Intelligent bus stop safety management system and method for protecting personal information

Abstract

The present disclosure relates to an intelligent bus stop safety management system and a method for protecting personal information. An intelligent bus stop safety management system according to one embodiment of the present disclosure may include a non-identification sensor unit that monitors a bus stop area, an identification sensor unit that captures the bus stop area to acquire identifiable evidence images, and a control unit linked with the non-identification sensor unit and the identification sensor unit. The control unit may be characterized by acquiring primary sensor data from the non-identification sensor unit, wherein the primary sensor data includes thermal image data and acoustic data, and by fusion analysis of the acquired thermal image data and acoustic data to calculate the risk level of the bus stop area, and when the calculated risk level exceeds a preset threshold, by conditionally activating the identification sensor unit in a continuous monitoring mode through the non-identification sensor unit to switch to an evidence securing mode.

Description

Technology Field

[0001] The present disclosure relates to an intelligent bus stop safety management system and method for protecting personal information, and more specifically, to a technology that accurately detects dangerous situations by fusion analysis of non-identifiable sensor data, conditionally activates an evidence securing device capable of personal identification, and prevents crimes and accidents while protecting privacy and personal information. Background Technology

[0002] Due to the nature of bus stops as public spaces used by an unspecified number of people 24 hours a day, they are susceptible to various social safety issues such as traffic accidents, violence, sexual crimes, homelessness, and falls. Citizens feel particularly anxious as the likelihood of becoming a target of crime is high, especially during late-night hours when there are few people around.

[0003] To address these issues, CCTVs (closed-circuit televisions) were previously installed and utilized as post-incident evidence. However, since CCTVs record identification information, including personal appearance, 24 hours a day, there are serious legal and social problems regarding the excessive infringement of personal information and privacy of ordinary citizens using bus stops.

[0004] Recently, there have been attempts to detect abnormal behavior in CCTV footage by incorporating AI; however, this approach also suffers from the inherent limitation of constantly collecting personally identifiable information, and the accuracy of video analysis significantly deteriorates at night when lighting is poor. Furthermore, sensors that detect only sound or movement have faced difficulties in ensuring system reliability due to high false positive rates, which misidentify non-hazardous situations—such as bus noise or the shouting of intoxicated individuals—as dangerous.

[0005] Meanwhile, sensor technologies that detect only sound or movement are being introduced, but these also have clear limitations. It is difficult to ensure system reliability due to a high false positive rate, where non-hazardous situations—such as bus noise or the loud shouting of intoxicated individuals—are mistaken for dangerous ones. These technical limitations further highlight the need for new technological solutions that can ensure public safety, thoroughly protect citizens' personal information, and accurately assess dangerous situations even amidst diverse environmental variables.

[0006] Therefore, there is an urgent need for a new technological solution that can ensure public safety while thoroughly protecting citizens' personal information and accurately assessing dangerous situations even amidst various environmental variables. The problem to be solved

[0007] The technical problem to be solved through the embodiments of the present disclosure is to solve the problems of the prior art as described above, and to accurately detect various dangerous situations such as assault, falling, and threats with weapons 24 hours a day without constantly collecting personal identification information (face, appearance, etc.) of bus stop users.

[0008] In addition, the technical problem to be solved through the embodiments of the present disclosure is to overcome the limitations of a single sensor (e.g., inability to identify at night, malfunction due to noise) by fusion analyzing non-identifiable sensor data of multiple modalities, such as thermal imaging sensors and acoustic sensors, and to dramatically improve the accuracy and reliability of risk assessment.

[0009] The technical problem to be solved by the embodiments of the present disclosure is to provide a technical procedure that activates an evidence securing device (CCTV) capable of conditional personal identification only when a real dangerous situation is determined through non-identifiable sensor analysis, and to protect third-party information other than that of the relevant parties during this process.

[0010] The technical problem to be solved through the embodiments of the present disclosure is to prevent crimes and accidents by detecting potential risk factors in advance, such as possession of weapons, loitering, and prolonged occupation, and responding actively (through guidance broadcasts and linked safe return home services).

[0011] The technical problems of the present disclosure are not limited to those mentioned above, and other unmentioned technical problems will be clearly understood by a person skilled in the art of the present disclosure from the description below. means of solving the problem

[0012] In one embodiment of the present disclosure, an intelligent bus stop safety management system may be provided. The intelligent bus stop safety management system may include a non-identification sensor unit that monitors a bus stop area, an identification sensor unit that captures the bus stop area to acquire identifiable evidence images, and a control unit linked with the non-identification sensor unit and the identification sensor unit. The control unit acquires primary sensor data from the non-identification sensor unit, wherein the primary sensor data includes thermal image data and acoustic data, and calculates the risk level of the bus stop area by fusion analysis of the acquired thermal image data and acoustic data. When the calculated risk level exceeds a preset threshold, the control unit may switch from a continuous monitoring mode through the non-identification sensor unit to an evidence securing mode by conditionally activating the identification sensor unit.

[0013] In one embodiment of the present disclosure, the control unit estimates the skeleton of a human body from the thermal image data, calculates a motion vector based on time-series data for a plurality of joints constituting the skeleton, and determines whether the calculated motion vector matches a predefined pattern of dangerous behavior and reflects this in the risk calculation.

[0014] In one embodiment of the present disclosure, the non-identification sensor unit may include a multi-microphone array composed of a plurality of microphones. When the acoustic data is detected, the control unit may identify the location of the sound source of the acoustic data by calculating the difference in sound arrival times reaching each microphone of the multi-microphone array.

[0015] In one embodiment of the present disclosure, the control unit may regard the acoustic data as external noise when the specified sound source generation location is determined to be outside the bus stop area.

[0016] In one embodiment of the present disclosure, the control unit may increase the weight of the risk level when the specified sound source generation location and the location of the skeleton match within a preset distance.

[0017] In one embodiment of the present disclosure, the control unit can calculate the risk level through a weight-based fusion model that calculates a first reliability score from the thermal image data analysis result, calculates a second reliability score from the acoustic data analysis result, and applies a pre-set weight to the first reliability score and the second reliability score, respectively.

[0018] In one embodiment of the present disclosure, when the identification sensor unit is activated, the control unit identifies a person related to the skeleton based on the thermal image data that caused the risk calculation, and can mask the body or face area of ​​a third party excluding the person related in real time from the evidence image acquired by the identification sensor unit.

[0019] In one embodiment of the present disclosure, the control unit immediately encrypts the unmasked original evidence image separately from the masked image, stores it in a separate storage space, and can decrypt it only according to a pre-set security procedure.

[0020] In one embodiment of the present disclosure, the control unit is configured to analyze the thermal silhouette around the hand joints of a human skeleton from the thermal image data and determine whether it matches a preset silhouette pattern of a dangerous object, and if the dangerous object is detected, the risk level can be immediately raised above the threshold value.

