Artificial intelligence-based automatic glass-breaking method and device

Abstract

The present invention relates to an artificial intelligence-based automatic glass-breaking method and device which analyzes images captured by an internal camera and an external camera to determine whether a vehicle or driver is in an emergency situation, and when the emergency situation occurs, breaks a glass window of the vehicle to enable the driver and passengers within the vehicle to quickly escape out of the vehicle, so that vehicle driver and passenger casualties can be minimized even though vehicle doors cannot be opened in an emergency situation such as a flood or fire.

Description

Artificial intelligence-based automatic glass breaking method and device

[0001] The present invention relates to an artificial intelligence-based automatic glass breaking method and device, and more specifically, to a method and device that recognizes an emergency situation using artificial intelligence and automatically breaks glass when an emergency situation occurs in a car.

[0002] Due to the popularization and widespread use of automobiles, the number of car owners is increasing every year. This increase in vehicles is also leading to an increase in traffic accidents, and the number of deaths and injuries from traffic accidents has also increased compared to the early days of automobile use.

[0003] With the recent increase in interest in automotive safety, peripheral technologies are also advancing to protect not only drivers but also passengers. Most vehicles are now equipped with front and side airbags to protect both drivers and passengers, and the number of fatalities from traffic accidents is gradually decreasing.

[0004] However, natural disasters caused by climate change, such as heavy rain, torrential downpours, floods, inundations, and landslides, are on the rise. Consequently, accidents involving vehicles drowning in torrential downpours and floods are on the rise.

[0005] When a vehicle is submerged in water, the water pressure outside the vehicle prevents passengers from opening the car doors, making it difficult to escape before the vehicle is completely submerged, often resulting in disaster.

[0006] In addition to flooding, accidents that involve hitting the side of a vehicle often result in the vehicle doors becoming deformed, making it difficult for passengers or outside rescuers to open the doors.

[0007] In the above-described case, there was a problem that it took a long time to rescue the passengers inside the vehicle because the vehicle door could not be opened easily, which resulted in the injuries of the injured person becoming more serious or, in severe cases, death.

[0008] Additionally, in the event of an emergency where the driver of the vehicle is in a critical condition such as a heart attack, fainting, or difficulty breathing, making it difficult for the driver to get out of the vehicle on their own, external rescuers have difficulty immediately rescuing the driver because the vehicle doors are locked.

[0009] The present invention provides a method and device for using artificial intelligence to assess an emergency situation in a vehicle and automatically destroy the vehicle's windshield in such an emergency. Furthermore, the present invention is not limited to the technical challenges described above, and other technical challenges may be derived from the following description.

[0010] An artificial intelligence-based automatic glass breaking device installed in a vehicle according to one embodiment of the present invention includes: an interior camera that photographs the interior of the vehicle and generates an interior image representing a driver's state inside the vehicle; at least one exterior camera that photographs the exterior of the vehicle and generates an exterior image representing an exterior state of the vehicle; an emergency situation determination module that determines, based on the interior image and the exterior image, whether an emergency situation has occurred in the vehicle or the driver and generates an emergency signal indicating whether an emergency situation has occurred; a control module that generates a signal for controlling the operation of a glass breaking module based on the emergency signal input from the emergency situation determination module; and a glass breaking module that is installed on a glass of the vehicle and breaks the glass of the vehicle based on a control signal input from the control module.

[0011] The above glass breaking module includes a heating wire attached to the automobile glass and a current supply device that supplies current to the heating wire.

[0012] The above emergency situation determination module divides the internal image input from the internal camera into a plurality of internal image frames representing the driver's condition, determines whether the driver's condition included in each of the plurality of internal image frames is an emergency situation, and if the driver of at least one internal image frame among the plurality of internal image frames is determined to be in an emergency situation, generates an emergency signal indicating that an emergency situation has occurred.

[0013] The above emergency situation judgment module includes a driver risk decision artificial neural network that is pre-trained to output a driver state included in an input internal image frame by a training set including a training internal image frame representing a driver state of a vehicle and a driver state represented by each training internal image frame.

[0014] The above emergency situation determination module divides an external image input from at least one external camera into a plurality of external image frames representing the external state of the vehicle, determines whether the external state of the vehicle included in each of the plurality of external image frames is an emergency situation, and, if it determines that the vehicle state of any one of the plurality of external image frames is an emergency situation, generates an emergency signal indicating that an emergency situation has occurred.

