Image analysis method and system for image security service using edge terminal

Abstract

The present invention relates to an image analysis method and system for an image security service using an edge terminal. The edge terminal according to the present invention reads a user request analysis function stored in a function setting DB, sets an execution target analysis function on the basis of the user request analysis function, calculates the probability of occurrence of a target event of the execution target analysis function through analysis of obtained images, generates image analysis results that include the probability of occurrence, and transmits the obtained images and the image analysis results to an image analysis server.

Description

Video analysis method and system for video security services using edge terminals

[0001] The present invention relates to a method and system for performing image analysis through collaboration between an edge terminal and an image analysis server.

[0002] Video security services utilizing edge terminals are attracting attention because they enable the real-time detection and response to abnormal situations, as video analysis is performed at the edge terminals. Generally, in such services, video analysis is conducted at the edge terminal, and the analysis results and video footage are transmitted to the video security system (server). The video security system then performs monitoring based on the analysis results and footage transmitted from the edge terminals.

[0003] However, these conventional methods have the disadvantage that not only do false negatives and false positives frequently occur because precise video analysis is difficult at edge terminals, but it is also difficult to apply various video analysis functions to video security services because resource limitations make it difficult to install multiple video analysis functions on edge terminals.

[0004] Therefore, a new image analysis method and system are needed that considers the resource limitations of edge terminals while reducing missed and false positives in the image analysis results of edge terminals.

[0005] The present invention aims to provide a video analysis method and system for a video security service using an edge terminal, which improves the video analysis miss detections and false positives that existed in conventional edge terminal-based video security services and applies various video analysis functions to the edge terminal-based video security service.

[0006] The objectives of the present invention are not limited to those mentioned above, and other unmentioned objectives will be clearly understood by those skilled in the art from the description below.

[0007] An edge terminal according to one embodiment of the present invention performs analysis of an acquired image. The edge terminal includes a first memory for storing computer-readable commands; and a first processor implemented to execute the commands.

[0008] The first processor is configured to read a first user request analysis function stored in a function setting DB in real time by executing the above commands, set an execution target analysis function according to a predetermined standard based on the user request analysis function, calculate the probability of occurrence of a target event of the execution target analysis function that is a detection target of the execution target analysis function through the analysis of the acquired image, generate an image analysis result including the probability of occurrence, and transmit the acquired image and the image analysis result to an image analysis server.

[0009] In one embodiment of the present invention, the first processor may be configured to set the common function of the operation capability analysis function of the edge terminal and the user request analysis function as the execution target analysis function during the process of setting the execution target analysis function.

[0010] In one embodiment of the present invention, the first processor transmits the image analysis result to an image analysis server, reads the second user request analysis function in real time from the function setting DB, and determines whether to update the execution target analysis function based on the second user request analysis function.

[0011] In one embodiment of the present invention, the target event of the execution target analysis function may include any one or a combination of collision, intrusion, assault, possession of a weapon, crowd density, loitering, flooding, and fire.

[0012]

[0013] An image analysis server according to one embodiment of the present invention performs image analysis.

[0014] The above-mentioned video analysis server includes a second memory for storing computer-readable commands; and a second processor implemented to execute said commands.

[0015] The second processor is configured to receive a user request analysis function from a user by executing the above commands, store the user request analysis function in a function setting DB, receive from an edge terminal an image acquired by the edge terminal and an image analysis result of the acquired image to which an execution target analysis function derived from the user request analysis function by the edge terminal is applied, and determine whether precise analysis of the acquired image is necessary based on the image analysis result.

[0016] In one embodiment of the present invention, the second processor may be configured to determine whether precise analysis of the acquired image is necessary based on the probability of occurrence of the target event of the execution target analysis function included in the image analysis result during the process of determining whether precise analysis of the acquired image is necessary.

[0017] In one embodiment of the present invention, the second processor may be configured to determine whether additional analysis of the acquired image is necessary based on the comparison result between the user request analysis function and the execution target analysis function.