[0021] In one embodiment of the present disclosure, the intelligent bus stop safety management system may further include an NFC module that communicates with a user terminal and a speaker that outputs information as voice. The control unit assigns at least one safety grade in real time according to the calculated risk level, and when a service is requested through the NFC module, it may transmit a safety status message including the real-time safety grade to the user terminal or a pre-registered guardian terminal.

[0022] In one embodiment of the present disclosure, an intelligent bus stop safety management system may be provided. The intelligent bus stop safety management system may include a non-identification sensor unit that monitors a bus stop area, an identification sensor unit that captures the bus stop area to acquire identifiable evidence images, and a control unit linked with the non-identification sensor unit and the identification sensor unit. The control unit acquires primary sensor data from the non-identification sensor unit, wherein the primary sensor data includes thermal image data and acoustic data, and calculates the risk level of the bus stop area by fusion analysis of the acquired thermal image data and acoustic data. When the calculated risk level exceeds a preset threshold, the control unit may switch from a continuous monitoring mode through the non-identification sensor unit to an evidence securing mode by conditionally activating the identification sensor unit.

[0023] In one embodiment of the present disclosure, the control unit estimates a human body skeleton from the thermal image data, calculates a motion vector based on time-series data for a plurality of joints constituting the skeleton, and determines whether the calculated motion vector matches a predefined pattern of dangerous behavior and reflects this in the risk calculation.

[0024] In one embodiment of the present disclosure, the non-identification sensor unit may include a multi-microphone array composed of a plurality of microphones. When the acoustic data is detected, the control unit may identify the location of the sound source of the acoustic data by calculating the difference in sound arrival times reaching each microphone of the multi-microphone array.

[0025] In one embodiment of the present disclosure, the control unit may regard the acoustic data as external noise when the specified sound source generation location is determined to be outside the bus stop area.

[0026] In one embodiment of the present disclosure, the control unit may increase the weight of the risk level when the specified sound source generation location and the location of the skeleton match within a preset distance.

[0027] In one embodiment of the present disclosure, the control unit can calculate the risk level through a weight-based fusion model that calculates a first reliability score from the thermal image data analysis result, calculates a second reliability score from the acoustic data analysis result, and applies a pre-set weight to the first reliability score and the second reliability score, respectively.

[0028] In one embodiment of the present disclosure, the control unit can identify a person related to the skeleton based on the thermal image data that caused the risk calculation when the identification sensor unit is activated, and can mask the body or face area of ​​a third party excluding the person related in real time from the evidence image acquired by the identification sensor unit.

[0029] In one embodiment of the present disclosure, the control unit immediately encrypts the unmasked original evidence image separately from the masked image, stores it in a separate storage space, and can decrypt it only according to a pre-set security procedure.

[0030] In one embodiment of the present disclosure, the control unit is configured to analyze the thermal silhouette around the hand joints of a human skeleton from the thermal image data and determine whether it matches a preset silhouette pattern of a dangerous object, and if the dangerous object is detected, the risk level can be immediately raised above the threshold value.

[0031] In one embodiment of the present disclosure, the intelligent bus stop safety management system may further include an NFC module that communicates with a user terminal and a speaker that outputs information as voice. The control unit assigns at least one safety grade in real time according to the calculated risk level, and when a service is requested through the NFC module, it may transmit a safety status message including the real-time safety grade to the user terminal or a pre-registered guardian terminal.

[0032] In one embodiment of the present disclosure, the control unit can transmit an automatic guidance broadcast regarding the dangerous behavior through the speaker when the risk level corresponds to a preset grade and the identification sensor unit is deactivated. Effects of the invention

[0033] According to the embodiments of the present disclosure, only a thermal imaging camera and a microphone that cannot identify individuals are operated continuously for four hours, thereby perfectly protecting the privacy and personal information of bus stop users.

[0034] In addition, according to the embodiments of the present disclosure, by fusing motion vector analysis of a thermal image skeleton with sound source location analysis of a multi-microphone array, external noise can be filtered without being affected by night or bad weather, thereby drastically reducing the false detection rate.

[0035] In addition, according to the embodiments of the present disclosure, potential dangers such as assault and falling, as well as possession of weapons and prolonged occupation, are detected in advance, and immediate intervention is provided through automatic guidance broadcasts, etc., resulting in excellent crime and accident prevention effects.

[0036] In addition, according to the embodiments of the present disclosure, the high-resolution CCTV is activated only when an actual dangerous situation occurs, and even then, third parties other than the relevant parties are 'dynamically privacy masked' to secure legal evidence while preventing secondary infringement of personal information.

[0037] In addition, according to embodiments of the present disclosure, the real-time safety rating of a bus stop can be linked with the user's 'safe return home service' to provide active safety information rather than simple bus arrival information. Brief explanation of the drawing

[0038] FIG. 1 is a drawing illustrating an implementation example of an intelligent bus stop safety management system according to various embodiments. FIG. 2 is a diagram illustrating the configuration of an intelligent bus stop safety management system according to various embodiments. FIG. 3 is a flowchart illustrating the operation method of an intelligent bus stop safety management system according to various embodiments. FIG. 4 is a diagram illustrating the operation of the components of an intelligent bus stop safety management system according to various embodiments. FIG. 5 is a flowchart illustrating a method for analyzing thermal image data according to various embodiments. FIG. 6 is a flowchart illustrating a method for calculating risk according to various embodiments. FIG. 7 is a flowchart illustrating a safety management method through the detection of hazardous objects according to various embodiments. FIG. 8 is a flowchart illustrating a method for providing a safe return home service according to various embodiments. Specific details for implementing the invention

[0039] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the accompanying drawings. In describing the embodiments, technical details that are well known in the art to which the present disclosure belongs and are not directly related to the present disclosure will be omitted. This is intended to convey the essence of the present disclosure more clearly without obscuring it by omitting unnecessary explanations.

[0040] The embodiments described in this specification are intended to clearly explain the concept of the invention to those skilled in the art to which the invention pertains, and therefore the invention is not limited to the embodiments described in this specification, and the scope of the invention should be interpreted to include modifications or variations that do not deviate from the concept of the invention.

[0041] The terms used in this specification have been selected to be as widely used as possible, taking into account their functions in the present invention; however, these terms may vary depending on the intent of those skilled in the art to which the present invention pertains, case law, or the emergence of new technologies. However, if a specific term is defined and used with an arbitrary meaning, the meaning of that term will be described separately. Accordingly, the terms used in this specification should be interpreted based on their actual meaning and the content throughout this specification, rather than merely their names.

[0042] The drawings attached to this specification are intended to facilitate the explanation of the present invention. The shapes depicted in the drawings may be exaggerated as necessary to aid in understanding the present invention, and therefore the present invention is not limited by the drawings.

[0043] In this specification, each of the phrases such as “A or B”, “at least one of A and B”, “at least one of A or B”, “A, B or C”, “at least one of A, B and C”, and “at least one of A, B, or C” may include any one of the items listed together in the corresponding phrase, or all possible combinations thereof.

