KR20130068334A - Automatic recognition and response system of the armed robbers and the methods of the same - Google Patents

Abstract

PURPOSE: An armed robber auto-recognition reaction system and a method thereof are provided to quickly stop a case progress by synthetically analyzing a surveillance camera and operating an emergency alarm in the spot and seek for a fast countermeasure by reporting a field situation in real time. CONSTITUTION: An armed robber auto-recognition reaction system includes a CCTV camera(100) which collects an image by photographing a field situation; an image transformer(200) which converts the image which is photographed through the CCTV camera into an analyzable image; an image analysis unit(300) which analyzes an image which is converted by the image transformer; a control unit(400) which stores a result which is analyzed by the image analysis unit in a circular queue and determines whether the analyzed result is transmitted; a transmission module(500) which transmits the analyzed result by the control unit; a central server(600) which receives a signal which is transmitted from the transmission module; and an alarm generation unit(700) which generates an alarm signal by the signal of the central server. [Reference numerals] (100) CCTV camera; (210) HSV transforming unit; (220) Noise removing unit; (230) Fourier transforming unit; (240) Edge extracting unit; (310) Recognizing unit; (320) Background recognizing unit; (330) User recognizing unit; (340) Pattern database; (350) Analysis result database; (410) Queue storing unit; (420) Measuring unit; (500) Transmitting module; (600) Central server; (610) Displaying unit; (700) Alarm releasing unit; (AA) Branch; (BB) Head center; (CC) Manager

Description

Automatic recognition and response system of the armed robbers and the methods of the same}

The present invention relates to a system for automatically recognizing and responding to armed strength, and more specifically, to a firearm invading a public institution such as a bank using a surveillance camera. And automatic detection and response system for armed strength to stop the progress of the incident by triggering the alarm on the scene and report the situation to the headquarters and related organizations in real time and to take quick countermeasures. It is about.

In the United States, about 7,200 bank robbery cases occur each year, and armed robbery incidents using firearms often occur as individual gun possessions increase through various routes in Korea.

In the event of such an armed robbery, the amount of damage is great, but in the event of personal injury, the customer's credibility to the institutions such as banks and post offices and the loss of the image of the institutions are not matters that can be converted into money.

Recently, security systems have also been continuously developed due to the development of technology, but the number of incidents of armed robbery incidents using firearms has been continuously increasing.

As countermeasures against armed robbery cases of public institutions, methods such as increased armed guards, installation of bulletproof glass and weapon detectors, CCTVs, and alarm systems have been proposed. In the case of weapons and weapons detectors, there is a problem in that the discomfort is impaired to the customers, and in the case of CCTV, the camouflage is indispensable for the resolution of the armed robbery, which is inefficient in solving the case. There is a problem that causes a life threat.

In addition, the prior art of the method for responding to the invasion of armed strength, Korean Patent Publication No. 10-0584481 discloses an armed strength intrusion monitoring system and method, the main technical configuration of the microphone 11, the camera An armed strength intrusion surveillance alarm system comprising a 12 and a system controller 15, comprising: a metal detector for detecting metal; A voice signal detection / voice recognition device embedded in the system controller capable of calculating a risk for an input voice signal; And a metal detection signal processing device embedded in the system controller capable of calculating a risk level for the input metal signal.

The configuration as described above is characterized in that the metal detector 13 detects the metal, and the video and audio through the video camera and microphone to comprehensively detect and monitor the intrusion situation of the armed strength, As such, metal detectors are not only disadvantageous to customers when they are used in institutions such as banks and post offices, but also to impair intimacy. There was a problem that could be a factor to inhibit.

Therefore, there is a need for a system and method with low cost to solve the problems described above and the ability to maintain intimacy with the customer, life safety and the cessation of incidents quickly.

The present invention has been made in order to solve the above problems, an object of the present invention is to synthesize a gun, armed robbery, the scene of the scene of the incident, employee gestures, etc. invading public institutions such as banks using a surveillance camera. Automatic recognition and response system of armed strength to analyze the situation and based on this, to stop the progress of the incident as soon as possible and to report the situation in real time to the headquarters and related organizations in order to take a quick response. In providing.

In addition, the present invention is able to quickly prevent the progress of armed robbery incidents while maintaining intimacy with the customer to provide a safer working environment and at the same time automatic recognition of armed strength to ensure a high price competitiveness with a simple system configuration And providing a corresponding system and method thereof.