[0018] In one embodiment of the present invention, the target event of the user request analysis function may include any one or a combination of collision, intrusion, assault, possession of a weapon, crowd density, loitering, flooding, and fire.

[0019]

[0020] According to one embodiment of the present invention, there is an effect of improving the detection of false positives and false negatives in video analysis of an edge terminal-based video security service.

[0021] According to one embodiment of the present invention, by applying various video analysis functions to an edge terminal-based video security service, the efficiency of the video security service is improved.

[0022] The effects obtainable from the present invention are not limited to those mentioned above, and other unmentioned effects will be clearly understood by those skilled in the art from the description below.

[0023] FIG. 1 is a block diagram showing the configuration of an image analysis system according to one embodiment of the present invention.

[0024] FIG. 2 is a flowchart illustrating an image analysis method according to an embodiment of the present invention.

[0025] FIG. 3 is a block diagram showing a computer system for performing an image analysis method according to an embodiment of the present invention.

[0026]

[0027] In conventional video security services using edge terminals, a method of performing video control based on the video analysis results of the edge terminal was mainly used. However, the method and system proposed in the present invention 1) reduce false positives and false negatives in video analysis by performing precise analysis on a video analysis server when necessary based on the reliability of the video analysis results performed on the edge terminal, and 2) use a method of performing the video analysis function on a video analysis server when the video analysis function desired by the user cannot be installed on the edge terminal due to resource limitations of the edge terminal.

[0028] The advantages and features of the present invention and the methods for achieving them will become clear by referring to the embodiments described below in detail together with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below but may be implemented in various different forms. These embodiments are provided merely to ensure that the disclosure of the present invention is complete and to fully inform those skilled in the art of the scope of the invention, and the present invention is defined only by the scope of the claims. Meanwhile, the terms used in this specification are for describing the embodiments and are not intended to limit the present invention. In this specification, the singular form includes the plural form unless specifically stated otherwise in the text. The terms "comprises" and / or "comprising" as used in this specification do not exclude the presence or addition of one or more other components, steps, actions, and / or elements in addition to the mentioned components, steps, actions, and / or elements.

[0029] Terms such as "first," "second," etc., may be used to describe various components, but said components should not be limited by said terms. These terms may be used for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, the first component may be named the second component, and similarly, the second component may be named the first component.

[0030] 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.

[0031] In describing the present invention, detailed descriptions of related prior art are omitted if it is determined that such descriptions may unnecessarily obscure the essence of the invention.

[0032] Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. In order to facilitate an overall understanding in describing the present invention, the same reference numerals will be used for the same means regardless of the drawing number.

[0033]

[0034] FIG. 1 is a block diagram showing the configuration of an image analysis system according to one embodiment of the present invention.

[0035] Referring to FIG. 1, an image analysis system (100) according to one embodiment of the present invention includes an edge terminal (110), an image analysis server (120), and a function setting DB (130). The function setting DB (130) may exist as an independent device as shown in FIG. 1, but may also be included in the image analysis server (120).

[0036] The image analysis system (100) illustrated in FIG. 1 is according to one embodiment, and the components of the image analysis system (100) according to the present invention are not limited to the embodiment illustrated in FIG. 1 and may be added, changed, or deleted as needed.

[0037]

[0038] The edge terminal (110) acquires images from a built-in camera or an external device. The edge terminal (110) generates an image analysis result through the analysis of the acquired images and transmits the acquired images and the image analysis result to an image analysis server (120). The image analysis server (120) receives the acquired images from the edge terminal (110) and performs a detailed analysis and / or additional analysis of the acquired images based on the image analysis result.

[0039] A user (50) of the video analysis server (120) inputs a user request analysis function into the video analysis server (120). The user request analysis function refers to information regarding the analysis function desired by the user (50). For example, the user (50) may request the video analysis system (100) to perform an analysis function regarding a car crash. As another example, the user (50) may request the video analysis system (100) to perform an analysis function regarding an assault. The user request analysis function has a target event, and the target event may include any one or a combination of collision, intrusion, assault, possession of a weapon, crowd density, loitering, flooding, and fire.