[0044] In cases where it is determined that a detailed description of known components or functions related to the present invention in this specification may obscure the essence of the invention, such detailed description shall be omitted as necessary. Furthermore, numbers used in the description of this specification (e.g., First, Second, etc.) are merely identification symbols to distinguish one component from another.

[0045] Furthermore, the suffixes "part" and "substance" used for components in the following description are assigned or used interchangeably solely for the ease of drafting the specification, and do not inherently possess distinct meanings or roles.

[0046] That is, the embodiments of the present disclosure are provided to make the present disclosure complete and to inform those skilled in the art of the scope of the present disclosure, and the invention of the present disclosure is defined only by the scope of the claims. Throughout the specification, the same reference numerals refer to the same components.

[0047] Terms such as “first” and / or “second” may be used to describe various components, but said components shall not be limited by said terms. For the sole purpose of distinguishing one component from another, for example, without departing from the scope of rights according to the concept of the present disclosure, the first component may be named the second component, and similarly, the second component may be named the first component.

[0048] When it is stated that one component is "connected" or "connected" to another component, it should be understood that while it may be directly connected or connected to that other component, there may also be other components in between. Conversely, when it is stated that one component is "directly connected" or "directly connected" to another component, it should be understood that there are no other components in between. Other expressions describing the relationship between components, such as "between" and "exactly between," or "adjacent to" and "directly adjacent to," should be interpreted in the same way.

[0049] In the drawings, each block of the process flow diagrams and combinations of the flow diagrams can be executed by computer program instructions. Since these computer program instructions can be loaded into the processor of a general-purpose computer, a special-purpose computer, or other programmable data processing equipment, the instructions executed through the processor of the computer or other programmable data processing equipment create means to perform the functions described in the flow diagram block(s). Since these computer program instructions can also be stored in computer-available or computer-readable memory that can be directed toward the computer or other programmable data processing equipment to implement the function in a specific way, the instructions stored in computer-available or computer-readable memory can also produce a manufactured item containing means of instruction to perform the function described in the flow diagram block(s). Since computer program instructions can be loaded onto a computer or other programmable data processing equipment, instructions that execute a computer or other programmable data processing equipment by performing a series of operation steps on the computer or other programmable data processing equipment to create a process executed by the computer may also provide steps for executing the functions described in the flowchart block(s).

[0050] Additionally, device-readable storage media may be provided in the form of non-transitory storage media. Here, 'non-transitory' simply means that the storage medium is a tangible device and does not contain a signal (e.g., electromagnetic waves), and this term does not distinguish between cases where data is stored semi-permanently and cases where it is stored temporarily.

[0051] Additionally, each block may represent a module, segment, or part of code containing one or more executable instructions for executing a specified logical function(s). It should also be noted that in some alternative execution examples, the functions mentioned in the blocks may occur out of order. For instance, two blocks described in succession may actually be executed substantially simultaneously, or the blocks may be executed in reverse order depending on the corresponding function. For instance, actions performed by a module, program, or other component may be executed sequentially, in parallel, iteratively, or heuristically; one or more of these actions may be executed in a different order, omitted, or one or more other actions may be added.

[0052] As used in this disclosure, the term “unit” refers to a software or hardware component such as a Field Programmable Gate Array (FPGA) or an Application Specific Integrated Circuit (ASIC). A “unit” performs specific roles but is not limited to software or hardware. A “unit” may be configured to reside in an addressable storage medium or may be configured to run one or more processors. Accordingly, according to some embodiments, a “unit” includes components such as software components, object-oriented software components, class components, and task components, as well as processes, functions, attributes, procedures, subroutines, segments of program code, drivers, firmware, microcode, circuits, data, databases, data structures, tables, arrays, and variables. The functions provided within the components and “units” may be combined into a smaller number of components and “units” or further separated into additional components and “units.” In addition, the components and 'parts' may be implemented to utilize one or more CPUs within the device or secure multimedia card. Furthermore, according to various embodiments of the present disclosure, the 'parts' may include one or more processors.

[0053] The operating principles of the present disclosure will be described in detail below with reference to the attached drawings. In describing the present disclosure below, specific descriptions of related known functions or configurations will be omitted if it is determined that such detailed descriptions would unnecessarily obscure the essence of the present disclosure. Furthermore, the terms described below are defined in consideration of their functions in the present disclosure, and these may vary depending on the intentions or practices of the user or operator. Therefore, their definitions should be based on the content throughout this specification.

[0054] A 'non-identification sensor unit' may refer to a sensor that collects data without identifying specific targets. Non-identification sensor units can primarily collect environmental information while protecting privacy. Non-identification sensor units monitor bus stop areas and can collect thermal and acoustic data. Non-identification sensor units operate in continuous monitoring mode and can contribute to activating identification sensor units if the risk level exceeds a threshold.

[0055] The 'identification sensor unit' may refer to a sensor that identifies a specific target or collects specific information. The identification sensor unit may include a high-resolution camera or biometric recognition technology. The identification sensor unit can acquire evidence footage by photographing the bus stop area. The identification sensor unit can be conditionally activated based on the data analysis results of the non-identification sensor unit to switch to an evidence acquisition mode.

[0056] 'Thermal imaging data' refers to temperature distribution data collected through an infrared camera. Thermal imaging data can primarily be used to analyze the thermal characteristics of the human body or objects. In bus stop areas, thermal imaging data can be utilized to estimate the human skeleton or detect hazardous objects. Thermal imaging data can be analyzed by a control unit to contribute to calculating risk levels.

[0057] "Acoustic data" may refer to data including the frequency, amplitude, and temporal characteristics of sound. Acoustic data can be used to analyze the location or pattern of sound generation. Acoustic data is collected through a multi-microphone array, and the control unit can analyze it to identify the location of the sound source. Acoustic data can be fused with thermal imaging data for analysis to calculate risk levels.

[0058] 'Continuous monitoring mode' may refer to a state of continuously monitoring the environment. Continuous monitoring mode can primarily be implemented through non-identifiable sensors. It may indicate a state where non-identifiable sensor units operate to continuously collect data from the bus stop area. Continuous monitoring mode may switch to evidence acquisition mode if the risk level exceeds a threshold.

[0059] 'Evidence Acquisition Mode' may refer to a state that operates to collect evidence in specific situations. Evidence Acquisition Mode can primarily be implemented through identification sensors. Evidence Acquisition Mode may be conditionally activated based on the data analysis results of non-identification sensor units. Evidence Acquisition Mode can be used by identification sensor units to acquire, store, or analyze evidence images.

[0060] A 'skeleton' can refer to a model that simplifies and represents the structure of the human body's joints and bones. Skeletons are primarily utilized for motion analysis or posture recognition. They are estimated based on thermal imaging data, and motion vectors can be calculated using time-series data of the joints. Skeletons can be incorporated into risk calculations by determining whether they match patterns of dangerous behavior.

