Method, apparatus, and program for detecting dangerous situation within edge device-based space

Abstract

Disclosed is a method for detecting a dangerous situation within an edge device-based space, according to various embodiments of the present invention. The method may comprise the steps of: receiving a dangerous situation detection signal from at least one device from among a plurality of edge devices installed in a dangerous situation detection target space; classifying the type and severity of the dangerous situation on the basis of the dangerous situation detection signal; and providing a warning notification on the basis of a notification policy corresponding to the type and severity of the dangerous situation.

Description

Method, device, and program for detecting hazardous situations within a space based on edge devices

[0001] The present invention relates to a method, device, and program for detecting dangerous situations within a space based on an edge device.

[0002] Due to factors such as urbanization, population aging, and climate change, the importance of detection and response technologies for various hazardous situations that may occur in public spaces is growing. In particular, in multi-use spaces such as parks, plazas, schools, hospitals, and mixed-use commercial facilities, rapid and precise detection, as well as appropriate alerts and responses, are essential in the event of an accident. To this end, various attempts are currently being made to combine sensor technology with artificial intelligence to automatically recognize and assess abnormal situations without human intervention.

[0003] Existing risk detection systems typically feature a structure where data collected via CCTV or environmental sensors is transmitted to a central server for integrated analysis. While this approach offers the advantage of enabling unified monitoring even in large-scale spaces, it suffers from poor real-time performance due to network latency or slow processing speeds during data transmission. In particular, in situations where a "golden hour" response immediately following an incident is critical, delays in detection and warning can pose a significant obstacle to ensuring safety. Furthermore, centralized systems involve high installation and operating costs, and there is a risk that the entire system could be paralyzed in the event of a server failure.

[0004] Some systems have limitations in recognizing complex risk situations or determining the severity of situations in detail because they utilize only specific types of sensors or perform simple detection based on fixed rules. For example, even though the severity of the same sound signal can vary depending on the surrounding environment, time of day, and repetition, existing technologies find it difficult to quantitatively distinguish or adaptively judge these factors. Furthermore, notifications provided to users when a risk situation occurs are often delivered in a uniform form, making it difficult to respond flexibly to the situation.

[0005] Therefore, there is a demand in the industry for a risk situation detection method capable of comprehensively detecting risk situations that may occur in various spatial environments and responding appropriately according to severity. In this regard, Korean Registered Patent No. 10-2647328 discloses an edge-type surveillance camera AI abnormal situation detection control device and method.

[0006] The technical problem that the present invention aims to solve is to provide a method, device, and program for detecting dangerous situations within a space based on edge devices.

[0007] The technical problems of the present invention are not limited to those mentioned above, and other unmentioned technical problems will be clearly understood by those skilled in the art from the description below.

[0008] According to an embodiment of the present invention for solving the problem described above, a method for detecting a dangerous situation within a space based on an edge device is disclosed. The method may include: receiving a dangerous situation detection signal from at least one device among a plurality of edge devices installed in a space to be detected for dangerous situations; classifying the type and severity of the dangerous situation based on the dangerous situation detection signal; and providing a warning notification based on a notification policy corresponding to the type and severity of the dangerous situation.

[0009] In an alternative embodiment, the method may further include: acquiring normal state data from each of the plurality of edge devices at preset intervals; generating judgment criteria information for detecting a dangerous situation based on the normal state data; and transmitting the judgment criteria information to each of the plurality of edge devices.

[0010] In an alternative embodiment, each of the plurality of edge devices converts monitoring data collected in real time into input data for inputting into a pre-trained risk detection model, and inputs the input data and the judgment criterion information into the risk detection model to detect a risk situation, and the risk detection model may include at least one of a behavior-based risk detection model that detects a risk situation based on image data, an acoustic-based risk detection model that detects a risk situation based on acoustic data, and an environment-based risk detection model that detects a risk situation based on environment data.

[0011] In an alternative embodiment, the method comprises: collecting monitoring data corresponding to the risk situation detection signal when receiving a risk situation detection signal from the at least one device; labeling the risk situation on the monitoring data to generate training data; and transmitting the training data to other devices of the same type as the at least one device, wherein the other devices of the same type as the at least one device can update the risk detection model based on the training data.

[0012] In an alternative embodiment, the step of classifying the type and severity of the risk situation based on the risk situation detection signal may include: a step of recognizing the type of a specific device that transmitted the risk situation detection signal; a step of classifying the type of the risk situation based on the type of the specific device; and a step of classifying the severity of the risk situation based on monitoring data corresponding to the risk situation detection signal.

[0013] In an alternative embodiment, the step of classifying the type of dangerous situation based on the type of the specific device comprises classifying the dangerous situation as an acoustic-based dangerous situation if the type of the specific device is a device type that collects acoustic data, classifying the dangerous situation as a visual-based dangerous situation if the type of the specific device is a device type that collects video data, and classifying the dangerous situation as an environment-based dangerous situation if the type of the specific device is a device type that collects environment data; and the step of classifying the severity of the dangerous situation based on monitoring data corresponding to the dangerous situation detection signal may comprise classifying the severity of the dangerous situation based on at least one of the intensity, duration, and rate of change of the signal included in the monitoring data.

[0014] In an alternative embodiment, the step of providing a warning notification based on a notification policy corresponding to the type and severity of the risk situation comprises at least one of the steps of: transmitting a warning notification to an administrator terminal; and outputting a warning notification through an electronic display or speaker provided in the space targeted for detection of the risk situation. The notification policy may include a first notification policy that transmits the warning notification only to the administrator terminal according to the type and severity of the risk situation, and a second notification policy that outputs the warning notification through the electronic display or speaker in conjunction with transmitting the warning notification to the administrator terminal, wherein the second notification policy controls the color of the warning notification output through the electronic display or the volume of the warning notification output through the speaker in correspondence with the severity of the risk situation.

[0015] In an alternative embodiment, the step of outputting a warning notification through an electronic display or speaker provided in the space targeted for detecting a dangerous situation may include: acquiring map information of the space targeted for detecting a dangerous situation; recognizing a point where a dangerous situation occurs based on the dangerous situation detection signal; recognizing an evacuation route that bypasses the point where a dangerous situation occurs based on the map information; and outputting a warning notification guiding the evacuation route based on a specific location where the electronic display or speaker is installed.

[0016] According to one embodiment of the present invention for solving the above-described problem, an apparatus is disclosed. The apparatus comprises: a memory for storing one or more instructions; and a processor for executing the one or more instructions stored in the memory, and the processor can perform the above-described methods by executing the one or more instructions.

[0017] According to one embodiment of the present invention for solving the above-described problem, a computer program stored on a computer-readable recording medium is disclosed, which is combined with a computer as hardware to perform the above-described methods.

[0018] Other specific details of the present invention are included in the detailed description and drawings.

[0019] The present invention can detect various types of dangerous situations occurring within a space based on a plurality of edge devices and provide customized warning notifications to administrators and users according to the type and severity of the dangerous situation.

[0020] The effects of the present invention are not limited to those mentioned above, and other unmentioned effects will be clearly understood by a person skilled in the art from the description below.

[0021] FIG. 1 is a drawing illustrating a system according to one embodiment of the present invention.

[0022] FIG. 2 is a hardware configuration diagram of a computing device according to one embodiment of the present invention.

[0023] FIG. 3 is a hardware configuration diagram of an edge device according to one embodiment of the present invention.

[0024] FIGS. 4 to 8 are drawings illustrating a method for detecting dangerous situations within a space based on an edge device according to an embodiment of the present invention.

