KR102287335B1 - Apparatus and method for monitoring tail-gating - Google Patents

Abstract

The present invention relates to a gate companion pass monitoring apparatus and a method thereof, and the gate accompanying passage monitoring apparatus according to an embodiment of the present invention includes: an image collecting unit for collecting a gate image captured through CCTV; a panel open detection unit for detecting an open state of the sliding panel through image processing on the gate image; an image feature extracting unit for extracting image features for image classification of the gate image according to detecting an open state of the sliding panel; an image classification unit for classifying the gate image into categories divided according to the number of people passing through the gate by applying the image feature to a pre-stored deep learning-based learning model; and a notification unit for receiving the classification result of the gate image from the image classification unit and notifying the manager terminal of the occurrence of the accompanying entry/exit state.

Description

Gate companion pass monitoring device and method {APPARATUS AND METHOD FOR MONITORING TAIL-GATING}

The present invention relates to a gate companion pass monitoring device and method, and more particularly, to implement a deep learning-based learning model for an image taken near a gate installed in an access control area, and to use the same to evaluate the gate companion pass situation It relates to a gate companion pass monitoring device and method for blocking an attempt to pass through without physical restrictions and cost burden on the existing door control system by monitoring.

A speed gate is installed and operated within a building or business establishment, allowing only authorized persons to enter and controlling the access of unauthorized persons.

This speed gate reads the identity of the visitor through a previously issued authentication means such as an IC card or fingerprint, authenticates the identity of the visitor, and then implements the function of automatically opening the gate.

In addition, the speed gate is configured to recognize an authorized person and an unauthorized person, and to control the opening and closing of the door according to the recognition result.

However, such a speed gate has a security loophole in that when an authorized person passes through an entrance door, when an unauthorized person tries to tailgating by attaching directly behind it, it cannot distinguish or block it.

That is, the speed gate is not equipped with a control system capable of adequately controlling the speed gate according to a number of unauthorized access situations by an authorized person.

However, in the past, existing systems are often used without improvement due to problems such as physical restrictions on entrances and cost burden, and because of this, it is always exposed to a situation in which it is difficult to effectively block an unauthorized person from attempting to pass through.

Recently, speed gates have also been applied to subway ticket gates. In the case of subway ticket gates, since many people try to pass together for free rides without purchasing a ticket, subway operators are contemplating ways to solve the cost deficit caused by losses caused by free rides.

However, since the number of ticket gates in each subway station is quite large for subway operators, it is not easy to block access by improving the subway ticket gates due to cost problems.

Since this situation leads to losses due to free riding for subway operators and causes a cost deficit, it is necessary to prepare a plan to effectively block the case of free riding.

Therefore, in the existing access control system, there is a need for a method to block an attempt to pass together without physically restricting the entrance and burdening the cost.

Korean Patent Publication No. 10-2010428 (registered on August 7, 2019)

It is an object of the present invention to implement a deep learning-based learning model for the image taken near the gate installed in the access control area, and monitor the accompanying passage situation of the gate using this, without physical limitations and cost burden on the existing door control system. An object of the present invention is to provide a gate companion pass monitoring apparatus and method for blocking a case of attempting a companion pass.

Gate companion pass monitoring device according to an embodiment of the present invention, an image collecting unit for collecting the gate image taken through CCTV; a panel open detection unit for detecting an open state of the sliding panel through image processing on the gate image; an image feature extracting unit for extracting image features for image classification of the gate image according to detecting an open state of the sliding panel; an image classification unit for classifying the gate image into categories divided according to the number of people passing through the gate by applying the image feature to a pre-stored deep learning-based learning model; and a notification unit for receiving the classification result of the gate image from the image classification unit and notifying the manager terminal of the occurrence of the accompanying access state.

The panel open detection unit extracts an edge image through edge detection with respect to the gate image, and checks whether an edge of the sliding panel formed across both sides of the gate body in the corresponding edge image exists to determine the open state of the sliding panel It may be Kim Ji-ha.

The panel open detection unit may detect the closed state of the sliding panel when the edge feature of the sliding panel is present, and detect the sliding panel as the open state when the edge feature of the sliding panel does not exist.

The panel open detection unit may apply a line fitting algorithm to determine the edge feature of the sliding panel.

