for example, while filming one end of a hallway, if it is rotated 180 degrees, it is configured to be able to film the other end of the hallway.

[0042] The integrated sensor module (500) is configured to include one or more of a heat detection sensor, a gas detection sensor, and a smoke detection sensor, and when a measurement value is generated from these sensors, it is provided to the alarm issuing unit (300).

[0043] The integrated sensor module (500) is configured to include a housing and a heat detection sensor, a gas detection sensor, a smoke detection sensor, etc., installed in the housing, and a communication unit is further configured to provide the measurement value to the alarm issuing unit (300). It is preferable that the integrated sensor module (500) be provided at multiple locations within the building, and if configured to transmit images using wireless communication, a wireless signal relay device may be separately configured for smooth data transmission.

[0044] The heat detection sensor, gas detection sensor, and smoke detection sensor of the integrated sensor module (500) are each configured to detect phenomena occurring during a fire. The heat detection sensor is configured to detect hot heat during a fire, and the gas detection sensor is configured to detect gases generated during a fire, such as carbon monoxide, carbon dioxide, and toxic gases. In this case, it is preferable to interpret the gas detection sensor as a set of sensors that detect each gas, rather than as a single sensor. Additionally, the smoke detection sensor is configured to directly sense the presence or absence of smoke generated during a fire.

[0045] The measurement value measured by the integrated sensor module (500) is provided to the alarm issuing unit (300) and is used as a factor to determine whether the alarm issuing unit (300) issues a second alarm.

[0047] The video analysis unit (200) receives video captured by the camera (100) in real time and determines whether smoke or flames have occurred in the video. The video analysis unit (200) may be a server, PC, laptop, or portable terminal owned by an administrator located in a building's management room or a remote control center, and any device capable of receiving video and performing calculations to determine whether smoke or flames have occurred is possible.

[0048] When an image is transmitted from the camera (100), the image analysis unit (200) compares it for each set frame to identify changes in pixels and obtain the change pattern. At this time, the image analysis unit (200) has been pre-learned information regarding how the movement pattern of smoke or flames is implemented within the image. Using this information, the image analysis unit (200) determines whether smoke or flames have occurred by determining whether the change pattern of pixels obtained from the real-time image matches the pre-learned movement pattern of smoke or flames. The determination result is then provided to the monitoring unit (400) and the alarm issuing unit (300).

[0049] The image analysis unit (200) learns the pixel change patterns of an actual image in which smoke or flames are generated using a dedicated AI model. It is desirable to learn the patterns of smoke or flames under as many different conditions as possible for such learning images, and at this time, environmental variable information of the space in which the smoke or flames are generated is also learned. The environmental variable information of the space in which the smoke or flames are generated may include two or more of the width, length, height, wind direction, and wind speed of the space in which the smoke or flames are generated. That is, the image analysis unit (200) learns the pixel change patterns of an actual image in which smoke or flames are generated for each item of the environmental variables. The AI ​​model is configured to enable the inference of pixel change patterns in unlearned environmental variables by using the pixel change patterns learned for each of these environmental variables. To simplify, for example, if a pattern is learned where smoke originates in the middle of a corridor 3 meters wide and 3 meters high, rises 2.7 meters at a 15-degree angle, spreads widely at an 85-degree angle, and is additionally learned where smoke originates in the middle of a corridor 4 meters wide and 4 meters high, rises 3.7 meters at a 15-degree angle, spreads widely at an 85-degree angle, then the principle is that when smoke originates in the middle of a corridor 4 meters wide and 3 meters high, the pattern of rising 2.7 meters at a 15-degree angle, spreading widely at an 85-degree angle, can be inferred. Of course, the spread of smoke or flames takes on a very complex pattern due to various causes such as ambient air temperature, humidity, wind direction, and wind speed; however, by continuously learning numerous videos of smoke or flame occurrences, discrimination errors can be gradually reduced, and the inference of pixel change patterns in unlearned environmental variables can also be performed with precision.

[0051] The alarm issuing unit (300) receives the determination result regarding whether smoke or flames have occurred from the image analysis unit (200), and if it is determined that smoke or flames have occurred, it issues a fire alarm. The alarm issuing unit (300) may be configured together with the image analysis unit (200) and may be a server, PC, laptop, or portable terminal owned by an administrator located in the building's management room or a remote control center. Any device capable of receiving the determination result regarding whether smoke or flames have occurred, generating an alarm signal, providing it to the monitoring unit (400), and generating and transmitting a control signal for the rotational operation of the camera (100) is possible.