[0015] The above emergency situation judgment module includes an automobile risk decision artificial neural network that is pre-trained to output the state of the exterior of the vehicle indicated by the input external image frame by a training set including a training external image frame indicating the state of the exterior of the vehicle and the state of the exterior of the vehicle indicated by each training external image frame.

[0016] The above emergency situation determination module extracts an internal danger frame, which is an internal image frame in which the driver's condition is determined to be an emergency situation, from among the plurality of internal image frames, and determines whether the driver's condition is an emergency situation based on the extracted internal danger frame.

[0017] Even if an accident occurs outside the vehicle, such as a car submerged in water, a fire ignited around the vehicle, or the driver collapses from cardiac arrest, the system can determine that an emergency exists for the vehicle or driver. If an emergency occurs, the system can shatter the vehicle's windshield, allowing the driver inside the vehicle to quickly escape through the shattered glass. This can reduce casualties from traffic accidents.

[0018] In addition, according to embodiments of the present invention, even if a driver of a car faints, it is determined using artificial intelligence that an emergency has occurred for the driver, and when an emergency occurs for the driver, the driver's seat and the windshield of the car are automatically destroyed, so that rescue teams or bystanders can quickly rescue the unconscious driver.

[0019] Additionally, according to embodiments of the present invention, when a fire breaks out on the outside of a vehicle or the vehicle is flooded on one side, the driver and passengers inside the vehicle can more safely evacuate by destroying the vehicle window located in the direction opposite to the direction in which the emergency situation occurred. For example, when a fire breaks out on the left side of a vehicle, an automatic glass breaking device according to embodiments of the present invention destroys the right side window instead of the left side, allowing the driver and passengers inside the vehicle to evacuate to the right side instead of the left side where the fire occurred, thereby further reducing casualties due to fire accidents.

[0020] Figure 1 is a configuration diagram of an artificial intelligence-based automatic glass breaking device according to one embodiment of the present invention.

[0021] FIG. 2 is a drawing illustrating an example of a vehicle in which an artificial intelligence-based automatic glass breaking device according to one embodiment of the present invention is installed.

[0022] Figure 3 is a flowchart of an artificial intelligence-based automatic glass breaking method according to one embodiment of the present invention.

[0023] An artificial intelligence-based automatic glass breaking device installed in a vehicle according to one embodiment of the present invention includes: an interior camera that photographs the interior of the vehicle and generates an interior image representing a driver's state inside the vehicle; at least one exterior camera that photographs the exterior of the vehicle and generates an exterior image representing an exterior state of the vehicle; an emergency situation determination module that determines, based on the interior image and the exterior image, whether an emergency situation has occurred in the vehicle or the driver and generates an emergency signal indicating whether an emergency situation has occurred; a control module that generates a signal for controlling the operation of a glass breaking module based on the emergency signal input from the emergency situation determination module; and a glass breaking module that is installed on a glass of the vehicle and breaks the glass of the vehicle based on a control signal input from the control module.

[0024] The advantages and features of the present disclosure, and methods for achieving them, will become clearer with reference to the embodiments described in detail below together with the accompanying drawings. However, the technical idea of ​​the present disclosure is not limited to the embodiments described below and may be implemented in various different forms. The following embodiments are provided only to complete the technical idea of ​​the present disclosure and to fully inform those skilled in the art of the present disclosure of the scope of the present disclosure, and the technical idea of ​​the present disclosure is defined only by the scope of the claims.

[0025] When assigning reference numerals to components in each drawing, it should be noted that identical components are assigned the same numerals whenever possible, even if they appear on different drawings. Furthermore, when describing the present disclosure, if a detailed description of a related known configuration or function is deemed likely to obscure the gist of the present disclosure, such detailed description will be omitted.

[0026] Unless otherwise defined, all terms (including technical and scientific terms) used herein may be used in the same sense as commonly understood by those of ordinary skill in the art to which this disclosure pertains. Furthermore, terms defined in commonly used dictionaries are not to be interpreted ideally or excessively unless explicitly and specifically defined otherwise. The terminology used herein is for the purpose of describing embodiments and is not intended to limit the disclosure. In this specification, singular forms also include plural forms, unless specifically stated otherwise.