[0025] The advantages and features of the present disclosure and the methods for achieving them will become clear by referring to the embodiments described in detail together with the accompanying drawings. However, the present disclosure is not limited to the embodiments presented below, but can be implemented in various different forms and should be understood to include all modifications, equivalents, and substitutions that fall within the spirit and scope of the present disclosure. The embodiments presented below are provided to make the present disclosure complete and to fully inform those skilled in the art of the scope of the invention. In describing the present disclosure, detailed descriptions of related prior art are omitted where it is determined that such detailed descriptions may obscure the essence of the present invention.

[0026] The terms used herein are used merely to describe specific embodiments and are not intended to limit the disclosure. Unless otherwise defined, all terms used herein have the same meaning as generally understood by those skilled in the art to which this disclosure pertains.

[0027] In this specification, singular expressions include plural expressions unless the context clearly indicates otherwise. Furthermore, terms such as "comprising" or "having" are intended to specify the existence of the features, numbers, steps, actions, components, parts, or combinations thereof described in the specification, and should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof.

[0028] Additionally, terms including ordinal numbers, such as "first" or "second" as used herein, may be used to describe various components, but the components should not be limited by the terms. The terms are used solely for the purpose of distinguishing one component from another.

[0029] Phrases such as "in one embodiment," "according to one embodiment," "related to one embodiment," or "according to an implementation of one embodiment" in this specification do not necessarily refer to the same embodiment. Furthermore, throughout this specification, "examples" are arbitrary distinctions to facilitate the description of the present disclosure, and each embodiment does not need to be mutually exclusive. For example, configurations mentioned for the description of one embodiment may be applied and / or implemented in other embodiments, and may be modified and applied and / or implemented to the extent that they do not depart from the scope of the present disclosure.

[0030] Some embodiments of the present disclosure may be represented by functional block configurations and various processing steps. Some or all of these functional blocks may be implemented by various numbers of hardware and / or software configurations that perform specific functions. For example, the functional blocks of the present disclosure may be implemented by one or more microprocessors or by circuit configurations for a specific function.

[0031] Additionally, for example, the functional blocks of the present disclosure may be implemented in various programming or scripting languages. The functional blocks may be implemented as algorithms executed on one or more processors. Furthermore, the present disclosure may employ prior art for electronic configuration, signal processing, and / or data processing, etc. Terms such as "mechanism," "element," "means," and "configuration" may be used broadly and are not limited to mechanical and physical configurations. Additionally, terms such as "-part," "-module," etc. refer to a unit that processes at least one function or operation, which may be implemented in hardware or software, or as a combination of hardware and software.

[0032] Furthermore, the connecting lines or connecting members between the components depicted in the drawings are merely illustrative of functional connections and / or physical or circuit connections. In the actual device, connections between components may be represented by various alternative or added functional connections, physical connections, or circuit connections.

[0033] In addition, some components in the drawings may be depicted with their size or proportions slightly exaggerated. Also, components depicted in one drawing may not be depicted in another drawing.

[0034] The present disclosure will be described in detail below with reference to the attached drawings.

[0035]

[0036] FIG. 1 is a drawing illustrating a system according to one embodiment of the present invention.

[0037] Referring to FIG. 1, a system according to one embodiment of the present invention may include a computing device (100), a plurality of edge devices (200), a user terminal (300), and an external server (400). The system illustrated in FIG. 1 is according to one embodiment, and its components are not limited to the embodiment illustrated in FIG. 1 and may be added, changed, or deleted as needed.

[0038] In one embodiment, when a computing device (100) detects a dangerous situation at an edge device (200) installed in a space targeted for detecting dangerous situations, it may provide a warning notification corresponding to the dangerous situation. For example, depending on the severity of the dangerous situation, the computing device (100) may provide multi-channel notifications, such as sending a notification to an administrator terminal, outputting a warning message through an electronic display, or providing an audible notification through a speaker.

[0039] Specifically, the computing device (100) can receive a risk situation detection signal from at least one of a plurality of edge devices (200) installed in a risk situation detection target space. Additionally, the computing device (100) can classify the type and severity of the risk situation based on the risk situation detection signal. Furthermore, the computing device (100) can provide a warning notification based on a notification policy corresponding to the type and severity of the risk situation.

[0040] Accordingly, the computing device (100) of the present invention can provide user-customized notifications to enhance real-time responsiveness to dangerous situations.

[0041] Hereinafter, an example of a method in which an edge device (200) detects a dangerous situation in a space targeted for detection of dangerous situations and a computing device (100) provides a warning notification regarding the dangerous situation will be described later with reference to FIGS. 4 to 8.

[0042] In various embodiments, the computing device (100) may provide Web or Application-based services. However, it is not limited thereto.

[0043] The computing device (100) may include any type of computer system or computer device, such as, for example, a microprocessor, a mainframe computer, a digital processor, a portable device, and a device controller. However, it is not limited thereto.

[0044] Hereinafter, the hardware configuration of the computing device (100) will be described with reference to FIG. 2.

[0045] In one embodiment, a plurality of edge devices (200) may be, for example, a camera module for collecting image data, a microphone sensor for collecting sound data, an environment sensor for measuring temperature, humidity, illuminance, gas concentration, etc., or a sensor device including at least one of such sensors.

[0046] Additionally, each edge device (200) may include a data processing unit (i.e., a processor) capable of analyzing collected data in real time and a pre-trained artificial intelligence model, and may communicate with a computing device (100) or a nearby device through a network (500). However, the form or configuration of the edge device (200) may vary depending on the application environment and is not limited thereto.

[0047] Hereinafter, the hardware configuration of the edge device (200) will be described with reference to FIG. 3.

[0048] Meanwhile, the user terminal (300) may be connected to the computing device (100) via the network (500) and may be a terminal of a user that uses a risk notification service or monitoring function provided by the computing device (100). For example, the user terminal (300) may include a terminal of a user that receives a risk situation notification, is guided to the location of the risk situation and evacuation route, or views situation-specific response guidelines.

[0049] Here, the user terminal (300) may include, for example, various types of computer devices. Specifically, for example, the user terminal (300) may refer to various terminal devices such as smartphones, tablet PCs, desktops, and laptops.

[0050] The user terminal (300) includes a display on at least a part of the terminal and may include an operating system for running applications or extension-based services provided by the computing device (100). For example, the user terminal (300) may be a smartphone, but is not limited thereto, and the user terminal (300) may include all kinds of handheld-based wireless communication devices such as navigation, PCS (Personal Communication System), GSM (Global System for Mobile communications), PDC (Personal Digital Cellular), PHS (Personal Handyphone System), PDA (Personal Digital Assistant), IMT (International Mobile Telecommunication)-2000, CDMA (Code Division Multiple Access)-2000, W-CDMA (W-Code Division Multiple Access), Wibro (Wireless Broadband Internet) terminal, smartpad, tablet PC, etc., as wireless communication devices that ensure portability and mobility.

[0051] An external server (400) can be connected to a computing device (100) via a network (500), and can transmit and receive various information / data necessary for the computing device (100) to provide various functions, technologies, or services related to the method of the present invention, and can store and manage various information / data generated as the computing device (100) provides various functions, technologies, or services related to the method of the present invention.

[0052] For example, the external server (400) may be a database server that stores information used in various functions, technologies, or services related to the method of the present invention. As another example, the external server (400) may be a server that provides information used in various functions, technologies, or services related to the method of the present invention.

[0053] The network (500) may refer to a connection structure capable of exchanging information between each node, such as computing devices, multiple terminals, and servers. For example, the network (500) includes a Local Area Network (LAN), a Wide Area Network (WAN), the World Wide Web (WWW), a wired and wireless data network, a telephone network, a wired and wireless television network, etc.

[0054] Wireless data communication networks include, but are not limited to, 3G, 4G, 5G, 3GPP (3rd Generation Partnership Project), 5GPP (5th Generation Partnership Project), LTE (Long Term Evolution), WIMAX (World Interoperability for Microwave Access), Wi-Fi, Internet, LAN (Local Area Network), Wireless LAN (Wireless Local Area Network), WAN (Wide Area Network), PAN (Personal Area Network), RF (Radio Frequency), Bluetooth network, NFC (Near-Field Communication) network, satellite broadcasting network, analog broadcasting network, DMB (Digital Multimedia Broadcasting) network, etc.