[0052] The alarm issuing unit (300) issues a first fire alarm when it receives information from the video analysis unit (200) that smoke or flames have been detected. A first fire alarm is a state in which the alarm can be deactivated depending on the manager's judgment. When the first fire alarm is issued, the manager visually checks the actual video of the area where the fire alarm occurred through the monitoring unit (400), and if it is determined to be a false alarm, deactivates the alarm. Deactivation of the alarm can be input through the monitoring unit (400), and the alarm issuing unit (300) deactivates the first fire alarm when it receives an alarm deactivation signal from the monitoring unit (400).

[0053] However, there is a risk of human error occurring when a manager deactivates the alarm without properly verifying the actual video footage upon the issuance of the first fire alarm. Human error can cause irreversible disasters, as evidenced by the fire at a mobile car wash in a parking lot in Cheonan, where the manager dismissed the alarm as a common false alarm and stopped the entire fire prevention system, resulting in amplified damage.

[0054] Accordingly, the present invention is configured to enable the additional issuance of a secondary fire alarm so as to prevent false alarms while eliminating human error. The alarm issuing unit (300) issues a secondary fire alarm if a measurement value greater than a set value is generated from the integrated sensor module (500) within a set time after the issuance of the primary fire alarm. That is, after issuing the primary fire alarm by determining the occurrence of smoke or flames from the captured image of the camera (100), since there is no case where any one of the heat, harmful gas, or smoke caused by the fire is detected within a set time except when an actual fire occurs, a secondary fire alarm is issued in this case, and the secondary fire alarm is configured so that it cannot be deactivated by the operation of the monitoring unit (400). This is valid even if the alarm is deactivated by the manager when the first fire alarm is issued, and even if the first fire alarm is deactivated, if a measurement value greater than the set value is generated from the integrated sensor module (500) within a set time from the time the first fire alarm is issued, a second fire alarm is issued. This is a configuration designed to prevent human error in which the manager deactivates the alarm without properly checking the first fire alarm, and then deactivates the second fire alarm by mistaking it for a common repeated false alarm.

[0055] It is desirable that the secondary fire alarm cannot be deactivated by the monitoring unit (400) located at a remote location, and that the alarm be deactivated only after the manager has verified the site. To this end, an alarm deactivation means is provided on one side of the housing of the integrated sensor module (500), and the secondary fire alarm can be deactivated by operating the alarm deactivation means. At this time, the deactivation of the secondary alarm should be possible only at the integrated sensor module (500) that generated the measurement value for issuing the secondary alarm, so that the manager can verify the site directly.

[0056] However, depending on the direction of the wind, the image analysis unit (200) determines that smoke or flames have occurred based on the image captured by the camera (100) installed in the integrated sensor module (500), and the alarm issuing unit (300) issues a first alarm. However, there may be cases where the wind blows in a direction away from the integrated sensor module (500), so the integrated sensor module (500) does not measure a value greater than the set value, and thus the conditions for issuing a second alarm cannot be satisfied. In such cases, even though an actual fire has occurred, the second alarm is not issued, and there is a possibility that the alarm may be deactivated due to a manager's mistake.

[0057] Therefore, even in such cases, a second alarm must be issued to prevent the administrator from accidentally deactivating the alarm. To this end, the present invention is configured such that when smoke or flames are detected as a result of determining an image captured by a camera (100) installed in a certain integrated sensor module (500), and a first alarm is issued by the alarm issuing unit (300) accordingly, if a measurement value greater than a set value is generated in another integrated sensor module (500) located within a set range from the said integrated sensor module (500), the alarm issuing unit (300) is configured to issue a second alarm.

[0058] To this end, the alarm issuing unit (300) stores the unique number and installation location of the integrated sensor module (500) together with matching, and also stores the unique number and installation location of the camera (100) together with matching. At this time, if the camera (100) is installed in the integrated sensor module (500), the unique number of the integrated sensor module (500) can be borrowed. Furthermore, when the image analysis unit (200) transmits the result of determining smoke or flame occurrence, the determination result is transmitted along with the unique number to indicate which camera (100) captured the image and the result of determination. When the measurement value from the integrated sensor module (500) is transmitted to the alarm issuing unit (300), the measurement value is transmitted along with the unique number to indicate which integrated sensor module (500) the measurement value is.

[0059] Meanwhile, since the integrated sensor module (500) of the present invention is configured to be attached to a desired location, such as an indoor wall or ceiling, and to receive data via wireless communication, there are cases where the installation location of the integrated sensor module (500) cannot be known even if the unique number of the integrated sensor module (500) is known. Therefore, when installing the integrated sensor module (500), a manager accompanies the installation to record which integrated sensor module (500) with which unique number is installed at which location, and stores this in advance in the alarm issuing unit (300).