According to an aspect of the present invention,

In the armed strength automatic recognition and response system, CCTV camera for collecting the image of the scene by collecting the scene, an image conversion unit for converting the image captured by the CCTV camera to the image that can be analyzed, and by the image conversion unit An image analyzer which analyzes the converted image, a controller which stores the result analyzed by the image analyzer in the form of a circular queue, and determines whether to transmit the analysis result, and analyzes the judgment by the controller It characterized in that it comprises a transmission module for transmitting a result, a central server for receiving a signal transmitted from the transmission module and an alarm generator for generating an alarm signal by the signal of the central server.

At this time, the image conversion unit HSV conversion unit for converting the image taken by the CCTV camera in the form of HSV (Hue Saturation Value), a noise removal unit for removing noise from the image converted into the HSV form, the image from which the noise is removed And an edge extractor for extracting an edge image from an image from which noise is removed.

In addition, the image analyzer analyzes the image converted by the image converter using an Ada-Boost algorithm to recognize and detect the shape of the firearm, and to recognize the characteristic background of the shooting site to detect It comprises a background recognition unit and an employee recognition unit for recognizing and detecting the gesture of the employee.

The image analyzer further includes a pattern database for storing pattern data to be used as basic data when analyzing an image using the Ada-Boost algorithm, and an analysis result database for storing the image analysis result using the Ada-Boost algorithm together with the image. It is characterized by.

In addition, the control unit normalizes the analysis results of the image analysis unit in the form of O / X and the cue storage unit for storing in the form of a circular queue, and confirms the frequency of O of the analysis results stored in the form of a circular queue to transmit the current situation Characterized in that it comprises a frequency measuring unit for determining whether or not.

On the other hand, armed strength automatic recognition and corresponding method according to the present invention, in the automatic armed strength automatic recognition and response method, a field shooting step of photographing the situation of the scene using a CCTV camera, and receiving the captured image image conversion unit An image conversion step of converting the image into an image that can be analyzed through the image conversion step, an image analysis step of analyzing the image converted in the image conversion step through an Ada-boost algorithm, and normalizing the analyzed image in an O / X form to form a circular queue. The storage judgment step of determining whether or not to transmit the current situation by checking the frequency of O from the stored analysis results, and if it is determined that transmission is necessary, the transmission module determines the current situation. A transmission step of transmitting the image and the judgment result to the central server and the person in charge and the image and the judgment result transmitted by the central server and the person in charge OK is characterized in that is configured to include an alert generation step of generating an alarm signal.

At this time, the image conversion step, HSV conversion step of converting the image taken by the CCTV camera in the form of HSV (Hue Saturation Value), noise removal step of removing noise from the image converted into HSV form, noise is removed And an edge extracting step of extracting an edge image from the image from which the noise is removed.

In addition, the image analysis step is characterized by a comprehensive analysis by recognizing the characteristic background of the armed robbery and the scene of the incident and the gesture of the staff through the Ada-Boost algorithm with reference to the previously stored pattern data of the accident situation. .

In the transmitting step, when an abnormality is detected in the transmission determining step, an image and a determination result of the current situation are transmitted to the central server, and the designated person is notified of the determination result by SMS.

According to the present invention, it is possible to quickly prevent the progress of armed robbery incidents while maintaining intimacy with the customer, thereby providing a safer working environment and having an excellent effect of securing a high price competitiveness with a simple system configuration. .

In addition, according to the present invention by using a surveillance camera to comprehensively analyze the gun, armed robbery of the incident site, incidents of the staff, and gestures of employees invading public institutions such as banks, and based on this, by operating an alarm on the incident site It has the effect of stopping the process as soon as possible and reporting the situation on the headquarters and related organizations in real time so that a quick response can be taken.

In addition, according to the present invention, the final alarm signal generation command is directly issued by the person in charge using the central server, thereby preventing the malfunction of the alarm generation unit at the source, and further has the effect of more flexible utilization.

1 is a view schematically showing a conventional armed strength alarm device.
2 is a conceptual diagram conceptually showing an armed strength automatic recognition and response system according to the present invention.
3 is a view schematically showing an armed strength automatic recognition and corresponding system according to the present invention.
Figure 4 is a schematic diagram illustrating the event processing of the automatic armed strength recognition and response system according to the present invention.
5 is a flowchart illustrating a method for automatically recognizing and responding to armed strength according to the present invention;

Hereinafter, with reference to the accompanying drawings will be described in detail preferred embodiments of the automatic armed strength recognition and response system and method according to the present invention.