[0055]

[0056] FIG. 2 is a hardware configuration diagram of a computing device according to one embodiment of the present invention.

[0057] Referring to FIG. 2, a computing device (100) according to one embodiment of the present invention may include one or more processors (110), a memory (120) for loading a computer program (151) executed by the processor (110), a bus (130), a communication interface (140), and a storage (150) for storing the computer program (151). Here, FIG. 2 illustrates only the components related to the embodiment of the present invention. Therefore, a person skilled in the art to which the present invention pertains will understand that other general-purpose components may be included in addition to the components illustrated in FIG. 2.

[0058] The processor (110) controls the overall operation of each component of the computing device (100). The processor (110) may be composed of one or more cores and may include processors for data analysis and deep learning, such as a central processing unit (CPU), a general purpose graphics processing unit (GPGPU), or a tensor processing unit (TPU) of the computing device. Alternatively, it may be configured to include any type of processor well known in the art of the present invention.

[0059] Additionally, the processor (110) can perform operations for at least one application or program for executing the method according to embodiments of the present invention, and the computing device (100) may have one or more processors.

[0060] In various embodiments, the processor (110) may further include RAM (Random Access Memory, not shown) and ROM (Read-Only Memory, not shown) for temporarily and / or permanently storing signals (or data) processed within the processor (110). Additionally, the processor (110) may be implemented in the form of a system-on-chip (SoC) comprising at least one of a graphics processing unit, RAM, and ROM.

[0061] Memory (120) stores various data, instructions and / or information. Memory (120) may load a computer program (151) from storage (150) to execute a method / operation according to various embodiments of the present invention. When the computer program (151) is loaded into memory (120), the processor (110) may perform the method / operation by executing one or more instructions constituting the computer program (151). Memory (120) may be implemented as volatile memory such as RAM, but the technical scope of the present invention is not limited thereto.

[0062] The bus (130) provides communication functions between components of the computing device (100). The bus (130) can be implemented as various types of buses, such as an address bus, a data bus, and a control bus.

[0063] The communication interface (140) supports wired and wireless internet communication of the computing device (100). Additionally, the communication interface (140) may support various communication methods other than internet communication. To this end, the communication interface (140) may be configured to include a communication module well known in the art of the present invention. In some embodiments, the communication interface (140) may be omitted.

[0064] Storage (150) can store a computer program (151) non-temporarily. When performing a process according to an embodiment of the present invention through a computing device (100), storage (150) can store various information necessary to perform a method according to the disclosed embodiment or to provide a service.

[0065] The storage (150) may be configured to include non-volatile memory such as ROM (Read Only Memory), EPROM (Erasable Programmable ROM), EEPROM (Electrically Erasable Programmable ROM), flash memory, a hard disk, a removable disk, or any form of computer-readable recording medium well known in the art to which the present invention belongs.

[0066] A computer program (151) may include one or more instructions that cause a processor (110) to perform a method / operation according to various embodiments of the present invention when loaded into memory (120). That is, the processor (110) may perform the method / operation according to various embodiments of the present invention by executing the one or more instructions.

[0067] In one embodiment, the computer program (151) may include one or more instructions to perform various methods related to various tasks related to learning a neural network model.

[0068] The steps of the method or algorithm described in connection with embodiments of the present invention may be implemented directly in hardware, implemented as a software module executed by hardware, or implemented by a combination thereof. The software module may reside in RAM (Random Access Memory), ROM (Read Only Memory), EPROM (Erasable Programmable ROM), EEPROM (Electrically Erasable Programmable ROM), Flash Memory, a hard disk, a removable disk, a CD-ROM, or any form of computer-readable recording medium well known in the art to which the present invention belongs.

[0069] The components of the present invention may be implemented as a program (or application) and stored on a medium to be executed in combination with a computer, which is hardware. The components of the present invention may be implemented as software programming or software elements, and similarly, embodiments may be implemented in programming or scripting languages ​​such as C, C++, Java, assembler, etc., including various algorithms implemented as combinations of data structures, processes, routines, or other programming configurations. Functional aspects may be implemented as algorithms executed on one or more processors.

[0070]

[0071] FIG. 3 is a hardware configuration diagram of an edge device according to one embodiment of the present invention.

[0072] Referring to FIG. 3, an edge device (200) according to one embodiment of the present invention may include one or more data collection units (210), a data processing unit (220) that processes data collected by the data collection unit (210) through a risk detection model (221), and a communication interface (230). Here, FIG. 3 illustrates only the components related to the embodiment of the present invention. Therefore, a person skilled in the art to which the present invention belongs will understand that other general-purpose components may be included in addition to the components illustrated in FIG. 3.

[0073] The data collection unit (210) can collect data such as video, sound, temperature, humidity, and gas concentration through one or more sensors installed outside the edge device (200). The collected data can be used to determine the type and likelihood of occurrence of a dangerous situation, and the type and installation location of the sensors can be configured in various ways depending on the spatial environment.

[0074] For example, the data collection unit (210) may include various types of sensors such as an image sensor (e.g., camera), an acoustic sensor (e.g., microphone), a temperature sensor, a humidity sensor, an illuminance sensor, a gas sensor, a vibration sensor, an infrared sensor, or a radar sensor. These sensors may be installed individually or in combination and may be selectively configured according to the characteristics of a specific space and the purpose of detection.

[0075] The data processing unit (220) can perform the role of preprocessing various types of data collected through the data collection unit (210) in real time and inputting them into a pre-trained risk detection model (221) to determine whether a risk situation has occurred. The preprocessing process may include operations such as noise removal, outlier correction, comparison with a normal state, or event pattern extraction.

[0076] In various embodiments, the data processing unit (220) classifies the type and severity of a dangerous situation based on the detection results and generates a warning signal or communication data accordingly, which can then be transmitted to a computing device (100) through a communication interface (230).

[0077] In addition to data processing, the data processing unit (220) can control the overall operation of each component of the edge device (200). That is, the data processing unit (220) may be composed of one or more cores as a processor of the edge device (200) and may include a processor for data analysis and deep learning, such as a central processing unit (CPU), a general purpose graphics processing unit (GPGPU), or a tensor processing unit (TPU) of a computing device. Alternatively, it may be configured to include any type of processor well known in the art of the present invention.

[0078] The risk detection model (221) may be an artificial intelligence-based model that determines whether a risk situation occurs by taking collected data as input. For example, the risk detection model (221) may include at least one of a behavior analysis model based on video data, an abnormal sound judgment model based on acoustic data, and a risk threshold judgment model based on environmental data.

[0079] The risk detection model (221) can be stored in a pre-trained form on the edge device (200) and executed independently, and can be updated or retrained by linking with an external server or computing device (100) according to a set period or specific conditions. Additionally, the risk detection model (221) can be implemented so that a single model or multiple models are configured in an ensemble structure to integrally analyze multiple data types.

[0080] The communication interface (230) supports wired and wireless internet communication of the edge device (200). Additionally, the communication interface (230) may support various communication methods other than internet communication. To this end, the communication interface (230) may be configured to include a communication module well known in the art of the present invention. In some embodiments, the communication interface (230) may be omitted.

[0081] FIGS. 4 to 8 are drawings illustrating a method for detecting dangerous situations within a space based on an edge device according to an embodiment of the present invention.

[0082] Referring to FIG. 4, the computing device (100) can receive a danger situation detection signal from at least one of a plurality of edge devices (200) installed in a danger situation detection target space (S110).