[0027] Additionally, terms such as first, second, A, B, (a), (b), etc. may be used to describe components of the present disclosure. These terms are only intended to distinguish the components from other components, and the nature, order, or sequence of the components are not limited by the terms. When a component is described as being "connected," "coupled," or "connected" to another component, it should be understood that the component may be directly connected or connected to the other component, but another component may also be "connected," "coupled," or "connected" between each component.

[0028] As used herein, the terms “comprises” and / or “comprising” do not exclude the presence or addition of one or more other components, steps, operations and / or elements.

[0029] Hereinafter, embodiments of the present invention will be described in detail together with drawings of the present invention.

[0030] FIG. 1 is a block diagram of an artificial intelligence-based automatic glass breaking device according to one embodiment of the present invention. Referring to FIG. 1, the artificial intelligence-based automatic glass breaking device (1) includes an interior camera (101) for capturing images inside a vehicle, at least one exterior camera (102) for capturing images outside the vehicle, an emergency situation determination module (103) for determining whether an emergency situation has occurred in the vehicle based on the captured images, a control module (104) for controlling a device for destroying vehicle glass based on whether an emergency situation has occurred, and a glass breaking module (105) for destroying vehicle glass.

[0031] The interior camera (101) is a camera that captures the interior of a vehicle, captures the driver driving the vehicle from inside the vehicle, and generates an interior image that indicates the driver's condition. The interior image is an image that represents the interior of the vehicle and indicates the driver's condition inside the vehicle. The artificial intelligence-based automatic glass breaking system can determine whether the driver is experiencing an emergency situation such as cardiac arrest, bleeding, or fainting based on the interior image generated by the interior camera (101).

[0032] At least one external camera (102) is a camera that photographs the exterior of the vehicle and generates an external image representing the external situation of the vehicle. At least one external camera (102) includes cameras that photograph the front, rear, left side, and right side of the vehicle, respectively. The external camera (102) photographs the front, back, left, and right sides of the vehicle and generates an external image representing the entire external situation of the vehicle. The artificial intelligence-based automatic glass breaking system can determine whether an emergency situation, such as a fire in the vehicle, a flood in the vehicle, or an accident in the vehicle, has occurred based on the external image generated by the external camera (102).

[0033] The emergency situation determination module (103) determines whether an emergency situation has occurred for the vehicle or the driver based on the internal image captured by the internal camera (101) and the external image captured by the external camera (102). The emergency situation determination module (103) determines whether an accident has occurred for the vehicle or an emergency rescue situation has occurred for the driver based on the image captured by the camera. If the emergency situation determination module (103) determines that an emergency situation has occurred for the vehicle or the driver, it generates an emergency signal indicating that an emergency situation has occurred and inputs the generated emergency signal to the control module (104).

[0034] The control module (104) controls the operation of the glass breaking module (105) based on an emergency signal input from the emergency situation judgment module (103). When an emergency signal indicating that an emergency situation has occurred for the vehicle or the driver is input from the emergency situation judgment module (103), the control module (104) operates the glass breaking module (105) to break the glass of the vehicle. When an emergency signal is input, the control module (104) generates a control signal that controls the operation of the glass breaking module (105) and inputs the generated control signal to the glass breaking module (105).

[0035] The control module (104) and the emergency situation judgment module (103) may be implemented by a separate module other than the ECU (electronic control unit) installed in the vehicle, and may be implemented by a computer program running in the ECU.

[0036] The glass breaking module (105) breaks the glass of the vehicle based on a control signal input from the control module (104). When a control signal for operating the glass breaking module (105) is input from the emergency situation judgment module (103), the glass breaking module (105) breaks the glass of the vehicle.

[0037] According to one embodiment of the present invention, a glass breaking module (105) includes a heating wire installed on a vehicle window and a current supply device that supplies current to the heating wire. When a control signal for operating the glass breaking module (105) is input from the emergency situation judgment module (103), the glass breaking module (105) operates the current supply device to supply current to the heating wire, thereby increasing the temperature of the heating wire.

[0038] As a high current flows through the heating wire, its temperature rises, causing the temperature surrounding the wire to rise rapidly. As the temperature of the glass surrounding the heating wire rises, the glass expands. This difference in expansion, caused by the temperature difference between the glass near the heating wire and the glass further away from the wire, generates stress, resulting in glass breakage.