[0040] The edge terminal (110) includes a first memory for storing computer-readable commands and a first processor implemented to execute said commands.

[0041] The first processor, by executing the above commands, reads the user request analysis function (hereinafter referred to as the 'first user request analysis function') stored in the function setting DB (130) in real time, and sets the execution target analysis function according to a predetermined standard based on the first user request analysis function. The execution target analysis function refers to the analysis function actually performed by the edge terminal (110). For example, the edge terminal (110) may set the common function of the 'operable analysis function' and the first user request analysis function, which is a function that the edge terminal (110) can perform, as the execution target analysis function. The target event of the execution target analysis function may include any one or a combination of collision, intrusion, assault, possession of a weapon, crowd density, loitering, flooding, and fire.

[0042]

[0043] The first processor is configured to calculate the probability of occurrence of a target event of an execution target analysis function, which is a detection target of an execution target analysis function, through the analysis of an acquired image, generate an image analysis result including said probability of occurrence, and transmit said acquired image and said image analysis result to an image analysis server.

[0044] Then, the first processor transmits the above video analysis results to the video analysis server and reads the user request analysis function (hereinafter referred to as the 'second user request analysis function') in real time from the function setting DB (130). The user request analysis function stored in the function setting DB (130) can be updated by the video analysis server (120) at any time based on the input of the user (50). Therefore, the first user request analysis function and the second user request analysis function read from the function setting DB (130) at different times may have different configurations. The edge terminal (110) compares the first user request analysis function and the second user request analysis function to determine whether to update the execution target analysis function. If the first user request analysis function and the second user request analysis function are identical, the edge terminal (110) does not update the execution target analysis function. If the first user request analysis function and the second user request analysis function are not identical, the edge terminal (110) updates the execution target analysis function based on the second user request analysis function. That is, the edge terminal (110) determines the common function of the second user request analysis function and the operation capability analysis function of the edge terminal (110), and sets the common function as a new execution target analysis function.

[0045]

[0046] The image analysis server (120) receives an image acquired by the edge terminal (110) and the image analysis results from the edge terminal (110), determines whether precise analysis of the image is necessary based on the image analysis results, and can perform precise analysis according to the determination result. In addition, the image analysis server (120) can perform additional analysis of the image if the execution target analysis function of the edge terminal (110) does not fully satisfy the configured user request analysis function.

[0047] The video analysis server (120) includes a second memory for storing computer-readable commands and a second processor for executing said commands.

[0048] The second processor receives a user request analysis function from the user (50) by executing the above commands, and stores the user request analysis function received from the user (50) in the function setting DB (130). The target event of the user request analysis function may include any one or a combination of collision, intrusion, assault, possession of a weapon, crowd density, loitering, flooding, and fire.

[0049] And the second processor receives from the edge terminal (110) the image acquired by the edge terminal (110) and the image analysis result of the acquired image, which is generated by applying the execution target analysis function derived from the user request analysis function of the edge terminal (110).

[0050] The second processor is configured to determine whether precise analysis of the acquired image is necessary based on the image analysis result. In the process of determining whether precise analysis of the acquired image is necessary, the second processor may be configured to determine whether precise analysis of the acquired image is necessary based on the probability of occurrence of the target event of the execution target analysis function included in the image analysis result.

[0051] And, the second processor may be configured to determine whether additional analysis of the acquired image is necessary based on the comparison result between the user request analysis function and the execution target analysis function of the edge terminal (110).

[0052]

[0053] Hereinafter, with reference to FIG. 2, the operation of the image analysis system (100) will be described in detail.

[0054] FIG. 2 is a flowchart illustrating an image analysis method according to an embodiment of the present invention.