[0083] Specifically, multiple edge devices (200) can be distributed and installed in various locations where people frequently pass through or crowds may form, such as parks, plazas, campuses, outdoor spaces of hospitals, pedestrian walkways within smart cities, and indoor and outdoor complex spaces. The installation locations can be determined by considering the structure of each space, the possibility of dangerous situations occurring, user density, etc., and individual devices can function as unit monitoring nodes that monitor an area within a certain radius.

[0084] Multiple edge devices (200) can convert monitoring data collected in real time through sensors into input data for input into a pre-trained risk detection model.

[0085] For example, a plurality of edge devices (200) can preprocess the collected image data into an image frame sequence, the acoustic data into a frequency spectrum or Mel spectrogram, and the environmental data into a numerical vector in time order, respectively.

[0086] And, multiple edge devices (200) can detect a dangerous situation by inputting input data and judgment criteria information into a risk detection model.

[0087] That is, input data is input into a risk detection model mounted on an edge device (200), and the model can detect whether a risk situation has occurred based on judgment criteria information stored together with the input data. Here, the judgment criteria information may consist of a predefined normal state pattern, a threshold range, or baseline data from a recent time period, and may be set independently at the edge device or received from a computing device (100).

[0088] The risk detection model may include at least one of a behavior-based risk detection model that determines abnormal behavior based on video data, an acoustic-based risk detection model that determines abnormal sounds such as sudden high-pitched sounds, impact sounds, and screams based on acoustic data, and an environment-based risk detection model that determines critical situations based on environmental data such as temperature, humidity, illuminance, vibration, and gas concentration. These models may be implemented as neural network structures or rule-based algorithms optimized for each data type and may be executed in real time within the edge device (200).

[0089] For example, multiple edge devices (200) can be configured with different time-based detection strategies to perform risk detection based on sound and environment at night and risk detection based on video during the day, and can perform more agile detection by temporarily increasing the detection sensitivity when a specific event (e.g., performance, gathering, etc.) occurs.

[0090] Meanwhile, the computing device (100) receives a risk situation detection signal transmitted from the edge device (200) and can use it as basic data to classify the type and severity of the risk situation in a subsequent step.

[0091]

[0092] In one embodiment, the computing device (100) can classify the type and severity of a dangerous situation based on a dangerous situation detection signal (S120).

[0093] Referring to FIG. 5, the computing device (100) can recognize the type of specific device that transmitted a danger situation detection signal (S121).

[0094] Specifically, the computing device (100) can determine the type of data collected by the device based on device identification information (Device ID), data collection type, installation location information, etc., that are pre-registered to each edge device (200).

[0095] For example, the edge device (200) may periodically transmit metadata including its own sensor configuration or additionally transmit device type information along with a danger situation detection signal. Accordingly, the computing device (100) can receive this additional information and identify type information regarding whether the device primarily collects video, audio, or environmental data.

[0096] Meanwhile, the computing device (100) can classify the type of dangerous situation based on the recognized type after recognizing the type of specific device (S122).

[0097] Specifically, the computing device (100) can classify a dangerous situation as an acoustic-based dangerous situation if the type of a specific device is a type of device that collects acoustic data. For example, the computing device (100) can classify this as an acoustic-based dangerous situation, such as 'scream', 'explosion', 'object collision', etc., if the occurrence of a high frequency or impact sound is detected in a signal received from an edge device that includes a microphone sensor, and the information includes that the device utilized an acoustic-based judgment model.

[0098] Additionally, the computing device (100) can classify a dangerous situation as a visual-based dangerous situation if the type of specific device is a type of device that collects video data. For example, the computing device (100) can classify a dangerous situation as a visual-based dangerous situation if the danger detection signal transmitted from the video-based detection device includes 'change in human body posture', 'abnormal crowd density', 'abnormal contact state with the ground', etc.

[0099] Additionally, the computing device (100) can classify a dangerous situation as an environment-based dangerous situation if the type of a specific device is a type of device that collects environmental data. For example, the computing device (100) can classify a dangerous situation as an environment-based dangerous situation if the danger detection signal received from a specific edge device includes environmental measurement values ​​such as 'sudden temperature rise', 'excessive gas concentration', or 'sudden change in illuminance'.

[0100] In this way, the computing device (100) can obtain basic information that allows setting notification policies and response strategies specialized for each situation by classifying risk situations according to the device type and the form of collected data.

[0101] In an additional embodiment, the computing device (100) may classify, in the step (S122) of classifying the type of phase situation, not only the type of the risk situation but also whether the risk situation can be responded to by local warning measures alone, or whether a wide-area warning that must be immediately conveyed to all users in the space is required. That is, the computing device (100) may classify the risk situation as an individual response type risk situation or a total response type risk situation.

[0102] Specifically, the computing device (100) can determine the warning range by comprehensively considering the location of the occurrence of the dangerous situation, the propagation range of the detection signal, the expected radius of impact, and the statistical pattern of similar accidents. For example, in cases where the likelihood of direct damage spreading to the surroundings is low, such as the detection of a fall (visual-based) or a one-time acoustic anomaly (acoustic-based) occurring in a relatively limited area, the computing device (100) can classify this as an individually respondable dangerous situation and respond by sending a notification only to the terminal of the area manager or by outputting a warning only to the electronic display board of the area.

[0103] Additionally, when the computing device (100) recognizes a dangerous situation with a high potential for spread or multiple casualties, such as high-output continuous impact sound, simultaneous detection by multiple devices, a rapid temperature rise, or a hazardous gas leak, it may classify this as a total response type dangerous situation. In this case, the computing device (100) may transmit an immediate warning notification not only through the manager terminal but also through electronic display boards, speakers, emergency broadcasting systems, etc., installed throughout the space.

[0104] In this way, the computing device (100) can provide accurate and timely warnings to the necessary targets while minimizing unnecessary confusion by classifying not only the type and severity of the dangerous situation but also the necessary range of warnings.

[0105] The computing device (100) can classify the type of dangerous situation and, based on monitoring data corresponding to a dangerous situation detection signal, classify the severity of the dangerous situation (S123). Here, the severity of the dangerous situation may refer to the degree of impact that the situation has on a person's life, body, property, or public safety. For example, the severity of the dangerous situation may be defined as a multi-level grade such as 'caution (low risk)', 'warning (medium risk)', and 'danger (high risk)'.

[0106] Specifically, the computing device (100) can classify the severity of a dangerous situation based on at least one of the intensity, duration, and rate of change of a signal included in the monitoring data. Here, the intensity of the signal can be quantified, for example, as the decibel magnitude in the case of acoustic data, the magnitude of the motion vector or acceleration value in the case of video data, or the absolute value of temperature or gas concentration in the case of environmental data. The duration is an indicator for determining whether the dangerous signal has been maintained for a certain period of time or longer, and the rate of change may be an indicator indicating how rapidly the signal increases or decreases over time.

[0107] For example, in the case of an acoustic-based danger situation, the computing device (100) can classify the severity of the danger situation as high risk when an impact sound of 100 dB or more persists for 3 seconds or more. Additionally, the computing device (100) can classify the severity as low risk or warning level when a single sound of 70 to 80 dB is detected.

[0108] For another example, in the case of a visual-based danger situation, the computing device (100) can determine the severity of the danger situation as medium or high risk if there is no movement for more than 10 seconds after a person's falling motion is detected, or if the crowd density in a specific area increases to more than 4 people per 1 m².

[0109] As another example, in the case of an environment-based risk situation, the computing device (100) can classify the risk level of the risk situation as a risk class if it detects that the rate of temperature increase is 10 degrees or more per minute or that the carbon dioxide concentration exceeds 1,500 ppm.