[0039] FIG. 2 is a drawing illustrating an example of a vehicle in which an AI-based automatic glass breaker device according to an embodiment of the present invention is installed. Referring to FIG. 2, the AI-based automatic glass breaker device is installed in a vehicle. An emergency situation determination module (103) is connected to an interior camera (101) installed with a camera lens facing inward to photograph the interior of the vehicle, and at least one exterior camera (102) installed with a camera lens facing the front, left and right sides, or rear of the vehicle to photograph the exterior of the vehicle. The emergency situation determination module (103) receives an interior image or an exterior image generated by the connected interior camera (101) and at least one exterior camera (102). The emergency situation determination module (103) determines whether an emergency situation has occurred for the vehicle or the driver based on the received interior image or exterior image.

[0040] The interior camera (101) is installed in a direction that can capture the driver and passengers of the vehicle inside the vehicle. According to the example illustrated in Fig. 2, the interior camera (101) may be installed on the front ceiling of the vehicle or on the inside of the windshield of the vehicle.

[0041] The external camera (102) is installed in a direction that can capture the situation outside the vehicle. According to the example illustrated in FIG. 2, the external camera (102) is installed at the front of the vehicle. Although not illustrated in FIG. 2, an external camera (102) may be installed at the left and right sides of the vehicle to capture the left and right sides of the vehicle, and an external camera (102) may be installed at the rear of the vehicle to capture the rear of the vehicle.

[0042] The control module (104) is connected to the emergency situation judgment module (103) and the glass destruction module (102). The control module (104) controls the operation of the glass destruction module (105) attached to each glass based on the emergency signal input from the emergency situation judgment module (103). When an emergency signal indicating that an emergency situation has occurred is input, the control module (104) generates a glass destruction signal to operate the glass destruction module (102) to destroy the automobile glass, and inputs the generated glass destruction signal to the glass destruction module (105).

[0043] The glass breaking module (105) is installed on the windshield, left and right side windows, and rear window of the vehicle, respectively. The glass breaking module (105) operates based on a control signal input from the control module (104). When the control module (104) inputs a glass breaking signal to the glass breaking module (105) to operate the glass breaking module (105), the glass breaking module (105) breaks the glass. For example, when the glass breaking module (105) is a heating wire and a current supply device, the heating wire is attached to a specific part of the vehicle glass. The current supply device supplies current to the heating wire attached to the vehicle glass.

[0044] Figure 3 is a flowchart of an AI-based automatic glass breaking method according to one embodiment of the present invention. The AI-based automatic glass breaking method according to one embodiment of the present invention is executed in a vehicle equipped with an AI-based automatic glass breaking device.

[0045] Referring to FIG. 3, in step 301, at least one external camera (102) generates an external image representing the external situation of the vehicle. The at least one external camera (102) photographs the exterior of the vehicle and generates an external image representing the external situation of the vehicle. The at least one external camera (102) inputs the generated external image into the emergency situation assessment module (103).

[0046] According to one embodiment of the present invention, four external cameras (102) capture the situations of the front, rear, left side, and right side of the vehicle, respectively, and generate a front external image, a rear external image, a left external image, and a right external image, respectively, which represent the situations of the front, rear, left side, and right side of the vehicle. The four external cameras (102) input the generated front external image, rear external image, left external image, and right external image into the emergency situation judgment module (103).

[0047] In step 302, the interior camera (101) generates an interior image representing the driver's condition inside the vehicle. The interior camera (101) captures the driver positioned inside the vehicle and generates an interior image representing the driver's condition. The interior camera (101) inputs the generated interior image into the emergency situation assessment module (103).

[0048] In step 303, the emergency situation assessment module (103) determines whether an emergency situation has occurred for the vehicle or the driver based on the external and internal images. The emergency situation assessment module (103) analyzes the external and internal images and, based on the analysis results, determines whether an emergency situation has occurred in at least one of the external and internal images. Here, an emergency situation includes a situation in which it is difficult to escape from the vehicle on one's own, such as when the vehicle falls into water, a fire breaks out in the vehicle, or the driver of the vehicle is unconscious or injured.

[0049] More specifically, the emergency situation determination module (103) determines whether an emergency situation has occurred for the driver of the vehicle based on the internal video. The emergency situation determination module (103) divides the internal video into multiple internal video frames. The emergency situation determination module (103) divides the acquired internal video into multiple internal video frames. For example, if the internal video has a play time of 20 seconds and a frame rate of 30 FPS (Frames Per Second), it consists of a total of 600 frames.