[0055] Referring to FIG. 2, an image analysis method according to one embodiment of the present invention is performed by an image analysis system (100) and includes steps S111 to S127. Steps S111 to S116 are performed by an edge terminal (110), and steps S121 to S127 are performed by an image analysis server (120). In this specification, steps S111 to S116 may be referred to as the 'method of operation of an edge terminal', and steps S121 to S127 may be referred to as the 'method of operation of an image analysis server'.

[0056] The image analysis method illustrated in FIG. 2 is according to one embodiment, and the steps of the image analysis method according to the present invention are not limited to the embodiment illustrated in FIG. 2 and may be added, changed, or deleted as needed.

[0057]

[0058] First, steps S111 to S116, which are the operation methods of the edge terminal (110), will be explained.

[0059] Step S111 is the analysis function setting step. This step is the step of setting the analysis function of the edge terminal (110) after starting the operation of the edge terminal (110).

[0060] The edge terminal (110) receives (or reads) information regarding an analysis function desired by a user (hereinafter referred to as "user request analysis function") from a function setting DB (130) that stores function setting data. The edge terminal (110) sets information regarding an analysis function that is installed on and operable on the edge terminal (110) among the above user request analysis functions (hereinafter referred to as "operable analysis function"), that is, an analysis function to be executed that the edge terminal (110) can perform. In this specification, the intersection of the user request analysis function and the operable analysis function is named the "analysis function to be executed". The edge terminal (110) stores the set analysis function to be executed in the function setting DB (130).

[0061] For example, if the edge terminal (110) has user request analysis functions obtained from the function setting DB (130) such as 'flooding', 'intrusion', 'assault', 'possession of a weapon', 'crowding', and 'loitering', and the operation possible analysis functions of the edge terminal (110) such as 'flooding', 'fire', 'collision', and 'assault', the analysis functions for 'flooding' and 'assault' are set as 'execution target analysis functions'.

[0062] The execution target analysis function of the edge terminal (110) can be described as a function that detects target events through the analysis of acquired images and calculates the probability of their occurrence. In the example above, the target events of the 'execution target analysis function' are 'flooding' and 'assault'. Of course, there may be one target event or three or more.

[0063]

[0064] Step S112 is the image acquisition step.

[0065] The edge terminal (110) can acquire video in real time from a camera built into the edge terminal (110) or from an external CCTV. The edge terminal (110) may be an edge-camera integrated terminal, such as an edge CCTV. In this case, the edge terminal (110) acquires video through the built-in camera. As another example, the edge terminal (110) may be a standalone edge terminal, such as an edge AI box, in which case it acquires video from an external CCTV.

[0066]

[0067] Step S113 is the edge-based image analysis step.

[0068] The edge terminal (110) performs image analysis based on the image obtained in step S112, which supports the execution target analysis function set in step S111, to determine whether a target event has occurred (type of event that has occurred) or to calculate the probability of the target event occurring.

[0069] That is, the edge terminal (110) determines whether a target event, which is the target of detection for the execution target analysis function, has occurred or calculates the probability of occurrence through the analysis process of the acquired image. For example, if the edge terminal (110) determines that a target event has occurred, it calculates the probability of occurrence for the event that has occurred.

[0070] Ultimately, the video analysis results generated in this stage are composed of the type of event that occurred and information on the probability of event occurrence (occurrence probability by event type).

[0071] In step S113, the artificial intelligence-based image analysis technique that the edge terminal (110) can take according to the target event of the execution target analysis function is as shown in the example in Table 1. Examples of artificial intelligence models that the edge terminal (110) can use for image analysis include a Convolutional Neural Network (CNN), a Recurrent Neural Network (RNN), and a Transformer.