[0061] 'Motion vector' refers to vector data representing the movement of an object. Motion vectors can primarily be used in motion analysis or image processing. Motion vectors can be calculated based on time-series data of a skeleton's joints. Motion vectors can contribute to the calculation of risk by being compared with predefined risk behavior patterns.

[0062] A 'weighted fusion model' can refer to an analysis model that derives results by reflecting the importance of various data. A weighted fusion model can enhance reliability by integrating diverse data sources. It can calculate risk levels by applying weights to the reliability scores of thermal imaging data and acoustic data, respectively. A weighted fusion model can be used to improve the accuracy of risk analysis in control units.

[0063] 'Encryption' can refer to the process of transforming data using a specific algorithm to protect it. Encryption can maintain data confidentiality and prevent unauthorized access. Encryption can be used to immediately protect unmasked original evidence footage. Encrypted footage is stored in a separate storage space and can be decrypted according to security procedures.

[0064] 'Decryption' can refer to the process of converting encrypted data back into its original form. Decryption can primarily be used to restrict access to data so that only authenticated users can access it. Decryption can be used to allow access to encrypted original evidence footage in accordance with security procedures. Decryption can operate as part of security procedures managed by a control unit.

[0065] "User terminal" may refer to an electronic device capable of transmitting and receiving information. User terminals may include smartphones, tablets, laptops, etc. User terminals can receive real-time safety status messages from bus stops via an NFC module. User terminals can also receive safety status messages via a pre-registered guardian terminal.

[0066] An 'NFC module' can refer to hardware that supports near-field wireless communication. NFC modules can primarily be used to transmit and receive data over short distances. An NFC module can communicate with a user terminal to transmit messages including real-time safety ratings. NFC modules can be utilized to provide safety status information upon a service request.

[0067] A 'speaker' can refer to a device that outputs sound. Speakers are primarily used to deliver voice guidance or warning sounds. Speakers can transmit automatic guidance broadcasts when the risk level corresponds to a specific grade. Speakers can deliver warning messages regarding dangerous behaviors according to commands from a control unit.

[0068] "Automatic guidance broadcast" may refer to an announcement automatically transmitted in specific situations. Automatic guidance broadcasts can primarily provide safety warnings or guidelines. Automatic guidance broadcasts may be transmitted when the risk level corresponds to a specific grade and the identification sensor unit is deactivated. Automatic guidance broadcasts can convey warnings about dangerous behavior through speakers.

[0069] FIG. 1 is a drawing illustrating an implementation example of an intelligent bus stop safety management system according to various embodiments.

[0070] Referring to FIG. 1, an intelligent bus stop safety management system (1) can be installed and operated in a bus stop area (R). The intelligent bus stop safety management system (1) can perform the role of detecting and responding to various safety threats that may occur in the bus stop area (R) in real time. The intelligent bus stop safety management system (1) performs constant monitoring in a manner that does not infringe on personal information, and can balance safety and privacy protection by activating the evidence securing function only when a dangerous situation occurs.

[0071] The intelligent bus stop safety management system (1) may be an integrated system designed to enhance the safety of the bus stop. The intelligent bus stop safety management system (1) may be configured to include a non-identification sensor unit, an identification sensor unit, and a control unit. Under normal circumstances, the safety status of the bus stop area (R) can be continuously monitored by collecting and analyzing data that does not identify individuals through a non-identification sensor unit, such as a thermal imaging camera or a microphone.

[0072] The intelligent bus stop safety management system (1) can quantitatively calculate the risk level through the fusion analysis of non-identifiable sensor data. For example, as illustrated in FIG. 1, if a situation occurs where a specific person screams or shouts loudly, the intelligent bus stop safety management system (1) can detect the sound through an acoustic sensor. At the same time, the intelligent bus stop safety management system (1) can analyze abnormal behavior patterns, such as violent movements or falling of the person, through a thermal imaging sensor, and comprehensively determine the two data to calculate the risk level. If the calculated risk level exceeds a preset threshold, the intelligent bus stop safety management system (1) can determine this as a dangerous situation and initiate the next step of the response procedure.

[0073] In one embodiment of the present disclosure, the intelligent bus stop safety management system (1) may be implemented in various forms such as a kiosk, a pole, digital signage, a smart shelter, a standalone surveillance tower, a wall-mounted unit, or a ceiling-embedded device. However, the present disclosure is not limited thereto.

[0074] The bus stop area (R) may refer to the target space where the intelligent bus stop safety management system (1) performs safety management. The bus stop area (R) can generally be defined as a zone including passenger waiting chairs, bus boarding and alighting points, the area around bus information boards, and a portion of an adjacent pedestrian walkway. The intelligent bus stop safety management system (1) can perform monitoring by focusing on objects and events within this bus stop area (R).

[0075] The intelligent bus stop safety management system (1) can prevent system malfunctions by clearly recognizing the boundaries of the bus stop area (R). For example, the location of sound generation can be tracked using a microphone array composed of multiple microphones. If the detected noise is determined to be vehicle noise or construction noise originating from a road outside the bus stop area (R), the intelligent bus stop safety management system (1) can classify it as external noise and exclude it from the risk calculation or apply a lower weight. This spatial filtering function can contribute to improving the accuracy of the system's judgment.

[0076] FIG. 2 is a diagram illustrating the configuration of an intelligent bus stop safety management system according to various embodiments.

[0077] Referring to FIG. 2 together with FIG. 1, details regarding the intelligent bus stop safety management system (1) that overlap with those described in FIG. 1 may be omitted. The intelligent bus stop safety management system (1) may include components for managing the safety of a bus stop area (R). For example, the intelligent bus stop safety management system (1) may include a non-identification sensor unit (100), an identification sensor unit (200), a control unit (300), an output unit (400), and a user interface unit (500).

[0078] The non-identification sensor unit (100) can collect information that cannot identify individuals and continuously monitor the bus stop area (R). The non-identification sensor unit (100) can perform primary detection. The non-identification sensor unit (100) can operate 24 hours a day without concerns about infringing on personal information. The non-identification sensor unit (100) may include a thermal imaging camera (110) and a multi-microphone array (120).

[0079] A thermal imaging camera (110) can detect infrared energy emitted from objects within a bus stop area (R). The thermal imaging camera (110) can generate thermal image data through the detected infrared energy. The thermal imaging camera (110) may not acquire identification information such as an individual's face or clothing. The thermal imaging camera (110) can acquire only temperature and silhouette information. Therefore, the thermal imaging camera (110) can effectively detect the number of people, their location, and their shape even in adverse conditions such as night, smoke, or fog.

[0080] A multi-microphone array (120) can collect acoustic data generated within a bus stop area (R). The multi-microphone array (120) may be composed of multiple microphones. For example, the multi-microphone array (120) may include at least three microphones. The multi-microphone array (120) can analyze the time difference of sound reaching each microphone. The control unit (300) can use this time difference information to precisely determine the direction and location of the sound source.