[0110] Meanwhile, in the examples described above, the signal intensity (e.g., sound pressure level, object velocity within the image, gas concentration), duration (e.g., duration in seconds of high-output sound, time of maintaining a stationary posture, time of exposure to a high-concentration environment), and rate of change (e.g., rate of temperature increase, rate of increase in density between frames, etc.) for classifying the severity of the dangerous situation can be pre-set based on normal data (i.e., reference data in a state where no dangerous situation occurs) received from a plurality of edge devices (200). This reference information can be generated by reflecting the average value, variance, and trend of the normal state over time for each edge device.

[0111] In this way, the computing device (100) can quantitatively evaluate the severity by comprehensively analyzing the numerical characteristics and temporal patterns of the collected real-time data. Through this, the computing device (100) can more accurately select the notification policy for subsequent stages and determine the target, means, intensity, and output cycle of the warning notification in a manner that corresponds to the characteristics and urgency of the dangerous situation.

[0112] For example, differentiated notification policies based on the level of response for each situation can be applied, such as sending warning notifications only to the administrator's terminal in low-risk situations, and providing real-time warning messages and audio notifications not only to the administrator's terminal but also through electronic display boards and speakers within the space in high-risk situations.

[0113]

[0114] In one embodiment, the computing device (100) can provide a warning notification based on a notification policy corresponding to the type and severity of the dangerous situation (S130).

[0115] The notification policy of the present invention may be defined as a set of rules or conditions for determining the target, method, intensity, and output medium of warning notifications according to the type and severity of dangerous situations. In other words, the purpose is to maximize the effectiveness of alerts and minimize confusion caused by unnecessary warnings by delivering warnings in an appropriate manner tailored to different situations.

[0116] For example, the notification policy may include a first notification policy that transmits warning notifications only to an administrator terminal depending on the type and severity of the dangerous situation. The first notification policy may be designed so that only the administrator can proactively recognize and make judgments without causing unnecessary alarm to field users in low-risk or routine warning situations.

[0117] Additionally, the notification policy may include a second notification policy that outputs a warning notification via an electronic display or speaker in conjunction with transmitting a warning notification to an administrator terminal, and controls the color of the warning notification output via the electronic display or the volume of the warning notification output via the speaker in response to the severity of the dangerous situation. The second notification policy can induce direct and rapid awareness of danger by the user in high-risk situations. For example, the second notification policy may include a method of outputting a red warning message on the electronic display or transmitting a high-output alarm sound through the speaker depending on the severity.

[0118] Referring to FIG. 6, the computing device (100) can send a warning notification to an administrator terminal (S131).

[0119]

[0120] Specifically, the computing device (100) can generate a warning message containing key information such as the type, severity, location of occurrence, and time of occurrence of the dangerous situation and transmit it to an administrator-only terminal (e.g., control center console, mobile app, notification server, etc.).

[0121] For example, the computing device (100) can support rapid situation recognition and decision-making regarding action by automatically generating a typed message format such as '[Caution] Area A acoustic anomaly detection (85dB, 12 seconds)' or '[Danger] Area B collapse detection (video-based)' and delivering it to the manager.

[0122] Additionally, the computing device (100) may output a warning notification through an electronic display or speaker installed in the space where the danger situation is detected (S132). This step may be omitted according to the notification policy, for example, if the severity is low or a private response is required, it may be omitted, and may be performed only in high-risk situations or when evacuation guidance is required.

[0123] Specifically, the computing device (100) can generate a control signal including text and color information of a warning message to be displayed on an electronic display board, or generate and transmit a signal including notification sound and volume intensity information to be sent to a speaker system.

[0124] For example, the computing device (100) can provide a direct warning to the user on site by displaying a phrase such as 'Please evacuate immediately - Danger detected in Zone C' on an electronic display board with a red background, or by playing an emergency alarm sound at maximum volume.

[0125] Accordingly, the computing device (100) of the present invention can provide appropriate and effective warning notifications to both managers and field users by applying a differentiated notification policy based on the nature and severity of the situation, along with real-time detection of a dangerous situation. This prevents unnecessary confusion and enables the realization of a fast and accurate alarm system in situations where a response is actually required.

[0126]

[0127] According to various embodiments of the present invention, the computing device (100) may, in step (S132), provide a warning notification to a number of users to guide them to bypass the location where the dangerous situation occurred.

[0128] Specifically, the computing device (100) can obtain map information of a space subject to danger detection. Here, the map information may be a 2D or 3D digital map including major paths, entrances, obstacle locations, and major facility locations (e.g., electronic display boards, speakers, emergency exits, etc.) within the space, and may be provided by a public safety system, a building management system (BMS), or a separate space modeling server. For example, the computing device (100) may obtain the map information through a space database linked in real-time from a digital twin-based space information system built within the space.

[0129] In addition, the computing device (100) can recognize the location of a dangerous situation based on a dangerous situation detection signal.

[0130] Specifically, the computing device (100) can recognize the location of a dangerous situation by referring to installation coordinate information associated with a unique identifier (Device ID) of the edge device (200) that transmitted a danger detection signal from the map information.

[0131] For example, when a high-risk detection signal is received from 'Device ID: C3', the computing device (100) can determine a point within the space corresponding to the coordinates (X=45.2, Y=17.8) where device C3 is installed as the location where the dangerous situation occurs.

[0132] In addition, the computing device (100) can recognize an evacuation route that bypasses the point where a dangerous situation occurs based on map information.

[0133] Specifically, the computing device (100) can calculate an evacuation path to an entrance, emergency exit, or safe zone within a space using a graph-based path search algorithm (e.g., Dijkstra, A*, BFS, etc.) after deactivating path nodes within the radius based on a minimum avoidance radius (e.g., 3m) set around the point where the dangerous situation occurs. Additionally, the computing device (100) can select an optimal evacuation path by assigning path weights to each evacuation path, taking into account distance, estimated time required, passage width, presence of obstacles, and real-time crowd density data.

[0134] For example, the computing device (100) can set an emergency passageway as a priority evacuation route in a wide and open environment, even if the distance to the nearest entrance from the point of occurrence of a dangerous situation is 30m, or the crowd density is high and the entrance is narrow.

[0135] And, the computing device (100) can output a warning notification guiding the evacuation route based on a specific location where an electronic display or speaker is installed.

[0136] Specifically, the computing device (100) can generate a warning message optimized for the user's visual and auditory perception direction based on the installation location and direction information of each display board or speaker.

[0137] For example, the computing device (100) can provide consistent visual and auditory-based evacuation direction guidance to users in all directions by displaying a phrase such as “South entrance *?* Evacuate in this direction” on an electronic display board installed facing west at point A with a red background and flashing effects, and simultaneously repeatedly broadcasting a voice message “This area is dangerous. Please move south according to the instructions” through a speaker installed facing north.

[0138] Accordingly, the computing device (100) of the present invention recognizes a point where a dangerous situation occurs, calculates an evacuation route that bypasses the point, and provides a directional warning notification through an electronic display and a speaker, thereby enabling safer evacuation even in large spaces or complex indoor and outdoor structures.

[0139]

[0140] According to an additional embodiment of the present invention, when a computing device (100) obtains map information of a space to be detected for a dangerous situation, it can divide the space into pre-set area units and monitor in real time whether a dangerous situation occurs for each area.

[0141] Specifically, the computing device (100) can divide the entire target space into multiple areas (cell, zone, area, etc.) based on spatial structure data included in the map information. This area division can be automatically generated in units of a spatial coordinate system-based grid (e.g., 5m x 5m) or manually set based on major facilities or entrances, and each area can be assigned a unique identifier (zone ID).

[0142] The computing device (100) can determine the data collection range and spatial coverage for each device by mapping the installation location and detection range of the edge devices (200) included in each area. Through this, the computing device (100) can determine whether a dangerous situation has occurred and the level of risk by integrating and analyzing the frequency, type, and severity of the risk detection signals received in a specific area.

[0143] For example, a computing device (100) can classify an area as a 'high-risk state' if it receives three or more high-pressure abnormality detection signals from an acoustic-based edge device within the last 5 minutes in the 'C3 zone' and detects a sudden change in a person's posture from an image-based device in the same zone.