[0072] Supplementary Explanation of Event AI-based Video Analysis Techniques Collision Object Detection / Recognition, Object Tracking, Motion Recognition Detects and tracks objects such as vehicles and people to determine their position, speed, and direction at the moment of collision. Determines whether a collision has occurred or the probability of a collision by assessing movement and deformation of the object's shape during the collision. Intrusion Object Detection / Recognition, Facial Recognition Detects objects moving within a defined surveillance area or objects crossing the boundary between a surveillance area and a non-surveillance area. Determines the probability of intrusion based on the detection results. Assault Object Detection / Recognition, Posture Estimation, Motion Recognition, Facial Expression Analysis Determines whether the detected object is a person. If it is a person, analyzes posture or movements of arms or legs to determine whether a violent act has occurred, whether the individual is a perpetrator or victim, or the probability. Adjusts the probability of occurrence by analyzing the facial expressions of the perpetrator or victim. Weapon-carrying Object Detection / Recognition, Image Classification, Posture Estimation, Motion Recognition Detects and classifies weapons such as knives, guns, steel pipes, and baseball bats. Specifically, the type of weapon is identified by analyzing its shape, size, color, etc. It is determined whether the object connected to the weapon is a person. The probability of weapon possession is determined by multiplying the probability that the detected object is a weapon by the probability that the connected object is a person. Risk is assessed by analyzing the posture and behavior of the person possessing the weapon. Crowd density object detection / recognition, object tracking, behavioral pattern analysis, motion recognition, area estimation, and density analysis determine whether an object within the video is a person. The area of ​​the captured region is estimated based on the object within the video. The number of people within the video is calculated, and density is calculated by dividing the calculated number of people by the estimated area. Congestion is estimated by analyzing the movement of people included in the video based on behavioral pattern analysis or motion recognition, and risk is predicted based on density, behavioral patterns, and motion recognition results. The probability of crowd density is determined based on the probability that the object within the video is a person and the probability distribution of the captured region area.Loitering Object Detection / Recognition, Object Tracking, Behavioral Pattern Analysis, Motion Recognition: Detects individuals who stay in a specific area for longer than a threshold time or engage in repetitive behaviors. Calculates the probability that the detected person is a suspicious individual based on the detection results and the frequency of interactions with other people over a predetermined period. Flooded Object Detection / Recognition, Object Tracking, Image Classification, Area Analysis, Area Estimation: Determines whether the detected object is water. If the detected object is water, analyzes the flooded area through area analysis. Distinguishes water from other objects based on color and texture. Estimates the area of ​​the region captured in the image based on objects within the image. Estimates the flooded area by comparing the estimated image area with the flooded area. Generates and analyzes time-series data representing changes in the flooded area over time. Calculates the probability of flooding occurrence based on the analysis of the flooded area or time-series data. Fire Thermal Imaging Camera Image Analysis, Object Detection / Recognition, Image Classification: Analyzes temperature changes based on thermal imaging camera images. Determines whether the detected object is smoke or flames. Determining whether a fire has occurred or the probability of its occurrence by analyzing the color, shape, and movement of the flames.

[0073]

[0074] Step S114 is the step of transmitting the video and video analysis results.

[0075] The edge terminal (110) transmits the image obtained through step S112 and the image analysis result generated through the execution of step S113 to the image analysis server (120).

[0076] Step S115 is a step for determining whether to change the execution target analysis function.

[0077] The edge terminal (110) receives a user request analysis function from the function setting DB (130), derives an execution target analysis function that is common to the user request analysis function and the operation capability analysis function of the edge terminal (110), and if the derived execution target analysis function is different from the currently set execution target analysis function, returns to step S111 to update the execution target analysis function, otherwise proceeds to step S116.

[0078]

[0079] Step S116 is the step for determining whether to terminate.

[0080] The edge terminal (110) determines whether there is a 'termination' user input, and if the user instructs to terminate, the operation of the edge terminal (110) is terminated, otherwise, it returns to step S112 and continues to perform image acquisition and image analysis.

[0081]

[0082] Next, steps S121 to S127, which are video analysis methods of the video analysis server (120), will be described.

[0083] Step S121 is the user function setting step.

[0084] This step is a step of performing user function settings after the image analysis server (120) according to one embodiment of the present invention starts operation.

[0085] The video analysis server (120) sets up video analysis functions for video security services based on input from a user (controller). Specifically, the video analysis server (120) receives user request analysis functions (information on analysis functions desired by the user) from the user (controller) and stores them in the function setting DB (130). The user request analysis functions stored in the function setting DB (130) can be read from the edge terminal (110).