[0081] The identification sensor unit (200) can be conditionally operated to acquire evidence video capable of personal identification. The identification sensor unit (200) can remain in a deactivated state under normal circumstances. The identification sensor unit (200) can prevent invasion of privacy. The identification sensor unit (200) can be activated according to a control signal from the control unit (300). The identification sensor unit (200) can be activated to perform secondary detection. The identification sensor unit (200) may include a CCTV (210).

[0082] The CCTV (210) can capture high-resolution color optical images. The CCTV (210) can start recording only when it receives an activation command from the control unit (300). The images obtained through the CCTV (210) can be used as legal evidence in the event of a crime or accident.

[0083] The control unit (300) can serve as a central processing unit that controls the overall operation of the intelligent bus stop safety management system (1). The control unit (300) may include a processor, memory, and a communication module. The control unit (300) may be electrically connected to a non-identification sensor unit (100), an identification sensor unit (200), an output unit (400), and a user interface unit (500). The control unit (300) may transmit and receive signals to and from each unit.

[0084] The control unit (300) can acquire primary sensor data from the non-identifiable sensor unit (100). The control unit (300) can calculate the risk level by fusion analyzing the acquired data. For example, the control unit (300) can estimate the skeleton of a human body from thermal image data. The control unit (300) can identify the location of a sound source from acoustic data. The control unit (300) can determine whether there is a dangerous situation in the bus stop area (R) based on the analyzed information. The control unit (300) can assign at least one safety level in real time according to the calculated risk level. In one embodiment of the present disclosure, the safety level may be divided into multiple stages such as 'Safe', 'Caution', 'Warning', and 'Danger'. However, the present disclosure is not limited thereto.

[0085] The control unit (300) may switch the operating mode of the system when the calculated risk level exceeds a preset threshold. For example, the control unit (300) may switch from a normal continuous monitoring mode to an evidence securing mode that activates the identification sensor unit (200). In one embodiment of the present disclosure, the control unit (300) may be implemented as an edge computing device located at a bus stop terminal. Alternatively, the control unit (300) may be implemented on a remote central control server (Cloud).

[0086] The output unit (400) can provide visual or auditory information to the user based on the judgment result of the control unit (300). The output unit (400) can perform the role of warning of dangerous situations or providing necessary information. The output unit (400) may include a speaker (410), a lighting device (420), and a DID (430).

[0087] The speaker (410) can output voice information under the control of the control unit (300). For example, the speaker (410) can transmit an automatic guidance broadcast regarding dangerous behavior when the risk level corresponds to a preset grade and the identification sensor unit (200) is deactivated. The automatic guidance broadcast may be a message such as "This is not a smoking area" or "Dangerous behavior has been detected." Additionally, the speaker (410) may transmit real-time voice messages delivered by a controller at a remote control center.

[0088] The lighting device (420) can provide a visual warning using light. The lighting device (420) can flash a warning light when a dangerous situation is detected. The lighting device (420) can attract the attention of the surroundings through flashing. In addition, the lighting device (420) can increase the brightness of the light. The lighting device (420) can secure a field of view for filming by the CCTV (210). The lighting device (420) can induce an effect that intimidates the psychology of potential criminals.

[0089] The DID (430) may be a Digital Information Display. The DID (430) may normally display lifestyle information such as bus arrival information, weather, and news. In the event of a crisis, the screen of the DID (430) may be switched according to the control of the control unit (300). The switched screen may display safety-related information such as evacuation instructions and warning messages.

[0090] The user interface unit (500) can provide a means for a bus stop user to interact with the intelligent bus stop safety management system (1). The user interface unit (500) can enable direct requests or service usage by the user. The user interface unit (500) may include a safety bell (510) and an NFC module (520).

[0091] The safety bell (510) may be a physical button that the user can press directly in an emergency situation. When the user presses the safety bell (510), the control unit (300) can immediately determine the situation to the highest risk level. Accordingly, the control unit (300) can activate the identification sensor unit (200). The control unit (300) can transmit the emergency situation to the control center.

[0092] The NFC module (520) may be a module for near-field communication. The NFC module (520) may provide a connection service with a user terminal, such as a user's smartphone. For example, when a service is requested through the NFC module (520), the control unit (300) may transmit a safety status message including a real-time safety level to the user terminal or a pre-registered guardian terminal.

[0093] FIG. 3 is a flowchart illustrating the operation method of an intelligent bus stop safety management system according to various embodiments.

[0094] Referring to FIG. 3 in conjunction with FIG. 2, the intelligent bus stop safety management system (1) may include a non-identification sensor unit (100), an identification sensor unit (200), and a control unit (300). The non-identification sensor unit (100) can monitor a bus stop area (R). The identification sensor unit (200) can photograph the bus stop area (R) to obtain identifiable evidence images. The control unit (300) can be linked with the non-identification sensor unit (100) and the identification sensor unit (200). The operation method of the intelligent bus stop safety management system (1) can be performed by the control unit (300).

[0095] In step S310, the control unit (300) can acquire primary sensor data from the non-identification sensor unit (100). The non-identification sensor unit (100) can continuously operate in an always-on monitoring mode to monitor the bus stop area (R). The primary sensor data may include thermal image data acquired from a thermal imaging camera (110) and acoustic data acquired from a multi-microphone array (120). By collecting non-identification data that cannot identify an individual, the control unit (300) can perform 24-hour safety monitoring without concerns regarding privacy infringement.

[0096] In one embodiment of the present disclosure, the control unit (300) may obtain primary sensor data from a non-identifiable sensor unit (100) that monitors a bus stop area (R). The primary sensor data may include thermal image data and acoustic data that cannot identify an individual.

[0097] In step S320, the control unit (300) can calculate the risk level of the bus stop area (R) by fusion-analyzing the acquired thermal image data and acoustic data. The control unit (300) can analyze patterns such as human shape, movement, and posture from the thermal image data. Additionally, the control unit (300) can analyze patterns of specific sounds, such as screaming, verbal abuse, and impact sounds, from the acoustic data. By synthesizing the two types of heterogeneous data, the control unit (300) can more accurately determine complex risk situations that are difficult to judge with a single sensor alone.

[0098] In one embodiment of the present disclosure, the control unit (300) can calculate the risk level as a quantitative value by combining the results of thermal image data analysis and acoustic data analysis. For example, if a scream occurs simultaneously at a location where a falling motion is detected, the control unit (300) can calculate the risk level as high.

[0099] In step S330, the control unit (300) can switch to an evidence securing mode by conditionally activating the identification sensor unit (200) from the continuous monitoring mode via the non-identification sensor unit (100) when the risk level exceeds a preset threshold. The control unit (300) can activate the identification sensor unit (200) only when the calculated risk level exceeds a predefined threshold. The CCTV (210) of the identification sensor unit (200) can capture high-resolution video to secure clear evidence of the incident. This conditional activation method can protect personal information during normal times. Additionally, the conditional activation method can balance safety and privacy protection by collecting necessary information only when an emergency occurs.