[0061] In addition, an alarm must be issued even when the above-described example occurs in reverse. For example, this is the case where no smoke or flames are generated within the monitoring range of the camera (100), but a measurement value exceeding a set value is generated from the integrated sensor module (500) where the camera (100) is installed.

[0062] For example, this is a case where a measurement value exceeding a set value is generated from a certain integrated sensor module (500), but the occurrence of smoke or flames is not detected from the image of the camera (100) installed in the integrated sensor module (500). This is a case where a fire occurs behind the monitoring area of ​​the camera (100) and toxic gases, etc., are carried by the wind into the integrated sensor module (500). In this case, the direction of the camera (100) must be rotated so that a monitoring area other than the current monitoring area is captured. To this end, the alarm issuing unit (300) is configured such that if a measurement value greater than a set value is generated from one integrated sensor module (500) but a result is not received from the image analysis unit (200) that smoke or flame has been generated, the camera (100) installed in the integrated sensor module (500) is rotated, and if a result is received from the image analysis unit (200) that smoke or flame has been generated after the camera (100) is rotated, a second fire alarm is issued immediately regardless of whether a first fire alarm is issued.

[0063] Meanwhile, when releasing the alarm through an alarm release means provided on one side of the integrated installation module, the release of the alarm is configured to be achieved only by the alarm release means of the integrated installation module that generated the measurement value or image for issuing the secondary alarm. This is intended to prevent the secondary alarm from being released without dispatching to the site for verification by operating the alarm release means of any one of the multiple integrated installation modules closest to the external entrance. In other words, when the secondary alarm is issued, the manager must dispatch to the area where the integrated installation module involved in issuing the secondary alarm is located and visually inspect the site in person.

[0064] Additionally, the alarm issuing unit (300) is configured to automatically notify the fire department of the fact when a second alarm is issued. This is to ensure a rapid response to an alarm issued when the manager is absent. Of course, the fastest way is for the manager to check the actual video of the monitoring area immediately after the first alarm is issued and to report directly to the fire department if it is determined that there is a fire, but since a second alarm has been issued, it is a situation where there is a fire or at least the scene must be checked directly, so the report can be made automatically to the fire department even when the manager is absent.

[0066] The monitoring unit (400) receives video footage captured by the camera (100) within the monitoring area from the video analysis unit (200) and displays it visually, and receives and displays the smoke or flame occurrence detection result from the video analysis unit (200). Additionally, the monitoring unit (400) displays the fire alarm when it is issued by the alarm issuing unit (300). The smoke or flame occurrence detection result and the fire alarm may be displayed visually on a display window, etc., or may be displayed through a speaker, etc.

[0067] Meanwhile, the monitoring unit (400) may be configured to display the measurement values ​​transmitted from each integrated sensor module (500) by item. In this case, the measurement values ​​measured by the integrated sensor module (500) are utilized as elements for monitoring indoor air quality. Additionally, as a configuration for monitoring indoor air quality, the integrated sensor module (500) may further include a fine dust detection sensor.

[0068] The monitoring unit (400) may be a PC, laptop, tablet, smartphone, etc., which is capable of communication and has a display window to display images, and may include various electronic devices, not limited thereto.

[0069] The monitoring unit (400) is configured so that video captured by the camera (100) is played back in real time and can be viewed by an administrator. However, for the protection of personal information, it is desirable to configure the monitoring unit so that a person's face or entire body is mosaiced. To this end, when video is transmitted from the camera (100), the video analysis unit (200) compares it at set frames to identify changes in pixels and obtain the change patterns. At this time, the video analysis unit (200) has pre-learned information on how a person's movement pattern is implemented within the video. Using this, the video analysis unit (200) determines the presence or absence of a person by determining whether the pixel change pattern obtained from the real-time video matches the pre-learned human movement pattern, and mosaics the pixel area having the pixel change pattern determined to be where a person exists, and then provides the mosaic-processed video to the monitoring unit (400).

[0070] However, in the event of a fire, it is necessary to clearly identify the movement paths of the victims in order to save lives, and at this time, it is necessary to determine how many victims are in which direction by verifying the identity of the victims or by confirming whether the victim who just passed in one direction and the victim who immediately passed in the other direction are the same person. Accordingly, the video analysis unit (200) is characterized by not applying mosaic processing to the video provided to the monitoring unit (400) when it is determined that smoke or flames have occurred as a result of analyzing the video.

[0072] The present invention, configured as described above, can detect the occurrence of a fire by analyzing an image captured by a camera (100) to determine whether smoke or flames are generated, and since the manager can directly check the image, it has the effect of preventing false fire alarms.