[0050] Here, the internal video is a video representing the driver's condition. The internal video frame is an image representing the driver's condition at rest. The internal video consists of multiple internal video frames.

[0051] The emergency situation determination module (103) extracts an internal image frame indicating a state in which the driver is collapsed from a plurality of segmented internal image frames. The emergency situation determination module (103) includes a driver risk determination artificial neural network that determines whether the driver's state included in each of the plurality of segmented internal image frames is a dangerous state. The driver risk determination artificial neural network outputs whether the driver's state included in the internal image frame indicating the driver's state is a dangerous state or not.

[0052] More specifically, the driver risk determination artificial neural network determines that the driver's state represented by the internal video frame is normal if the driver's state is normal, and determines that the driver's state represented by the internal video frame is abnormal (for example, the driver is not sitting in the driver's seat but has collapsed), and determines that the driver's state is dangerous. The driver risk determination artificial neural network is an artificial neural network that is trained in advance to output the driver's state represented by the input internal video frame by a training set including training internal video frames representing the driver's state of the vehicle and the driver's state (normal state or dangerous state) represented by each training internal video frame.

[0053] The emergency situation judgment module (103) determines an internal video frame in which the driver's condition is determined to be dangerous as an internal danger frame and extracts the determined internal danger frame.

[0054] The emergency situation determination module (103) determines whether the driver is in an emergency situation based on the extracted internal risk frames. The emergency situation determination module (103) calculates the number of extracted internal risk frames and calculates the time the driver remains in a dangerous state using the number of extracted internal risk frames and the FPS of the internal video. The time the driver remains in a dangerous state is calculated using the following mathematical equation 1.

[0055]

[0056] Here, t danger is the time the driver remains in the dangerous state, a is the number of extracted internal dangerous frames, and b is the FPS of the internal video.

[0057] For example, if the number of internal risk frames is 1,800 and the FPS of the internal video is 60 FPS, the time the driver remained in the risk state was 30 seconds.

[0058] The emergency situation judgment module (103) compares the calculated time for which the driver maintained the dangerous state with the preset dangerous standard time, and if the time for which the driver maintained the dangerous state is longer than the dangerous standard time, it determines that the driver has encountered an emergency situation, and if not, it determines that the driver has not encountered an emergency situation.

[0059] In addition, the emergency situation judgment module (103) determines whether an emergency situation has occurred outside the vehicle based on external images input from at least one external camera (102). The emergency situation judgment module (103) divides each external image into a plurality of external image frames. The emergency situation judgment module (103) divides the acquired external image into a plurality of external image frames. For example, if the external image has a play time of 20 seconds and a frame rate of 30 FPS (Frames Per Second), it consists of a total of 600 frames.

[0060] Here, the exterior video is a video depicting the situation outside the car. The exterior video frame is an image depicting the situation outside the car at a standstill. The exterior video consists of multiple exterior video frames.

[0061] The emergency situation determination module (103) extracts an external image frame indicating that the external situation of the vehicle is an emergency situation from a plurality of segmented external image frames. The emergency situation determination module (103) includes an automobile risk determination artificial neural network that determines whether the external situation of the vehicle included in each of the plurality of segmented external image frames is an emergency situation. The automobile risk determination artificial neural network outputs whether the external situation of the vehicle included in the external image frame indicating the external situation of the vehicle is an emergency situation or not.

[0062] More specifically, the automobile risk determination artificial neural network determines that the external situation of the automobile represented by the external image frame is normal if the external situation of the automobile represented by the external image frame is normal, and determines that the external situation of the automobile represented by the external image frame is abnormal (for example, when the automobile is submerged or when the automobile is on fire), and determines that the external situation of the automobile represented by the external image frame is dangerous. The automobile risk determination artificial neural network is an artificial neural network that is trained in advance to output the external situation of the automobile represented by the input external image frame by a training set including training external image frames representing the external situation of the automobile and the external state of the automobile (normal state or dangerous state) represented by each training external image frame.

[0063] The emergency situation judgment module (103) determines an external video frame in which the situation outside the vehicle is determined to be a dangerous situation as an external danger frame and extracts the determined external danger frame.