2 is a conceptual diagram conceptually illustrating an automatic arming strength recognition and response system according to the present invention, and FIG. 3 is a view schematically showing an automatic arming strength recognition and response system according to the present invention. Figure 4 is a schematic diagram showing the event processing process of the automatic armed strength recognition and response system according to the present invention, Figure 5 is a flow chart showing a method of automatic recognition and armed strength according to the present invention.

The present invention comprehensively analyzes the weapons of armed robbery, the scene of the crime scene, the gesture of the staff, etc. using the surveillance camera, and triggers an alarm on the scene of the incident to stop the progress of the incident. The present invention relates to an automatic armed strength recognition and response system and a method thereof for enabling quick countermeasures by reporting the site situation to the head office and related organizations in real time. As shown in 3, the CCTV camera 100, the image conversion unit 200, the image analysis unit 300, the control unit 400, the transmission module 500, the central server 600 and the alarm generator 700 It is configured to include.

In more detail, the CCTV camera 100 is installed in a public place such as a bank or a post office and serves to collect images by photographing the situation of the site, and a plurality of blind spots are installed so as not to occur. It is configured to monitor all areas of the site in real time.

Next, the image converting unit 200 is electrically connected to the CCTV camera 100 and serves to convert the image captured by the CCTV camera 100 into an image that can be analyzed, and HSV (Hue Saturation Value). The conversion unit 210, the noise removing unit 220, Fourier (Fourier) conversion unit 230 and the edge extraction unit 240 is included.

That is, the HSV converting unit 210 converts the color image signal of the RGB type photographed from the CCTV camera 100 into an HSV type image signal expressing hue, saturation, and contrast as a light source such as illumination and sunlight. This prevents the captured image from being inaccurate and obtains a more accurate image. The noise removing unit 220 is connected to the HSV converter 210 to convert the image into an HSV form. It removes the noise to get a clearer and more accurate image.

Next, the Fourier transform unit 230 is connected to the noise removing unit 220 to convert an image from which the noise is removed into a frequency domain through Fourier transform, thereby removing unnecessary images and obtaining a clearer image. The edge extractor 240 is connected to the noise remover 220 to extract an edge image from the image from which the noise is removed.

That is, the edge extractor 240 extracts an outline (edge image) of an object based on a boundary between a background and an object from an HSV-type image from which noise is removed, so that the image analysis unit 300, which will be described later, provides a more accurate image. So that the analysis can be done.

Next, the image analysis unit 300 is connected to the image conversion unit 200 is installed to analyze the image converted by the image conversion unit 200 to grasp the situation of the scene taken by the CCTV camera 100. In this case, the image analyzer 300 analyzes the image converted by the image converter 200 through an Ada-Boost algorithm.

In this case, since a technique of recognizing a specific shape through the Ada-boost algorithm is conventionally used, a detailed description thereof will be omitted.

In more detail, the image analyzer 300 includes a gun recognizer 310, a background recognizer 320, and an employee recognizer 330, and the gun recognizer 310 includes an edge extractor. A role of detecting and recognizing the shape of the firearm by comparing the image extracted from the 240 or the image converted by the Fourier transform unit 230 and the pattern data of the firearm shape previously stored in the pattern database 340 to be described later. The background recognizer 320 compares the image extracted by the edge extractor 240 or the image converted by the Fourier transform unit 230 with the pattern data of the photographing site previously stored in the pattern database 340. It serves to detect and recognize the characteristic background of the shooting site, the employee recognition unit 330 is an image extracted from the edge extraction unit 240 or the image and pattern database (converted by the Fourier transform unit 230) ( 34 By comparing the movement pattern data of the employees already stored in 0), it plays a role to detect and recognize simple gestures (for example, promised hand gestures) of the employees.

That is, the image analyzer 300 may perform various patterns to be used as basic data when analyzing the image using the Ada-Boost algorithm, that is, the shape of the firearm, the background where people are lying down or the objects are disturbed, and the staff may perform the emergency. And a pattern database 340 for storing a database of pattern data of an image or an image such as a promised gesture, and converting the image from the Fourier transform unit 230 or the edge extractor 240 during image analysis. The comparison is made to analyze the site situation more accurately.

In this case, the image analyzer 300 further includes an analysis result database 350 for storing the image analysis result through the Ada-boost algorithm together with the corresponding image.

Next, the control unit 400 is connected to the image analysis unit 300 and is to determine whether or not to transmit the results analyzed by the image analysis unit 300 to the central server 600 installed in the head office. It includes a queue (Queue) storage unit 410 and the frequency measuring unit 420.