[0144] Additionally, the computing device (100) can display the risk status of each area in real time on a digital map using colors, flashing effects, boundary line highlighting, etc., according to the risk level. At this time, the risk level can be composed of multiple stages such as 'normal', 'caution', 'warning', and 'danger', and can support intuitively checking the risk trend by area on the UI of the user terminal (300).

[0145] Meanwhile, the computing device (100) can detect related events that may spread between areas based on area-unit risk information. For example, if 'Area B2' and 'Area C3' are adjacent, and crowd density increases rapidly in B2 and an acoustic anomaly occurs simultaneously in C3, the computing device (100) can analyze the correlation between these areas to determine the possibility of a wider range of risk spread and issue an early warning to the adjacent areas.

[0146] Additionally, the computing device (100) can trigger an automated response scenario based on area-unit risk information. For example, if a specific area is classified as being in a 'danger' state, a scenario can be executed to output an evacuation guidance message via an electronic display within that area and send a warning notification to an administrator terminal in an adjacent area. Here, the response scenario can be set through a predefined policy or an artificial intelligence-based recommendation algorithm.

[0147] Accordingly, the computing device (100) of the present invention manages the entire space in area units and comprehensively determines the risk situation of each area, thereby supporting a more systematic and rapid response to risk situations even in a large-scale space.

[0148]

[0149] According to various embodiments of the present invention, a computing device (100) can generate judgment criteria information based on normal data periodically received from a plurality of edge devices (200).

[0150] The judgment criterion information of the present invention is reference information for defining the normal range or normal pattern of monitoring data collected by an edge device, and may refer to reference data that supports determining whether there is an abnormality (i.e., the possibility of a dangerous situation occurring) by comparing it with data collected in real time by each device.

[0151] This judgment criterion information can be set at the individual device level or at the group level of devices of the same type, taking into account the sensor type, installation location, and characteristics by time period of each device, and can be configured in various forms such as average values, median values, standard deviations, minimum / maximum allowable ranges, normal distribution models, or pre-trained normal class models.

[0152] Based on this judgment criterion information, the computing device (100) can determine whether real-time data received from the edge device (200) has deviated from the standard range and quantitatively classify the severity of the risk situation by considering the degree of deviation.

[0153] Additionally, the edge device (200) can store the judgment criteria information locally or receive it in real time, thereby comparing and analyzing the collected monitoring data with the judgment criteria information, and if an abnormal sign is detected, use it as an input value for a risk detection model or directly detect a dangerous situation. Specifically, the edge device (200) can apply the judgment criteria information as a threshold, a normal distribution range, or a reference parameter of an anomaly detection algorithm, so that a pre-trained risk detection model can perform accurate predictions without false positives or false negatives.

[0154] Referring to FIG. 7, the computing device (100) can obtain normal state data from each of the plurality of edge devices (200) at preset intervals (S210).

[0155] Specifically, the computing device (100) may periodically request data from the point in time when the result of the risk detection model is determined to be normal among the monitoring data collected by each edge device (200), or may periodically receive normal data selected by the edge device itself. Here, 'normal state data' refers to raw sensor values ​​or preprocessed input values ​​at a point in time when it is determined that no dangerous situation, such as abnormal user behavior or sudden environmental changes, has occurred.

[0156] For example, the computing device (100) can recognize data collected under conditions with few crowds and limited external stimuli as normal state data based on park environment data for a specific time period each day (e.g., 2:00 AM to 4:00 AM), and can receive temperature, humidity, illuminance, sound pressure, video frame analysis values, etc., for this time period from multiple edge devices. In addition, the computing device (100) can periodically update long-term accumulated data to reflect differences in normal state by time period, day of the week, and season, and after removing noise or outliers, use it as input for calculating reference values ​​based on statistics or machine learning.

[0157] Additionally, the computing device (100) can generate judgment criteria information for detecting dangerous situations based on normal state data (S220).

[0158] Specifically, the computing device (100) can define a numerical range of normal state by classifying normal state data received from a plurality of edge devices (200) by time, location, and device type, and by calculating statistical characteristic values ​​(e.g., mean value, standard deviation, variance, quantile value, etc.) of each data.

[0159] Additionally, the computing device (100) can set a reference threshold based on the corresponding characteristic value or generate a reference distribution model (e.g., Gaussian model, IQR range, Mahalanobis distance-based reference value, etc.) for automatically detecting abnormal signs outside the normal range.

[0160] Such judgment criteria information can be individually configured for each device type (e.g., sound-based, video-based, environment-based), and even for the same device, it may be configured differently depending on the installation location or time of day.

[0161] For example, the computing device (100) can generate judgment criteria information for a 'nighttime normal state' based on the fact that the average motion vector of video-based normal data received from the park's east CCTV is 0.5 or less, and the average sound pressure during nighttime hours (e.g., 22:00 to 5:00) is less than 40 dB. As another example, if the normal concentration value in the data received from an edge device including a CO₂ sensor is stably maintained between 600 and 800 ppm, the computing device (100) can define judgment criteria information based on this by setting an upper 95% confidence interval so that a warning is generated when it exceeds 1,000 ppm.

[0162] And, the computing device (100) can transmit judgment criteria information to each of the multiple edge devices (200) (S230).

[0163] Specifically, the computing device (100) can package judgment criteria information by device type or by individual device and distribute and transmit it to each edge device via a network (500). At this time, the judgment criteria information may include different threshold values, tolerances, normal ranges, etc., depending on the type of sensor data (e.g., sound, video, environment) collected by each device, and may be configured to be transmitted together with metadata so that the edge device can appropriately interpret and apply it according to its environment.

[0164] Meanwhile, each of the multiple edge devices (200) that receive judgment criteria information from the computing device (100) can convert the monitoring data collected in real time into input data for input into a pre-trained risk detection model.

[0165] Specifically, the edge device (200) can preprocess the collected raw data, perform processes such as removing outliers, normalizing, and extracting features, and then form an input vector or time series data that can be compared with judgment criteria information.

[0166] For example, an edge device including a microphone sensor can convert real-time acoustic data into the frequency domain and configure spectral feature values ​​reflecting the difference from the normal frequency distribution defined in the judgment criteria as input data. As another example, a video-based device can map a person's movement path into coordinates to calculate the average speed and movement deviation, and use these as input vectors for comparison with the judgment criteria.

[0167] Additionally, each edge device (200) can detect a dangerous situation by inputting input data and judgment criteria information into a risk detection model. That is, the risk detection model embedded within the edge device determines the dangerous situation, and the edge device can generate a detection signal based on the result determined by the model and transmit it to the computing device (100).

[0168] For example, an edge device converts acoustic data collected in real time into input data and inputs it into an acoustic-based risk detection model along with judgment criterion information; if the model determines that a pattern in which a sound pressure level exceeding a certain threshold persists for a certain period of time is an anomaly, it can detect the corresponding event as a dangerous situation.

[0169] Accordingly, the computing device (100) and edge device (200) of the present invention generate judgment criteria information based on normal state data and utilize this for detecting dangerous situations at the edge device, thereby enabling the detection of dangerous situations by flexibly responding to changes in the situation in various spatial environments.

[0170]

[0171] According to various embodiments of the present invention, a computing device (100) can generate learning data that reflects the characteristics and occurrence conditions of a risk situation based on a risk situation detection signal received from an edge device (200), and propagate this to edge devices of the same type.

[0172] Referring to FIG. 8, when the computing device (100) receives a danger situation detection signal from at least one device, it can collect monitoring data corresponding to the danger situation detection signal (S310).

[0173] Specifically, the computing device (100) can retrieve monitoring data corresponding to a certain time interval before and after the occurrence of a dangerous situation based on the device identifier (Device ID) and occurrence time (Time Stamp) included in the detection signal, or obtain it by directly requesting it from the edge device (200).