[0100] In one embodiment of the present disclosure, the control unit (300) may terminate the continuous monitoring mode and initiate the evidence acquisition mode when the calculated risk level exceeds a threshold value. In the evidence acquisition mode, the control unit (300) may activate the identification sensor unit (200) to acquire identifiable evidence images. Additionally, the control unit (300) may automatically establish communication with an external security system, such as a police station or a control center. The control unit (300) may transmit a real-time video stream or recorded evidence images to the security system to induce a rapid response.

[0101] In one embodiment of the present disclosure, step S330 may further include the step of controlling the output unit (400) to output a warning signal while the control unit (300) switches to an evidence securing mode. For example, the control unit (300) may broadcast a warning message through a speaker (410). Alternatively, the control unit (300) may flash a lighting device (420) to draw attention to the surroundings.

[0102] FIG. 4 is a diagram illustrating the operation of the components of an intelligent bus stop safety management system according to various embodiments.

[0103] Referring to FIG. 4, a diagram may be illustrated to explain the operation of the components of an intelligent bus stop safety management system (1). Regarding the non-identification sensor unit (100) and the identification sensor unit (200), details that overlap with those described in FIG. 2 may be omitted. The non-identification sensor unit (100) can transmit thermal image data and acoustic data to the control unit (300). The control unit (300) can analyze the received data to calculate the risk level. If the calculated risk level exceeds a preset threshold, the control unit (300) can switch to an evidence securing mode by conditionally activating the identification sensor unit (200) in a continuous monitoring mode through the non-identification sensor unit (100).

[0104] The control unit (300) can control the overall operation of the system by linking with the non-identification sensor unit (100) and the identification sensor unit (200). The control unit (300) can calculate the risk level of the bus stop area by acquiring primary sensor data from the non-identification sensor unit (100). The primary sensor data may include thermal image data and acoustic data. The control unit (300) may include a data analysis unit (310), a fusion model (320), a risk object detection module (330), and a privacy masking module (340).

[0105] The data analysis unit (310) can analyze thermal image data and acoustic data obtained from the non-identifiable sensor unit (100), respectively. The data analysis unit (310) can perform a deep learning-based thermal image skeleton estimation algorithm for the analysis of thermal image data. The data analysis unit (310) can generate a human body skeleton by estimating the positions of major joints, such as the head, shoulders, elbows, pelvis, and knees, within the thermal image silhouette without personal identification information. Subsequently, the data analysis unit (310) can calculate motion vectors representing the direction and speed of movement of each joint over time. The data analysis unit (310) can classify acts such as assault, falling, and loitering by comparing the calculated motion vectors with predefined motion vector patterns of dangerous acts.

[0106] In one embodiment of the present disclosure, the non-identification sensor unit (100) may include a multi-microphone array composed of a plurality of microphones. A data analysis unit (310) may process sound received from the multi-microphone array for acoustic data analysis. The data analysis unit (310) may detect specific acoustic events such as screams, explosions, and glass breaking sounds. When an acoustic event is detected, the data analysis unit (310) may calculate the Time Difference of Arrival (TDOA) of the sound reaching each microphone. The data analysis unit (310) may determine the three-dimensional location of the sound source through the calculated Time Difference of Arrival. If the data analysis unit (310) determines that the determined location of the sound source is outside the bus stop area, it may filter the acoustic data by considering it as external noise.

[0107] The fusion model (320) can calculate the final risk level by synthesizing the analysis results of the data analysis unit (310). The fusion model (320) can calculate a first reliability score from the thermal image data analysis results. Additionally, the fusion model (320) can calculate a second reliability score from the acoustic data analysis results. The fusion model (320) can calculate the final risk level through a weight-based fusion model that applies a pre-set weight to the first reliability score and the second reliability score, respectively. In one embodiment of the present disclosure, the fusion model (320) can increase the risk weight when the location of a specific sound source and the location of a skeleton match within a pre-set distance. For example, if a motion vector suspected of assault and a scream are detected simultaneously, and the location of the sound source and the location of the skeleton match, the fusion model (320) can calculate the risk level as 'severe' grade.

[0108] The danger object detection module (330) can detect whether a danger object is being possessed based on thermal imaging data. The danger object detection module (330) can intensively analyze the thermal imaging silhouette around the hand joints of the human skeleton estimated by the data analysis unit (310). The danger object detection module (330) can determine whether the analyzed silhouette matches a predefined silhouette pattern of a danger object. For example, the danger object may be a knife, a blunt instrument, a fuel container, a firearm, an explosive, a chemical container, or a laser pointer. However, the present disclosure is not limited thereto. When a danger object is detected, the control unit (300) can immediately raise the risk level above a threshold value and initiate a response procedure regardless of the reliability of other sensor data.

[0109] The privacy masking module (340) can perform the function of protecting the personal information of third parties other than the relevant parties when the identification sensor unit (200) is activated. The control unit (300) can activate the identification sensor unit (200) only when the calculated risk level exceeds a preset threshold. At this time, the privacy masking module (340) can identify the relevant parties corresponding to the skeleton based on the thermal image data that caused the risk level calculation. The privacy masking module (340) can mask the body or face area of ​​all third parties in real time, excluding the identified relevant parties, in the evidence image acquired by the identification sensor unit (200).

[0110] In one embodiment of the present disclosure, the privacy masking module (340) can immediately encrypt the unmasked original evidence image and store it in a separate storage space, separate from the masked image. The encrypted original evidence image can be decrypted only according to a pre-established security procedure, such as the presentation of a warrant by the police. Through this dual storage and access control method, the privacy masking module (340) can prevent secondary infringement of personal information while ensuring the integrity of the evidence.

[0111] FIG. 5 is a flowchart illustrating a method for analyzing thermal image data according to various embodiments.

[0112] Referring to FIG. 5 together with FIG. 2 and FIG. 4, the method for analyzing thermal image data can be performed by the data analysis unit (310).

[0113] In step S510, the data analysis unit (310) can estimate the skeleton of the human body from thermal image data. The data analysis unit (310) can identify the temperature distribution of the human body from the thermal image data acquired from the thermal image camera (110). The data analysis unit (310) can apply a deep learning-based pose estimation algorithm to the identified human body region. The data analysis unit (310) can specify the positions of the major joints of the human body as coordinates through the pose estimation algorithm. The data analysis unit (310) can generate a skeleton using the specified coordinates. For example, the major joints of the human body may be the head, shoulders, elbows, wrists, pelvis, knees, ankles, neck, spine, and toes. However, the present disclosure is not limited thereto. The method of using thermal image data may not directly collect personal information such as a person's face or personal details. Therefore, the method of using thermal image data can protect personal information.