[0086] For example, the user request video analysis function can be 'collision', 'intrusion', 'assault', 'possession of a weapon', 'crowd density', 'loitering', 'flooding', 'fire', etc. The corresponding information is stored in the function setting DB (130).

[0087]

[0088] Step S122 is the step of receiving the video and video analysis results.

[0089] The video analysis server (120) receives video and video analysis results from the edge terminal (110). The video analysis results include the type of event that occurred and the probability of occurrence for each type of event.

[0090]

[0091] Step S123 is the step for determining whether detailed analysis is necessary.

[0092] The video analysis server (120) determines whether precise analysis of the video analysis server (120) is required based on the video analysis results transmitted by the edge terminal (110).

[0093] Specifically, the video analysis server (120) determines whether to perform a detailed analysis of the video acquired by the edge terminal (110) based on the types of events included in the video analysis results and the probability of occurrence for each type of event.

[0094] For example, if the event included in the video analysis result is ‘assault’ and the probability of the event occurring is 70% or higher, the video analysis server (120) determines that ‘assault’ has occurred and determines that no detailed analysis is needed for the ‘assault’ event.

[0095] Additionally, if the event included in the video analysis result is ‘assault’ and the probability of the event occurring is in the range of 30% or more and less than 70%, the video analysis server (120) determines that a detailed analysis of the video acquired by the edge terminal (110) is required for the ‘assault’ event.

[0096] In addition, if the event included in the video analysis result is ‘assault’ and the probability of the event occurring is 30% or less, the video analysis server (120) determines that ‘assault’ did not occur and determines that detailed analysis is not necessary.

[0097] If it is determined in step S123 that a detailed analysis is required, the video analysis server (120) performs a detailed analysis on the event that is determined to require a detailed analysis (S124). After performing step S124, the video analysis server (120) proceeds to step S125.

[0098] In step S123, if it is determined that detailed analysis of the event that occurred is not necessary, the video analysis server (120) proceeds to step S125.

[0099]

[0100] Step S124 is the precision analysis step.

[0101] The video analysis server (120) performs a detailed analysis on a target event that is determined to require detailed analysis, and determines whether the target event has occurred.

[0102] For example, if the video analysis server (120) determines that a detailed analysis of an 'assault' event is necessary, the video analysis server (120) performs a detailed analysis of the video transmitted by the edge terminal (110) to determine whether an 'assault' event has occurred.

[0103] The video analysis server (120) can perform the artificial intelligence-based video analysis technique described in Table 1 in step S124. If there is a method among the video analysis techniques described in Table 1 that the edge terminal (110) has not performed, the video analysis server (120) can perform it in the precise analysis step to more clearly determine whether the target event has occurred.

[0104] When step S124 is completed, step S125 is performed.

[0105]

[0106] Step S125 is the step for determining whether additional analysis is necessary.

[0107] This step is a step for determining whether additional analysis is required for the image acquired by the edge terminal (110).

[0108] The video analysis server (120) reads the execution target analysis function (meaning information about the execution target analysis function) that the edge terminal (110) is performing from the function setting DB (130), and determines whether additional analysis is necessary by determining whether the execution target analysis function matches the user request analysis function. That is, the process of determining whether additional analysis is necessary is a process of determining whether the execution target analysis function performed by the edge terminal (110) satisfies all user request analysis functions stored in the function setting DB (130), and the video analysis server (120) can be said to determine whether additional analysis of the video acquired by the edge terminal (110) is necessary by comparing the configured user request analysis function and the execution target analysis function of the edge terminal (110).

[0109] For example, if the user request analysis function stored in the function setting DB (130) is 'collision', 'intrusion', 'assault', and 'possession of a weapon', and the execution target analysis function performed at the edge terminal (110) is 'collision' and 'intrusion', the video analysis server (120) may determine that additional analysis is needed for 'assault' and 'possession of a weapon'.