[0174] For example, when a high-risk risk situation detection signal is received from the 'ID-D7' device, the computing device (100) can obtain video, audio, and environmental sensor data for 10 seconds before and after the risk detection time (e.g., from T-5 seconds to T+5 seconds) from the local storage or real-time data stream of the device 'ID-D7'.

[0175] Additionally, the computing device (100) can collect monitoring data from adjacent edge devices installed in the same area together and expand it into multi-source based learning data for complex situation awareness or reliability reinforcement.

[0176] Additionally, the computing device (100) can generate training data by labeling dangerous situations in the monitoring data (S320).

[0177] Specifically, the computing device (100) can generate label information by combining the monitoring data collected in step (S310), the type of edge device that generated the data (e.g., video, audio, environment), the type of detected dangerous situation (e.g., crowd density, screaming, sudden temperature rise, etc.), and the severity level, and structure this into a training data format linked to the raw data. Here, the labeling can be applied as a categorical class or a numerical risk level in the case of structured data, and as an annotation for an event point in time within a time series interval in the case of unstructured data.

[0178] For example, the computing device (100) may assign a label of ‘impact sound_high risk’ to the result of spectrum analysis of acoustic data received from a specific edge device, and store the data sample along with the label in the form of ‘acoustic_01.wav’, label ‘class=shock, level=high’. As another example, in the case of video-based monitoring data, the computing device (100) may assign labels such as ‘fall’, ‘severity: medium’ to the ‘frames 150~180’ section, thereby constructing training data in which dangerous situations are clearly identified by section.

[0179] And, the computing device (100) can transmit training data to other devices of the same type as at least one device (S330).

[0180] Specifically, the computing device (100) classifies edge devices having the same sensor configuration or risk detection purpose as the edge device where the training data was generated, and can transmit the training data to these devices. Here, the transmitted training data includes raw monitoring data, label information attached to the data (type and severity of risk), time of occurrence of the situation, and sensor metadata, and can be transmitted after being preprocessed according to a predefined format or model input format. In addition, the training data can be transmitted in a real-time streaming manner or in batch units, and a communication protocol between devices (TLS-based encryption, device authentication, etc.) can be applied to maintain security and integrity.

[0181] For example, the computing device (100) can transmit the training data generated for the ‘explosion_high risk’ situation from a specific acoustic-based edge device (e.g., ID-A1) to the same microphone sensor-based edge devices (e.g., ID-A2, ID-A3, etc.) so that the risk detection model of each device can learn to recognize the same situation or correct its performance.

[0182] Additionally, the computing device (100) can overcome the local learning limitations of individual edge devices and improve the overall system's cognitive accuracy by sharing learning data related to the same type of crowd density or fall event among image-based edge devices placed in locations with similar installation environments (e.g., doorways).

[0183] Meanwhile, other edge devices of the same type as at least one device can update the risk detection model based on training data. Here, the update can be performed by readjusting the weights of the existing model through new training data or by improving the detection sensitivity for risk types or patterns that were not previously reflected. That is, each edge device (200) can dynamically improve risk recognition performance by partially retraining (fine-tuning) the locally embedded model or, if necessary, through incremental learning, online learning methods, etc.

[0184] For example, if training data is delivered to a specific acoustic-based edge device (e.g., ID-A1) in which high-frequency acoustic patterns corresponding to 'impact sounds' are labeled as hazards, another edge device of the same type (e.g., ID-A2) can utilize this to redefine the criteria for judging abnormal acoustics in the existing model and update the hazard detection model by incorporating new feature vectors. As another example, a video-based edge device can use training data obtained from cases of multiple user falls to further incorporate detailed features of changes in human posture and calibrate the model to detect 'falling' situations earlier in a shorter time compared to the existing model.

[0185] Accordingly, the computing device (100) and edge device (200) of the present invention can generate training data by utilizing data at the time of a risk situation occurrence, and by utilizing this for sharing among similar devices and updating models, the risk detection accuracy and adaptability of the overall system can be continuously improved.

[0186]

[0187] According to various embodiments of the present invention, a computing device (100) can visualize the state of a space subject to danger detection in real time and provide an integrated control screen through a user terminal (300).

[0188] Specifically, the computing device (100) can model a space for detecting actual dangerous situations on a digital twin-based 3D virtual space and generate a visualized control screen by mapping the location and status information of multiple edge devices (200) installed in the space in real time. This screen can be provided as a UI in the form of a web or a dedicated application accessible from a user terminal (300), and can intuitively display key information such as location, time of occurrence, detection device ID, type of danger, and severity when a dangerous situation occurs.

[0189] Additionally, the computing device (100) manages 3D spatial data of the area under control (e.g., entrances, passages, major facilities, topographical features, etc.) and can display in real time environmental information (e.g., temperature, humidity, illuminance, gas concentration, etc.) linked to data from edge devices within the space and status information of major facilities. For example, intuitive status monitoring is possible by displaying sensor measurement values ​​and operational status (whether it is open, whether there is an equipment malfunction, etc.) for major facilities such as restrooms, outdoor performance venues, and underground security rooms as icons or colors.

[0190] Additionally, the computing device (100) may include a scenario builder function that, when it detects a dangerous situation related to visitors (e.g., falls, screaming, crowd density, etc.) along with visitor location or density information, visually highlights the event on the control screen and automatically suggests a response scenario for the dangerous situation. For example, it may provide sequential actions such as “fall detection in Zone A, call nearby manager, broadcast warning” as presets, or support the manager in directly creating an action flow.

[0191] In addition, the computing device (100) provides an interface for integrated management of data collected from each edge device and external system, thereby enabling efficient control and visualization of diversified data flows on a single integrated control system. Accordingly, an administrator using a user terminal (300) can receive a real-time advanced control environment that allows for the rapid recognition of risk factors within a complex space and the execution of appropriate responses.

[0192]

[0193] According to various embodiments of the present invention, a computing device (100) can monitor dangerous situations in real time regarding major facilities within a park and control related facilities as needed.

[0194] Specifically, the computing device (100) can collect real-time monitoring data from edge devices (200) installed in key facility areas such as outdoor performance venues, restrooms, rest areas, and electrical equipment rooms, and recognize dangerous situations or abnormal conditions by comparing event occurrence conditions or abnormal signs with judgment criteria.

[0195] For example, if the noise level of an outdoor performance venue rises rapidly or crowd density is detected in an adjacent space, the computing device (100) can automatically identify the area and display it as an alarm on the control screen, and generate a warning event so that the manager can immediately recognize it. In addition, if a sudden change in lighting, prolonged congestion, or abnormal operation patterns are detected based on sensors or video, the condition can be determined as a precursor and a priority inspection notification can be sent to the manager.

[0196] In addition, the computing device (100) can perform linked control with facility systems such as lighting, broadcasting, and access control equipment in conjunction with detected events. For example, if a dangerous situation occurs during nighttime hours, it can output a control signal to automatically turn on lights in adjacent areas or transmit an emergency broadcast. In addition, it can be systemically linked with doors, security sensors, and automatic control devices within the facility to enable subsequent control, such as automatic closing or calling an administrator, when necessary.

[0197] In this way, the computing device (100) of the present invention can enhance the safety of the entire park facilities and improve the effectiveness of emergency response by integrally performing event detection, alarm display, system linkage, and control for each major facility.

[0198]

[0199] According to various embodiments of the present invention, the computing device (100) can detect risk factors in a user-dense area within a park in real time and perform monitoring and response functions to ensure the safety of users.

[0200] Specifically, the computing device (100) collects real-time data from multiple edge devices (200) linked with a LiDAR sensor, an artificial intelligence-based image detection system, an acoustic analysis module, etc., and analyzes the density of people, the possibility of wild animals appearing, and whether abnormal sounds occur based on this, thereby enabling early detection of the possibility of danger.