[0073] In addition, the present invention has the effect of protecting privacy by applying mosaic processing to a person's face or entire body when monitoring the image of the camera (100).

[0074] In addition, the present invention has the effect of allowing for more accurate verification of the fire situation by not applying mosaic processing to the monitored video when it is determined that a fire has occurred.

[0075] In addition, the present invention is equipped with an integrated sensor module (500) capable of detecting heat, gas, smoke, etc., and since the occurrence of a fire can be detected from the measurement value of the integrated sensor module (500), the actual presence of a fire can be determined more accurately, and accordingly, the effect of preventing false fire alarms can be achieved.

[0076] In addition, the present invention has the effect of preventing the fire alarm from being deactivated due to human error by ensuring that when the detection of a fire by the camera (100) image and the detection of a fire by the integrated sensor module (500) overlap, the manager cannot arbitrarily deactivate the fire alarm. Explanation of the symbols

[0078] 100 : Camera 200 : Video Analysis Department 300 : Alarm Issuance Unit 400 : Monitoring Department 500: Integrated sensor module

Claims

Claim 1 A camera (100) that films within a surveillance area; an image analysis unit (200) that receives images captured by the camera (100) in real time and determines whether smoke or flames have occurred in the images; an alarm issuing unit (300) that issues a fire alarm when it is determined by the image analysis unit (200) that smoke or flames have occurred; and a monitoring unit (400) that receives the images from the image analysis unit (200) and displays them visually, and receives and displays the smoke or flame occurrence determination results from the image analysis unit (200) or displays the fire alarm issued by the alarm issuing unit (300). A fire monitoring system equipped with a non-fire alarm prevention function, comprising: an integrated sensor module (500) that provides a measurement value to the alarm issuing unit (300) by including one or more of a heat detection sensor, a gas detection sensor, and a smoke detection sensor; wherein the image analysis unit (200) acquires a pixel change pattern from the image and determines whether smoke or flame has occurred by determining whether the acquired pixel change pattern matches a previously learned smoke or flame movement pattern; and the alarm issuing unit (300) issues a first fire alarm when information determined to have occurred by the image analysis unit (200) is received, releases the first fire alarm when an alarm release signal is received from the monitoring unit (400), issues a second fire alarm when a measurement value greater than a set value is generated from the integrated sensor module (500) within a set time after the first fire alarm is issued, and the second fire alarm cannot be released by operation of the monitoring unit (400). Claim 2 A fire monitoring system equipped with a false fire alarm prevention function, wherein, in claim 1, the image analysis unit (200) learns the change pattern of pixels in an image in which actual smoke or flames are generated using an AI model. Claim 3 In paragraph 2, the above-mentioned image analysis unit (200) is characterized by learning the change pattern of pixels in an image of actual smoke or flame generated for each item of environmental variables including two or more of the width, length, height, wind direction, and wind speed of the space where smoke or flame is generated, in a fire monitoring system equipped with a false fire alarm prevention function. Claim 4 In paragraph 3, the above image analysis unit (200) is characterized by being able to infer a pixel change pattern in an unlearned environmental variable using a pixel change pattern learned for each environmental variable, thereby providing a fire monitoring system equipped with a non-fire alarm prevention function. Claim 5 A fire monitoring system equipped with a non-fire alarm prevention function, wherein, in claim 1, the image analysis unit (200) acquires a pixel change pattern from the image when providing the image to the monitoring unit (400), determines the presence or absence of a person by determining whether the acquired pixel change pattern matches a previously learned human movement pattern, and mosaics the pixel area having the pixel change pattern determined to be present of a person. Claim 6 A fire monitoring system equipped with a non-fire alarm prevention function, wherein, in claim 5, the image analysis unit (200) does not mosaic the image provided to the monitoring unit (400) when it is determined that smoke or flames have been generated as a result of analyzing the image. Claim 7 delete Claim 8 delete Claim 9 A fire monitoring system equipped with a non-fire alarm prevention function, characterized in that, in claim 1, an alarm release means is provided on one side of the housing of the integrated sensor module (500), and by operating the alarm release means, a secondary fire alarm can be released. Claim 10 A fire monitoring system equipped with a non-fire alarm prevention function, wherein, in claim 1, the camera (100) is rotatably mounted on one side of the integrated sensor module (500), and the alarm issuing unit (300) rotates the camera (100) when a result determining that smoke or flame has occurred is not received from the image analysis unit (200) even though a measurement value greater than a set value is generated from the integrated sensor module (500), and issues a secondary fire alarm when a result determining that smoke or flame has occurred is received from the image analysis unit (200) after the camera (100) has rotated.

Images