[0110] In addition, as another example, if the user request analysis function stored in the function setting DB (130) and the execution target analysis function performed by the edge terminal (110) are the same, the video analysis server (120) determines that no additional analysis is required.

[0111] If it is determined that additional analysis is necessary in step S125, step S126 is performed; otherwise, step S127 is performed.

[0112]

[0113] Step S126 is the step for performing additional video analysis.

[0114] If the video analysis server (120) determines that additional analysis is needed in step S125, it performs additional analysis on the video transmitted by the edge terminal (110).

[0115] Specifically, the video analysis server (120) performs additional analysis on the video with respect to the functions that are not the target events of the edge terminal (110) among the user request analysis functions stored in the function setting DB (130), namely the execution target analysis functions. That is, the video analysis server (120) calculates whether an event occurs and / or the probability of occurrence of an event that is not a target event of the execution target analysis function but is designated as a target event of the user request analysis function, based on the video transmitted by the edge terminal (110).

[0116] The video analysis server (120) can perform the artificial intelligence-based video analysis technique described in Table 1 at step S126. If there is a method among the video analysis techniques described in Table 1 that the edge terminal (110) has not performed, the video analysis server (120) can perform it at the additional analysis step.

[0117] When step S126 is completed, step S127 is performed.

[0118]

[0119] Step S127 is the step for determining whether to terminate.

[0120] The video analysis server (120) determines whether there is a 'termination' user input, and if the user instructs termination, the operation of the video analysis server (120) is terminated, and if not, it returns to step S122.

[0121]

[0122] The aforementioned image analysis method has been described with reference to the flowchart presented in the drawings. For simplicity of explanation, the method has been illustrated and described in a series of blocks; however, the present invention is not limited to the order of said blocks, and some blocks may occur in a different order or simultaneously with other blocks as illustrated and described herein, and various other branches, flow paths, and sequences of blocks may be implemented to achieve the same or similar results. Furthermore, not all illustrated blocks may be required for the implementation of the method described herein.

[0123]

[0124] Meanwhile, in the description with reference to FIG. 2, each step may be further divided into additional steps or combined into fewer steps according to an embodiment of the present invention. Also, some steps may be omitted as necessary, and the order between steps may be changed. Furthermore, even if other omitted details are included, the contents of FIG. 1 or FIG. 3 may be applied to the contents of FIG. 2. Also, the contents of FIG. 2 may be applied to the contents of FIG. 1 or FIG. 3.

[0125]

[0126] FIG. 3 is a block diagram showing a computer system for implementing an image analysis method according to an embodiment of the present invention. An edge terminal (110), an image analysis server (120), or a function setting DB (130) may be implemented in the form of the computer system shown in FIG. 3.

[0127] Referring to FIG. 3, a computer system (1000) may include at least one of a processor (1010), memory (1030), an input interface device (1050), an output interface device (1060), and a storage device (1040) that communicate via a bus (1070). The computer system (1000) may also further include a communication device (1020) coupled to a network. The processor (1010) may be a central processing unit (CPU) or a semiconductor device that executes computer-readable instructions stored in memory (1030) or storage device (1040). Memory (1030) and storage device (1040) may include various forms of volatile or non-volatile storage media. For example, memory (1030) may include read-only memory (ROM) and random access memory (RAM). In the embodiment of the present description, the memory (1030) may be located inside or outside the processor (1010), and the memory (1030) may be connected to the processor (1010) through various known means. The memory (1030) is a volatile or non-volatile storage medium of various forms, and for example, the memory (1030) may include read-only memory (ROM) or random access memory (RAM).

[0128] Accordingly, embodiments of the present invention may be implemented as a method implemented on a computer or as a non-transient computer-readable medium in which computer-executable instructions are stored. In one embodiment, when executed by a processor (1010), the computer-readable instructions may perform a method according to at least one aspect of the present description.

[0129] The communication device (1020) can transmit or receive wired or wireless signals.