[0201] For example, the computing device (100) can identify abnormal human density patterns or unexpected animal movement paths within a specific area in real time based on crowd distribution information or object recognition results extracted from an edge device linked to an artificial intelligence-based image detection system, and if abnormal sounds such as high-frequency screams or animal cries are detected, it can consider this as a signal of an abnormal situation and immediately generate a warning notification.

[0202] Additionally, the computing device (100) can be configured to visually display these image-based detection results on a 3D digital twin-based control screen and to store risk detection history and location-based records in parallel, thereby enabling subsequent pattern analysis or incident response history management.

[0203] For example, if rapid movement of a large number of people is detected within a short period of time at the north entrance of the park, or if a series of short high-pitched sounds are detected within the same area, warning indications can be enhanced around the location, and visualization information from a 3D viewpoint accessible from multiple angles can be provided to the control screen to support the manager's situational judgment.

[0204] In addition, the computing device (100) of the present invention can be linked with an existing video surveillance system (CCTV, etc.) to utilize high-resolution real-time video streams of major surveillance areas in parallel, or implement a mutual verification structure between detection results and existing video. Through this, the computing device (100) can reduce the false positive rate and increase the reliability of the response.

[0205]

[0206] Meanwhile, embodiments according to the present disclosure may be implemented in the form of a computer program that can be executed through various components on a computer, and such a computer program may be recorded on a computer-readable medium. In this case, the medium may include, but is not limited to, magnetic media such as hard disks, floppy disks, and magnetic tapes, optical recording media such as CD-ROMs and DVDs, magneto-optical media such as floptical disks, and hardware devices specifically configured to store and execute program instructions such as ROM, RAM, and flash memory.

[0207] Meanwhile, the above computer program may be one specifically designed and configured for the present disclosure or one known and available to those skilled in the art of computer software. Examples of computer programs may include machine code, such as that produced by a compiler, as well as high-level language code that can be executed by a computer using an interpreter, etc.

[0208] According to one embodiment, the method according to various embodiments of the present disclosure may be provided by being included in a computer program product. The computer program product may be traded between a seller and a buyer as a product. The computer program product may be distributed in the form of a device-readable storage medium (e.g., compact disc read-only memory (CD-ROM)), or distributed online (e.g., download or upload) through an application store (e.g., Play Store™) or directly between two user devices. In the case of online distribution, at least a portion of the computer program product may be temporarily stored or temporarily created in a device-readable storage medium, such as the memory of a manufacturer's server, an application store's server, or a relay server.

[0209] Unless explicitly stated otherwise, the steps constituting the method according to the present disclosure may be performed in a suitable order. The present disclosure is not necessarily limited by the order in which the steps are described. The use of any examples or exemplary terms (e.g., etc.) in the present disclosure is merely for the purpose of describing the present disclosure in detail and, unless limited by the claims, the scope of the present disclosure is not limited by such examples or exemplary terms. Furthermore, those skilled in the art will understand that various modifications, combinations, and changes may be made according to design conditions and factors within the scope of the claims or equivalents to which they are added.

[0210] Accordingly, the scope of the present disclosure should not be limited to the embodiments described above, and all scopes equivalent to or equivalently modified from the claims set forth below, as well as the claims set forth below, shall be considered to fall within the scope of the scope of the present disclosure.

Claims

1. A method performed by a computing device comprising at least one processor, A step of receiving a risk situation detection signal from at least one device among a plurality of edge devices installed in a risk situation detection target space; A step of classifying the type and severity of the risk situation based on the risk situation detection signal; and A step of providing a warning notification based on a notification policy corresponding to the type and severity of the above-mentioned risk situation; including, Edge device-based method for detecting hazardous situations within a space.

2. In Paragraph 1, The above method is, A step of acquiring normal state data from each of the plurality of edge devices at preset intervals; A step of generating judgment criterion information for detecting a dangerous situation based on the above normal state data; and A step of transmitting the above judgment criteria information to each of the plurality of edge devices; including, Edge device-based method for detecting hazardous situations within a space.

3. In Paragraph 2, Each of the above plurality of edge devices is, Converting real-time collected monitoring data into input data for input into a pre-trained risk detection model, and The above input data and the above judgment criteria information are input into the above risk detection model to detect a risk situation, and The above risk detection model is, A model comprising at least one of a behavior-based risk detection model that detects dangerous situations based on video data, an acoustic-based risk detection model that detects dangerous situations based on acoustic data, and an environment-based risk detection model that detects dangerous situations based on environment data. Edge device-based method for detecting hazardous situations within a space.

4. In Paragraph 3, The above method is, When a risk situation detection signal is received from at least one device, a step of collecting monitoring data corresponding to the risk situation detection signal; A step of generating training data by labeling the above-mentioned monitoring data with the above-mentioned risk situation; and A step of transmitting the training data to other devices of the same type as the at least one device; Includes, Other devices of the same type as the above-mentioned at least one device are, Updating the risk detection model based on the above training data, Edge device-based method for detecting hazardous situations within a space.

5. In Paragraph 1, The step of classifying the type and severity of the risk situation based on the risk situation detection signal is, A step of recognizing the type of a specific device that transmitted the above-mentioned danger situation detection signal; A step of classifying the type of the risk situation based on the type of the specific device; and A step of classifying the severity of the risk situation based on monitoring data corresponding to the risk situation detection signal; including, Edge device-based method for detecting hazardous situations within a space.

6. In Paragraph 5, The step of classifying the types of dangerous situations based on the type of the specific device mentioned above is, A step of classifying the risk situation as an acoustic-based risk situation if the type of the specific device is a device type that collects acoustic data, classifying the risk situation as a visual-based risk situation if the type of the specific device is a device type that collects video data, and classifying the risk situation as an environment-based risk situation if the type of the specific device is a device type that collects environment data; Includes, The step of classifying the severity of the risk situation based on monitoring data corresponding to the risk situation detection signal is, A step of classifying the severity of the risk situation based on at least one of the intensity, duration, and rate of change of the signal included in the monitoring data; including, Edge device-based method for detecting hazardous situations within a space.

7. In Paragraph 1, The step of providing warning notifications based on a notification policy corresponding to the type and severity of the above-mentioned risk situation is: A step of sending a warning notification to an administrator terminal; and A step of outputting a warning notification through an electronic display or speaker provided in the above-mentioned danger situation detection target space; Includes at least one of the steps, The above notification policy is, A first notification policy that transmits the warning notification only to the administrator terminal according to the type of the above-mentioned risk situation and the severity of the above-mentioned risk situation, and A second notification policy that outputs the warning notification through the electronic display or speaker in conjunction with transmitting the warning notification to the administrator terminal, and controls the color of the warning notification output through the electronic display or the volume of the warning notification output through the speaker in response to the severity of the dangerous situation. Edge device-based method for detecting hazardous situations within a space.

8. In Paragraph 7, The step of outputting a warning notification through an electronic display or speaker provided in the space targeted for detecting the above-mentioned dangerous situation is: A step of obtaining map information of the space targeted for detecting the above-mentioned dangerous situation; A step of recognizing the point of occurrence of a dangerous situation based on the above dangerous situation detection signal; A step of recognizing an evacuation route that bypasses the point where the dangerous situation occurs based on the above map information; and A step of outputting a warning notification guiding the evacuation route based on a specific location where the above-mentioned electronic display or speaker is installed; including, Edge device-based method for detecting hazardous situations within a space.

9. Memory for storing one or more instructions; and A processor that executes one or more instructions stored in the memory. Including, The above processor executes the above one or more instructions, A device that performs the method of claim 1.

10. A computer program stored on a computer-readable recording medium that is combined with a computer, which is hardware, to perform the method of claim 1.

Images