In more detail, the queue storage unit 410 serves to store a plurality of data items in a form of a queue by normalizing the analysis result of the image analyzer 300 in an O / X form. In the present invention, the analysis results normalized in O / X form are divided into predetermined unit times and stored in the form of a circular queue.

That is, the queue storage unit 410 is based on, for example, the results of the firearm recognition unit 310, the background recognition unit 320 and the employee recognition unit 330 of the image analysis unit 300, armed strength firearms Are categorized into various items and normalized so that they can be judged in O / X form, such as whether they are in possession of, a large number of customers are not moving, and whether employees are taking special gestures as promised. It is stored in the form of a circular queue.

Next, the frequency measuring unit 420 is to check the frequency of O of the analysis results stored in the form of a circular cue in order to determine whether an abnormal situation occurs at the shooting site, for example, If the frequency of O is three or more among the analysis results classified into ten items and stored in the form of a circular queue, it may be determined that an abnormal situation has occurred and transmitted to the central server 600 and the person in charge about the current situation.

At this time, among the analysis results normalized in O / X form as described above, it is possible to determine by weighting important items such as whether or not the gun is possessed, and the frequency measuring unit 420 divides the frequency of O by dividing by unit time. By confirming, the reliability of the judgment result can be improved.

Next, the transmission module 500 transmits the analysis result of the image analyzer 300 to the central server 600 and the person in charge at the head office based on the result determined by the frequency measuring unit 420 of the controller 400 described above. If it is determined that the abnormal situation has occurred, the analysis result stored in the analysis result database 350 and the corresponding image and the O / X-type analysis result stored in the queue storage unit 410 corresponding to the central server To 600.

In addition, the transmission module 500 is configured to automatically notify the fact that the occurrence of the abnormal situation to the mobile phone in charge of the headquarters or related organizations when it is determined that the abnormal situation occurred in the shooting site is configured to make a faster action have.

On the other hand, the central server 600 serves to receive a signal transmitted from the transmission module 500, it may be installed in the relevant authority, such as the control room or police station of the head office.

At this time, the central server 600 is connected to a plurality of points at the same time installed through the network, in order to facilitate the immediate situation determination, in addition to the report on the abnormal situation transmitted from the branch side directly from the head office or related organizations The display unit 610 for monitoring by connecting to the CCTV camera 100 of each point is provided.

In addition, even when monitoring the CCTV camera 100 at a specific point, the display unit 610 preferentially displays a photographed image of a corresponding area when a signal for an abnormal situation is received from the transmission module 500 at another point. It is configured to be.

Next, the alarm generating unit 700 serves to generate an alarm signal such as a warning light and an alarm sound by a command of the central server 600, and is installed at each branch and a related authority such as a police station.

That is, when an abnormal situation is received through the central server 600 and the display unit 610 provided in the control room of the head office or related organizations, the person in charge checks this and finally determines whether an alarm signal is generated. By lowering the signal through 600 is configured to generate an alarm signal through the alarm generator 700.

Therefore, the armed robbery automatic recognition and response system according to the present invention as described above is to be installed in branches and headquarters of financial institutions or public institutions where armed robbery events occur, such as banks, post offices, CCTV cameras 100, video The converting unit 200, the image analyzing unit 300, the control unit 400, the transmission module 500 and the alarm generating unit 700 is installed at a point, and finally receives the data of the field situation and finally generates an alarm signal. The central server 600 to determine the can be installed in the control room of the head office.

In addition, the final alarm signal generation command is directly issued by the person in charge using the central server 600, thereby preventing the malfunction of the alarm generation unit at the source, and has the advantage of more flexible utilization.

Hereinafter, with reference to the accompanying drawings will be described in detail the armed strength automatic recognition corresponding method according to the present invention.

In accordance with the present invention, the method for automatically recognizing armed strength according to the present invention has a large field shooting step (S10), an image conversion step (S20), an image analysis step (S30), a transmission determination step (S40), and a transmission step. It comprises a (S50) and the alarm generating step (S60).

In more detail, the field shooting step (S10) relates to the step of taking a picture of the situation in real time using the CCTV camera 100 installed in each point, the image taken by the CCTV camera 100 is a network It is configured to be displayed on the display provided in the central server 600 installed in the head office.

Next, the image conversion step (S20) relates to the step of receiving the image taken by the CCTV camera 100 and converting it into an image that can be analyzed by the image conversion unit 200, HSV conversion step (S22), Noise removal step S24, Fourier transform step S26, and edge extraction step S28.