[0114] In one embodiment of the present disclosure, step S510 may include the step of first identifying a temperature region corresponding to the human body in thermal image data received from a thermal image camera (110). Step S510 may include the step of applying a deep learning-based pose estimation algorithm within the identified region to determine the two-dimensional or three-dimensional coordinates of major joints.

[0115] In step S520, the data analysis unit (310) can calculate motion vectors based on time-series data for multiple joints constituting the skeleton. The data analysis unit (310) can generate time-series data by tracking changes in the position of each joint over consecutive time frames. The data analysis unit (310) can calculate motion vectors representing the direction and speed of movement of each joint between the previous frame and the current frame using the generated time-series data. For example, the motion vector of the wrist joint may be information indicating how fast and in which direction the wrist moved during a specific period of time.

[0116] In step S530, the data analysis unit (310) can determine whether the calculated motion vector matches a pattern of a predefined dangerous behavior. The data analysis unit (310) can reflect the result of the determination in the calculation of the risk level. The data analysis unit (310) can compare and analyze the currently calculated motion vector with motion vector patterns for each dangerous behavior that are stored in advance in a database. For example, dangerous behaviors may include assault, falling, vandalism, loitering, throwing, smoking, and fighting. However, the present disclosure is not limited thereto. For example, the data analysis unit (310) can detect a motion vector pattern in which the arm joint of one skeleton moves rapidly and repeatedly toward the head or torso of another skeleton while two skeletons are in close proximity. The data analysis unit (310) can determine the detected motion vector pattern as an act of 'assault'. Additionally, the data analysis unit (310) can detect a motion vector in which the head or pelvic joint of a skeleton rapidly descends at a speed exceeding a set speed. Afterward, the lowered skeleton may remain in a horizontal state for a certain period of time. In this case, the data analysis unit (310) may determine the action as a 'falling' action.

[0117] In one embodiment of the present disclosure, the data analysis unit (310) may be a part of the control unit (300).

[0118] In one embodiment of the present disclosure, the control unit (300) can estimate the skeleton of a human body from thermal image data. The control unit (300) can calculate motion vectors based on time-series data for a plurality of joints constituting the skeleton. The control unit (300) can determine whether the calculated motion vectors match a pattern of predefined dangerous behavior. The control unit (300) can reflect the result of the determination in the calculation of the risk level.

[0119] FIG. 6 is a flowchart illustrating a method for calculating risk according to various embodiments.

[0120] Referring to FIGS. 1 to 6, the steps for calculating the risk level can be performed by a control unit (300).

[0121] In step S610, the control unit (300) can calculate a first reliability score from the results of the thermal image data analysis. The control unit (300) can acquire thermal image data from the thermal image camera (110). The control unit (300) can analyze the acquired thermal image data to estimate the skeleton of a person within the bus stop area (R). The control unit (300) can calculate a motion vector based on the time series data of the estimated skeleton. The control unit (300) can quantitatively evaluate the degree to which the calculated motion vector matches a dangerous behavior pattern. For example, the dangerous behavior pattern may be assault, falling, loitering for a long time, vandalism, fighting, sudden running, or an abnormal stopping state. However, the present disclosure is not limited thereto. The control unit (300) can calculate a first reliability score based on the evaluated degree of match.

[0122] In step S620, the control unit (300) can calculate a second reliability score from the results of the acoustic data analysis. The control unit (300) can extract feature points from the acoustic data collected through the multi-microphone array (120). The control unit (300) can determine whether the extracted feature points match the pattern of a specific dangerous sound. For example, the specific dangerous sound may be a scream, a glass breaking sound, a distress call, an explosion sound, a gunshot, a vehicle crash sound, or a sharp metallic sound. However, the present disclosure is not limited thereto. Based on the determination result, the control unit (300) can calculate a second reliability score indicating the probability that the currently detected sound is associated with a dangerous situation.

[0123] In step S630, the control unit (300) can calculate the risk level through a weight-based fusion model that applies pre-set weights to the first reliability score and the second reliability score, respectively. The control unit (300) can evaluate the results of thermal image data analysis and acoustic data analysis by fusing them. The control unit (300) can improve the accuracy of risk determination by fusing two types of heterogeneous data. For example, the control unit (300) can dynamically adjust the weights for each reliability score according to situational information such as time of day, day of the week, or the congestion level of the bus stop area (R).

[0124] In one embodiment of the present disclosure, the control unit (300) can calculate a first reliability score from the results of thermal image data analysis. The control unit (300) can calculate a second reliability score from the results of acoustic data analysis. The control unit (300) can calculate the risk level through a weight-based fusion model (320) that applies a pre-set weight to the first reliability score and the second reliability score, respectively. Fusion analysis can reduce the false detection rate that may occur when relying on information from only one sensor. Fusion analysis can improve the accuracy of judgment regarding complex situations.

[0125] In one embodiment of the present disclosure, the step of calculating the risk level (S630) may include additional steps. The control unit (300) may additionally collect environmental data of the current bus stop area (R). The control unit (300) may dynamically reset the weights for the first reliability score and the second reliability score based on the collected environmental data. For example, the control unit (300) may recognize that it is currently nighttime through an illuminance sensor. The control unit (300) may determine that the reliability of the thermal image information is high during nighttime. The control unit (300) may adjust the weight of the first reliability score upward according to the determination. The control unit (300) may recognize that the average noise level measured by the multi-microphone array (120) exceeds a preset threshold. The control unit (300) may determine that the reliability of the acoustic data is low. The control unit (300) may adjust the weight of the second reliability score downward according to the determination.

[0126] FIG. 7 is a flowchart illustrating a safety management method through the detection of hazardous objects according to various embodiments.

[0127] Referring to FIGS. 3 to 7, a safety management method through the detection of hazardous objects can be performed by a control unit (300).

[0128] In step S310, the control unit (300) can analyze the thermal silhouette around the hand joints of the human skeleton from the thermal image data. The control unit (300) can extract the outline of an object from the thermal image data obtained from the thermal image camera (110) of the non-identification sensor unit (100). The control unit (300) can particularly focus on analyzing the temperature distribution and shape around the area presumed to be the human hand. The control unit (300) can obtain silhouette information of the object from the analyzed temperature distribution and shape.

[0129] In step S720, the control unit (300) can determine whether it matches a preset silhouette pattern of a dangerous object. The control unit (300) can compare the currently analyzed silhouette with silhouette pattern information of dangerous objects stored in a database. For example, the dangerous object may be a knife, a blunt instrument, a firearm, a hammer, a brick, a glass bottle, or a syringe. However, the present disclosure is not limited thereto. The control unit (300) can quantitatively calculate the similarity of the silhouette using a pattern matching algorithm or a deep learning-based object recognition model.