[0130] In addition, the image analysis method according to an embodiment of the present invention may be implemented in the form of program instructions that can be executed through various computer means and may be recorded on a computer-readable medium.

[0131] The above computer-readable medium may include program instructions, data files, data structures, etc., either individually or in combination. The program instructions recorded on the computer-readable medium may be specially designed and configured for embodiments of the present invention, or they may be known and available to a person skilled in the art of computer software. The computer-readable recording medium may include a hardware device configured to store and execute program instructions. For example, the computer-readable recording medium may be 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; ROM; RAM; flash memory, etc. The program instructions may include not only machine code, such as that generated by a compiler, but also high-level language code that can be executed by a computer through an interpreter, etc.

[0132] The processor (1010) can perform all or part of the steps of the image analysis method by executing computer-readable commands stored in memory (1030) or storage device (1040).

[0133]

[0134] Although the present invention has been described above with reference to preferred embodiments, those skilled in the art will understand that various modifications and changes can be made to the invention without departing from the spirit and scope of the invention as described in the following claims.

[0135]

[0136] [Explanation of the symbol]

[0137] 50: User

[0138] 100: Video Analysis System

[0139] 110: Edge terminal

[0140] 120: Video Analysis Server

[0141] 130: Function Configuration DB

[0142] 1000: Computer System

[0143] 1010: Processor

[0144] 1020: Communication device

[0145] 1030: Memory

[0146] 1040: Storage device

[0147] 1050: Input interface device

[0148] 1060: Output interface device

[0149] 1070: Bus

Claims

1. An edge terminal that performs analysis of an acquired image, A first memory for storing computer-readable commands; and It includes a first processor implemented to execute the above commands, and The first processor above, by executing the above instructions, Reads the first user request analysis function stored in the function settings DB in real time, and Based on the above user request analysis function, the execution target analysis function is set according to predetermined criteria, and Calculate the probability of occurrence of a target event of the execution target analysis function, which is a detection target of the execution target analysis function, through the analysis of the acquired image, and generate an image analysis result including the probability of occurrence. Configured to transmit the above-mentioned acquired image and the above-mentioned image analysis results to an image analysis server. In-edge terminal.

2. In paragraph 1, the first processor is, In the process of setting the above execution target analysis function, The common function of the operation capability analysis function of the edge terminal and the user request analysis function is configured to be set as the execution target analysis function. In-edge terminal.

3. In paragraph 1, the first processor is, After transmitting the above video analysis results to the video analysis server, reading the second user request analysis function in real time from the above function setting DB, and determining whether to update the above execution target analysis function based on the above second user request analysis function. In-edge terminal.

4. In paragraph 1, the target event of the execution target analysis function is, Includes any one or a combination of collision, intrusion, assault, possession of a weapon, crowding, loitering, flooding, and fire. In-edge terminal.

5. In a video analysis server that performs video analysis, A second memory for storing computer-readable commands; and It includes a second processor implemented to execute the above commands, and The second processor above, by executing the above instructions, Receive user request analysis functions from the user, and store the said user request analysis functions in the function settings DB, Receiving from an edge terminal an image acquired by the edge terminal and an image analysis result of the acquired image to which an execution target analysis function derived from the user request analysis function of the edge terminal is applied, Based on the above image analysis results, configured to determine whether precise analysis of the acquired image is necessary. Video analysis server.

6. In paragraph 5, the second processor is, In the process of determining whether precise analysis of the above-mentioned acquired image is necessary, It is configured to determine whether precise analysis of the acquired image is necessary based on the probability of occurrence of the target event of the execution target analysis function included in the image analysis result. Video analysis server.

7. In paragraph 5, the second processor is, It is configured to determine whether additional analysis of the acquired image is necessary based on the comparison result between the user request analysis function and the execution target analysis function. Video analysis server.

8. In paragraph 5, the target event of the above user request analysis function is, Includes any one or a combination of collision, intrusion, assault, possession of a weapon, crowding, loitering, flooding, and fire. Video analysis server.

Images