In more detail, the HSV conversion step S22 converts the RGB image photographed by the CCTV camera 100 into an HSV type image through the HSV conversion unit 210, thereby excluding the influence of the light source, thereby more precisely and vividly. This is a step for obtaining an image, and the noise removing step (S24) is to remove the noise from the image converted into HSV form using the noise removing unit 220 in the image analysis step (S30) to be described later This step is to play a more accurate analysis.

Next, the Fourier transforming step (S26) is a step of eliminating unnecessary images and obtaining a clearer image by converting an HSV-shaped image from which noise is removed into a frequency domain through Fourier transform, and the edge In the extracting step (S28), the image analysis unit 300, which will be described later, may analyze the image more accurately by extracting an outline (edge image) of the object based on the boundary between the background and the object from the HSV-shaped image from which the noise is removed. That's the step.

Next, the image analysis step (S30) relates to the step of determining the situation of the shooting scene by analyzing the image converted in the image conversion step (S20), in the present invention analyzes the image through the Ada-Boost algorithm. .

In more detail, the image analysis step (S30) compares the image transformed in the Fourier transform step (S26) or the edge extraction step (S28) with the pattern data previously stored in the pattern database 340 to perform the ada-boost algorithm. In the pattern database 340, the shape of the gun that can possess armed strength, the situation of the site that may occur when armed strength is invaded, the gesture that an employee can perform when armed strength is invaded, etc. The patterns of various situations are stored in advance, and the scenes of the scene are comprehensively compared with the pattern data pre-stored in the pattern database 340 with the images converted in the Fourier transform step S26 or the edge extraction step S28. Can be analyzed.

Next, the transmission determination step (S40) is normalized to determine the items that can determine the image analyzed in the image analysis step (S30) in the form of O / X so that the frequency of O generated during a certain unit time is a specific value Determining that the abnormal situation has occurred in the site when it is exceeded and relates to the step of determining to transmit the abnormal situation of the site to the central server 600 and the person in charge.

In more detail, in the transmission determination step (S40), first, the result analyzed in the image analysis step (S30) is normalized in the form of O / X and stored in the form of a circular queue, in the form of a circular queue In case of storing the analysis result, it is easy to insert or delete additional data when overflow occurs or else.

Next, the O / X-type analysis result stored in the form of a circular queue is judged to check the frequency of O during the unit time. When the frequency of O during the unit time is less than a specific value, an abnormal situation is caused. It may be determined that no occurrence occurs, and an abnormal situation may be determined when the frequency of O is greater than a specific value.

In addition, the analysis result that plays an important role in determining the abnormal situation, such as the presence of guns in the armed strength, may be determined that the abnormal situation occurred immediately if a specific weight is assigned to O.

Next, in the transmission step (S50), if it is determined that the current site situation should be transmitted as a result of transmission determination, the head office is provided with an image and a determination result of the site where an abnormal situation occurs using the transmission module 500. It relates to the step of transmitting to the central server 600 and the person in charge, wherein the central server 600 is provided with a display unit 610 is determined on the image and abnormal situation of the scene transmitted through the transmission module 500 It is configured to check in real time.

In addition, when it is determined that the abnormal situation in the transmission determination step (S40), in the transmission step (S50) by using the transmission module 500 notifies the predetermined accident occurrence officer by the SMS to determine the result of the quick action is made. It is configured to be.

Next, the alarm generating step (S60) is the central server 600 and the person in charge checks the image and the determination result transmitted in the transmission step (S50) when it is determined that the emergency situation in which the armed strength actually invaded, the central server Regarding the step of generating an alarm signal from the alarm generating unit 700 provided in the relevant institutions such as branches and police stations using the 600, the generation of such an alarm signal is determined by the central server 600 Since it is made by the judgment of the person who confirmed the occurrence of the accident from the central server 600, it is possible to prevent the malfunction of the alarm signal at the source.

Therefore, according to the automatic arming strength recognition and response system and method according to the present invention as described above, it is possible to quickly prevent the progress of armed robbery incidents while maintaining intimacy with the customer to provide a safer working environment Simple system configuration ensures high price competitiveness, and comprehensively analyzes guns of armed robbery who broke into public institutions such as banks, characteristics of the crime scene, and gestures of employees using surveillance cameras. By operating an alarm on the site, it is possible to stop the progress of the incident as soon as possible and to report the situation on the spot in real time to the head office and related organizations to take a quick response.

Although the preferred embodiments of the present invention have been described for illustrative purposes, those skilled in the art will readily appreciate that many modifications are possible in the exemplary embodiments without materially departing from the novel teachings and advantages of this invention.