[0064] The emergency situation judgment module (103) determines that an emergency situation has occurred outside the vehicle if an external risk frame is extracted, and if not, determines that no emergency situation has occurred outside the vehicle.

[0065] The emergency situation determination module (103) determines that an emergency situation has occurred when it determines that an emergency situation has occurred in at least one of the external and internal images.

[0066] In step 304, if it is determined that an emergency situation has occurred, the emergency situation determination module (103) generates an emergency signal indicating that an emergency situation has occurred and inputs the generated emergency signal to the control module (104).

[0067] An emergency situation determination module (103) according to another embodiment of the present invention determines an emergency image representing an image in which an emergency situation has occurred among internal and external images. For example, if a car includes an internal camera (101) and four external cameras (102) that respectively capture the front, rear, left, and right sides, an emergency image is determined to represent whether the image in which an emergency situation has occurred is an internal image captured by the internal camera (101) or an external image captured by any of the four external cameras (102). The emergency situation determination module (103) inputs emergency image data representing the determined emergency image to the control module (104) along with an emergency signal.

[0068] At step 305, the control module (104) generates a glass breaking signal to operate the glass breaking module (105) based on the emergency signal, and inputs the generated glass breaking signal into the glass breaking module (105).

[0069] According to one embodiment of the present invention, when an emergency signal is input from an emergency situation judgment module (103), the control module (104) generates a glass destruction signal that operates all glass destruction modules (105) to destroy all glass of the vehicle, and inputs the generated glass destruction signal to all glass destruction modules.

[0070] According to one embodiment of the present invention, when an emergency signal including emergency image data representing an emergency image that indicates an image in which an emergency situation has been determined to have occurred is input from an emergency situation determination module (103), the control module (104) generates a glass destruction signal for operating a glass destruction module (105) based on the position of the camera that captured the emergency image, and inputs the generated glass destruction signal to the glass destruction module (105).

[0071] In case the camera that recorded the emergency video is an external camera (102), in order to destroy the glass located on the opposite side of the external camera (102) that recorded the emergency video, a glass destruction signal is generated to operate the glass destruction module (105) installed on the opposite side of the external camera (102) that recorded the emergency video, and the generated glass destruction signal is input to the glass destruction module (105) installed on the opposite side of the external camera (102) that recorded the emergency video.

[0072] For example, if the emergency situation judgment module (103) determines that an emergency situation has occurred from a camera filming the right side of the car, in other words, if a fire has occurred on the right side of the car, the control module (104) generates a glass destruction signal that operates a glass destruction module (105) installed on the left side window of the car, and inputs the generated glass destruction signal to the glass destruction module (105) installed on the left side window.

[0073] If the camera that recorded the emergency video is an internal camera (101), a glass destruction signal is generated to operate the glass destruction module (105) installed on the front and driver's side windows of the vehicle to destroy the front and driver's side windows of the vehicle, and the generated glass destruction signal is input to the glass destruction module (105) installed on the front and driver's side windows of the vehicle.

[0074] According to another example, when the emergency situation judgment module (103) determines that an emergency situation has occurred from a camera filming the inside of a car, in other words, when an emergency situation has occurred for the driver of the car, the control module (104) generates a glass destruction signal that operates a glass destruction module (105) installed on the front and left side windows of the car, and inputs the generated glass destruction signal to the glass destruction module (105) installed on the front and left side windows.

[0075] At step 306, the glass breaking module (105) breaks the glass of the vehicle based on the glass breaking signal. For example, if the glass breaking module (105) is a heating wire and a current supply device, the current supply device supplies current to the heating wire. As a high current flows through the heating wire, the temperature of the heating wire increases, and accordingly, the temperature around the heating wire increases rapidly. As the temperature of the glass around the heating wire increases, the glass around the heating wire undergoes thermal expansion. At this time, stress is generated due to the difference in expansion caused by the temperature difference between the glass around the heating wire and the glass far from the heating wire, causing the glass to break.

[0076] According to the embodiments of the present invention described above, even if an accident occurs outside the vehicle, such as the vehicle submerging in water, a fire igniting around the vehicle, or the driver collapsing from cardiac arrest, the system determines that an emergency has occurred for the vehicle or the driver, and if an emergency occurs, the vehicle window is broken, allowing the driver inside the vehicle to quickly escape to the outside through the broken window after the accident. Accordingly, casualties due to traffic accidents can be reduced.