[0130] In step S730, the control unit (300) can immediately raise the risk level above a threshold when a dangerous object is detected. The control unit (300) can determine an immediate threat situation if the similarity exceeds a specific threshold based on the silhouette comparison results. The control unit (300) can adjust the risk level to the highest level. The control unit (300) can quickly initiate follow-up measures, such as activating the identification sensor unit (200), according to the adjusted risk level. Such immediate raising measures can contribute to preventing the occurrence of serious crimes, such as assault or robbery, in advance.

[0131] In one embodiment of the present disclosure, the control unit (300) can analyze the thermal silhouette around the hand joints of a human skeleton from the thermal image data. The control unit (300) may be configured to determine whether it matches a preset silhouette pattern of a dangerous object. If a dangerous object is detected, the control unit (300) may immediately raise the risk level above a threshold value.

[0132] FIG. 8 is a flowchart illustrating a method for providing a safe return home service according to various embodiments.

[0133] Referring to FIG. 8 together with FIG. 3 to 7, a method for providing a safe return home service can be performed by a control unit (300).

[0134] In step S320, the control unit (300) may assign at least one safety grade in real time according to the calculated risk level. The control unit (300) may compare the risk score calculated through fusion analysis with a preset threshold value for each interval. Based on the comparison result, the control unit (300) may determine a grade indicating the safety status of the current bus stop area. For example, the safety grade may be Safe, Caution, Alert, Danger, Severe, Emergency, or Evacuation grade. However, the present disclosure is not limited thereto. The real-time safety grade assignment can quantitatively evaluate the severity of the situation. Additionally, the real-time safety grade assignment can be used as a criterion for determining the level of follow-up action.

[0135] In step S810, when a service is requested via the NFC module (520), the control unit (300) can transmit a safety status message including a real-time safety rating to a user terminal or a pre-registered guardian terminal. The user can tag a user terminal, such as their smartphone, to the NFC module (520) installed at the bus stop. The control unit (300) can recognize the user's tagging action as a service request. The control unit (300) can generate a safety status message based on the recognized request. The safety status message may include the current safety rating of the bus stop, location information, and the current time. The control unit (300) can transmit the generated safety status message. The safety status message may be sent to a guardian terminal of a family member or acquaintance that the user has previously registered.

[0136] In one embodiment of the present disclosure, the intelligent bus stop safety management system (1) may further include a Near Field Communication (NFC) module (520) that communicates with a user terminal and a speaker (410) that outputs information as voice. The control unit (300) may assign at least one safety grade in real time according to the calculated risk level. When a service is requested through the NFC module (520), the control unit (300) may transmit a safety status message including the real-time safety grade to the user terminal or a pre-registered guardian terminal. Through this, users using the bus stop during late-night hours can inform the outside of their safety status and use public transportation with peace of mind.

[0137] In one embodiment of the present disclosure, step S810 may include the step of including a security link in the transmitted safety status message that allows checking the real-time status of the current bus stop. The guardian can view the real-time video with privacy masking through the security link in the received message. Additionally, the safety status message may include an emergency contact function that can immediately connect to the police or a control center in the event of an emergency. The user can request help by touching a button in the message.

[0138] Although embodiments of the present disclosure have been described above with reference to the attached drawings, those skilled in the art will understand that the present invention may be practiced in other specific forms without altering the technical concept or essential features thereof. Therefore, the embodiments described above should be understood as illustrative in all respects and not restrictive. The scope of protection of the present invention shall be interpreted by the claims below, and all technical concepts within the equivalent scope shall be interpreted as being included within the scope of rights of the technical concepts defined by the present disclosure. Explanation of the symbols

[0139] 1: Intelligent Bus Stop Safety Management System 100: Non-identifiable sensor unit 200: Identification sensor unit 300: Control Unit 400: Output unit 500: User Interface Unit

Claims

Claim 1 In an intelligent bus stop safety management system, a non-identification sensor unit that monitors a bus stop area; and an identification sensor unit that captures the bus stop area to acquire identifiable evidence images; and a control unit linked with the non-identification sensor unit and the identification sensor unit; wherein the control unit acquires primary sensor data from the non-identification sensor unit, and the primary sensor data includes thermal image data and acoustic data; and calculates the risk level of the bus stop area by fusion analysis of the acquired thermal image data and acoustic data, wherein the fusion analysis includes calculating a first reliability score from the thermal image data analysis result, calculating a second reliability score from the acoustic data analysis result, and calculating the risk level through a weight-based fusion model that applies a pre-set weight to the first reliability score and the second reliability score, respectively; and if the calculated risk level exceeds a pre-set threshold, the identification sensor unit is conditionally activated in the continuous monitoring mode through the non-identification sensor unit to switch to an evidence securing mode; and when the identification sensor unit is activated, the control unit identifies a specific related person through the thermal image data analysis, and in real-time, the body or face area of ​​a third party excluding the related person in the evidence image acquired by the identification sensor unit An intelligent bus stop safety management system characterized by masking the original unmasked evidence video, storing it in a separate storage space separately from the masked video, and decrypting it only according to a pre-set security procedure. Claim 2 In claim 1, the control unit estimates a human skeleton from the thermal image data, calculates a motion vector based on time-series data for a plurality of joints constituting the skeleton, and the calculated motion vector is predefined dangerous behavior An intelligent bus stop safety management system characterized by determining whether it matches the pattern and reflecting this in the risk calculation. Claim 3 An intelligent bus stop safety management system according to claim 1, wherein the non-identification sensor unit comprises a multi-microphone array composed of a plurality of microphones, and the control unit, when the acoustic data is detected, calculates the difference in the time of arrival of the sound reaching each microphone of the multi-microphone array to identify the location of the sound source of the acoustic data. Claim 4 An intelligent bus stop safety management system according to paragraph 3, wherein the control unit considers the acoustic data as external noise when the identified sound source generation location is determined to be outside the bus stop area. Claim 5 An intelligent bus stop safety management system according to claim 4, wherein the skeleton of a human body is estimated from the thermal image data, and the control unit increases the weight of the risk level when the identified sound source location and the location of the skeleton match within a preset distance. Claim 6 delete Claim 7 delete Claim 8 delete Claim 9 An intelligent bus stop safety management system according to claim 1, wherein the control unit is configured to analyze the thermal silhouette of the hand joints of a human skeleton from the thermal image data and determine whether it matches a preset silhouette pattern of a dangerous object, and, when the dangerous object is detected, immediately raises the risk level above the threshold value. Claim 10 An intelligent bus stop safety management system according to claim 1, further comprising: an NFC module communicating with a user terminal; and a speaker outputting information as voice; wherein the control unit assigns at least one safety grade in real time according to the calculated risk level, and when a service is requested through the NFC module, transmits a safety status message including the at least one safety grade to the user terminal or a pre-registered guardian terminal. Claim 11 An intelligent bus stop safety management system according to claim 10, wherein the control unit transmits an automatic guidance broadcast regarding dangerous behavior through the speaker when the risk level corresponds to a preset grade and the identification sensor unit is deactivated.

Images