The present invention relates to a system for automatically recognizing and responding to armed strength, and more specifically, to a firearm invading a public institution such as a bank using a surveillance camera. And automatic detection and response system for armed strength to stop the progress of the incident by triggering the alarm on the scene and report the situation to the headquarters and related organizations in real time and to take quick countermeasures. It is about.

100: CCTV camera 200: video conversion unit
210: HSV converter 220: noise removing unit
230: Fourier transform unit 240: edge extraction unit
300: image analysis unit 310: firearm recognition unit
320: background recognition unit 330: employee recognition unit
340: pattern database 350: analysis result database
400: control unit 410: queue storage unit
420: frequency measurement unit 500: transmission module
600: central server 610: display unit
700: alarm generator
S10: field shooting step S20: video conversion step
S22: HSV conversion step S24: noise removal step
S26: Fourier transform step S28: Edge extraction step
S30: Image analysis step S40: Transmission decision step
S50: Transmission step S60: Alarm generation step

Claims (9)

In the armed strength automatic recognition and response system,
CCTV cameras to capture the scene of the scene and collect video,
An image converter for converting an image captured by the CCTV camera into an image that can be analyzed;
An image analyzer for analyzing the image converted by the image converter;
A control unit for determining whether to transmit the analysis result by storing the result analyzed by the image analyzer in the form of a circular queue;
A transmission module for transmitting an analysis result by the control unit;
A central server for receiving a signal transmitted from the transmission module;
And an alarm generation unit for generating an alarm signal by the signal of the central server.
The method of claim 1,
The image conversion unit HSV conversion unit for converting the image taken by the CCTV camera in the form of HSV (Hue Saturation Value),
A noise removing unit for removing noise from the image converted into the HSV form,
A Fourier transform unit for converting an image from which noise is removed into a frequency domain;
And an edge extracting unit for extracting an edge image from the noise-removed image.
The method of claim 1,
The image analyzer analyzes the image converted by the image converter through an Ada-Boost algorithm and recognizes and detects the shape of the firearm, and the background to recognize and detect the characteristic background of the shooting site. Arm recognition strength automatic recognition and response system comprising a recognition unit and an employee recognition unit for recognizing and detecting the gesture of the employee.
The method of claim 3,
The image analyzing unit may further include a pattern database for storing pattern data to be used as basic data when analyzing an image using the Ada-Boost algorithm, and an analysis result database for storing the image analysis result using the Ada-Boost algorithm together with the image. Armed strength automatic recognition and response system.
The method of claim 1,
The control unit normalizes the analysis result of the image analysis unit in the form of O / X and stores the cue storage unit in the form of a circular cue, and checks the frequency of O among the analysis results stored in the form of the circular cue to determine whether to transmit the current situation. Armed strength automatic recognition and response system, characterized in that it comprises a frequency measuring unit for determining.
In the armed strength automatic recognition and response method,
On-site shooting step of shooting the situation of the site using a CCTV camera,
An image conversion step of receiving the captured image and converting the image into an image that can be analyzed by an image conversion unit;
An image analysis step of analyzing the image converted in the image conversion step through an ada-boost algorithm;
Determination step of normalizing the analyzed image in the form of O / X and storing it in the form of a circular queue, and determining whether to transmit the current situation by checking the frequency of O among the stored analysis results;
As a result of the transmission decision, if it is determined that the transmission is necessary, the transmission step of transmitting the image and the determination result of the current situation to the central server and the person in charge using the transmission module and
And an alarm generation step of generating an alarm signal by checking the transmitted image and the determination result by the central server and the person in charge.
The method according to claim 6,
The image conversion step,
HSV conversion step of converting the image taken by the CCTV camera into the form of HS S (Hue Saturation Value),
A noise removing step of removing noise from the image converted into the HSV form;
A Fourier transform step of converting the image from which the noise is removed to a frequency domain;
And an edge extraction step of extracting an edge image from the noise-removed image.
The method according to claim 6,
In the image analysis step, armed robbery is characterized by comprehensively recognizing the characteristic background of the armed robbery and the scene of the incident and the gesture of the employee through an Ada-Boost algorithm by referring to the pattern data of the accident situation stored in advance. Automatic recognition and countermeasures.
The method according to claim 6,
In the transmitting step, if abnormality is detected in the transmission determining step, the image and the determination result of the current situation are transmitted to the central server, and the designated personnel are automatically notified of the determination result by SMS. How to respond.

Images