[0077] In addition, according to embodiments of the present invention, even if a driver of a car faints, it is determined using artificial intelligence that an emergency has occurred for the driver, and when an emergency occurs for the driver, the driver's seat and the windshield of the car are automatically destroyed, so that rescue teams or bystanders can quickly rescue the unconscious driver.

[0078] Additionally, according to embodiments of the present invention, when a fire breaks out on the outside of a vehicle or the vehicle is flooded on one side, the driver and passengers inside the vehicle can more safely evacuate by destroying the vehicle window located in the direction opposite to the direction in which the emergency situation occurred. For example, when a fire breaks out on the left side of a vehicle, an automatic glass breaking device according to embodiments of the present invention destroys the right side window instead of the left side, allowing the driver and passengers inside the vehicle to evacuate to the right side instead of the left side where the fire occurred, thereby further reducing casualties due to fire accidents.

[0079] The present invention has been described with a focus on preferred embodiments. Those skilled in the art will appreciate that the present invention can be implemented in modified forms without departing from its essential characteristics. Therefore, the disclosed embodiments should be considered illustrative rather than limiting. The scope of the present invention is set forth in the claims, not the foregoing description, and all differences within the scope equivalent thereto should be construed as being encompassed by the present invention.

Claims

1. In an artificial intelligence-based automatic glass breaking device installed in a car, An interior camera (101) that photographs the interior of the vehicle and generates an interior image representing the driver's condition inside the vehicle; At least one external camera (102) for photographing the exterior of the vehicle and generating an external image representing the external condition of the vehicle; An emergency situation judgment module (103) that determines whether an emergency situation has occurred in the vehicle or the driver based on the internal image and the external image, and generates an emergency signal indicating whether an emergency situation has occurred; A control module (104) that generates a signal for controlling the operation of the glass breaking module (105) based on an emergency signal input from the emergency situation judgment module (103); and An artificial intelligence-based automatic glass breaking device characterized by including a glass breaking module (105) installed on the glass of the automobile and breaking the glass of the automobile based on a control signal input from the control module (104).

2. In paragraph 1, The above glass breaking module (105) An artificial intelligence-based automatic glass breaking device characterized by including a heating wire attached to the automobile glass and a current supply device supplying current to the heating wire.

3. In paragraph 2, The above emergency situation judgment module (103) The internal image input from the internal camera (101) is divided into a plurality of internal image frames representing the driver's status, Determine whether the driver's condition included in each of the above multiple internal video frames is in an emergency situation, An artificial intelligence-based automatic glass breaking device characterized in that, when the driver of at least one internal image frame among the plurality of internal image frames determines that an emergency situation has occurred, an emergency signal is generated indicating that an emergency situation has occurred.

4. In paragraph 3, The above emergency situation judgment module (103) An artificial intelligence-based automatic glass breaking device, characterized in that it includes a driver risk determination artificial neural network that is pre-trained to output a driver state included in an input internal image frame by a training set including a training internal image frame representing a driver state of a vehicle and a driver state represented by each training internal image frame.

5. In paragraph 3, The above emergency situation judgment module (103) Divide the external image input from at least one external camera (102) into a plurality of external image frames representing the external state of the vehicle, Determine whether the external condition of the vehicle included in each of the above multiple external image frames is an emergency situation, An artificial intelligence-based automatic glass breaking device characterized in that, when it is determined that the state of a vehicle in any one of the plurality of external image frames is an emergency situation, an emergency signal is generated indicating that an emergency situation has occurred.

6. In paragraph 5, The above emergency situation judgment module (103) An artificial intelligence-based automatic glass breaking device characterized in that the vehicle risk decision artificial neural network includes a vehicle risk decision artificial neural network that is pre-trained to output a state of the exterior of a vehicle indicated by an input external image frame, by a training set including training external image frames indicating a state of the exterior of a vehicle, and a state of the exterior of a vehicle indicated by each training external image frame.

7. In paragraph 3, The above emergency situation judgment module (103) Among the above multiple internal video frames, an internal danger frame is extracted, which is an internal video frame in which the driver's condition is determined to be in a critical situation, and An artificial intelligence-based automatic glass breaking device characterized in that it determines whether the driver's condition is an emergency situation based on the extracted internal risk frame.

Images