The panel open detection unit may detect an open state of the sliding panel by detecting a color of an LED lamp identified on the upper surface of the gate body through the gate image.

The panel open detection unit detects that the sliding panel is opened when the color of the LED lamp is red or green, and when the color of the LED lamp is white, the sliding panel is closed. It may be a case of detection.

The image feature extractor may define a filter region for extracting an image feature with respect to the gate image in advance so that the filter performs a stride.

The filter area is a rectangular area passing through the intersection of a gate body and a sliding panel that face each other and form a passage, and two end points located in a diagonal direction in the rectangular area, two end points located in a diagonal direction of the gate body It may be positioning.

The gate image may be captured to include the filter region.

The above categories are: the first case in which no person passes through the gate (non-human), the second case in which one person passes through the gate (single-human), and the third case in which two or more people pass through the gate ( multi-human).

The image classification unit generates a count index indicating the classification result of the gate image and transmits it to the notification unit, wherein '0' is assigned to the count index in the first case, and '1' is assigned to the second case is allocated, and '2' may be allocated in the third case.

The notification unit, when the count index transmitted from the image classification unit exceeds a threshold value, may be to notify the occurrence of the accompanying access state to the manager terminal.

The threshold value may be 1.

According to an embodiment, a database for storing and managing the photographing location of the gate image in a table format by matching the location information of the CCTV, location information of the access control area, and manager terminal information with each other; further comprising, the notification unit, Checking the manager terminal information through the database corresponding to the gate image may be to inform the manager terminal the occurrence of the accompanying access state.

The database may store the learning model generated through machine learning based on the amount of learning data for classifying the gate image by category according to the number of people passing through the gate.

In addition, according to an embodiment of the present invention (a) collecting the gate image taken through the CCTV; (b) detecting an open state of the sliding panel through image processing on the gate image; (c) extracting image features for image classification of the gate image according to detecting the open state of the sliding panel; (d) classifying the gate image into categories divided according to the number of people passing through the gate by applying the image feature to a pre-stored deep learning-based learning model; and (e) notifying the manager terminal of the occurrence of an accompanying entry/exit state according to the classification result of the gate image.

In the step (b), after extracting an edge image through edge detection with respect to the gate image, it is checked whether there is an edge of the sliding panel formed across both sides of the gate body in the corresponding edge image to open the sliding panel. It may be Kim Ji-ji of the state.

In step (c), a filter region for extracting image features is defined in advance with respect to the gate image, and the filter performs a stride.

The step (e) generates a count index indicating the classification result of the gate image, wherein, as for the count index, '0' is assigned in the first case and '1' is assigned in the second case, In the third case, '2' may be assigned.

In step (e), when the count index exceeds a threshold value, it may be to notify the occurrence of a companion access state to the manager terminal.

In addition, according to an embodiment of the present invention, there is provided a gate companion pass monitoring apparatus, comprising: at least one processor; and a memory for storing computer readable instructions, wherein the instructions, when executed by the at least one processor, cause the gate companion pass monitoring device to collect a gate image captured through CCTV, An open state of the sliding panel is detected through image processing on the gate image, and an image feature for image classification of the gate image is extracted according to the sensed open state of the sliding panel, and the image feature is stored in advance. Applied to a deep learning-based learning model to classify the gate image into categories divided according to the number of people passing through the gate, and to inform the manager terminal of the occurrence of an accompanying entry/exit state according to the classification result of the gate image there is.

The present invention implements a deep learning-based learning model for the captured images near the gate installed in the access control area, and monitors the accompanying passage situation of the gate using the same, thereby passing through the existing door control system without physical limitations and cost burden. You can block the attempt.

In addition, the present invention can provide installation convenience because the accompanying pass can be determined using only the image without interlocking with a separate authentication device.

1 is a view showing a gate companion pass monitoring device according to an embodiment of the present invention;
2 is a view showing a gate image;
3 is a view showing an image classification category;
4 is a view showing a gate companion pass monitoring method according to an embodiment of the present invention.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. However, detailed descriptions of well-known functions or configurations that may obscure the gist of the present invention in the following description and accompanying drawings will be omitted. Also, it should be noted that throughout the drawings, the same components are denoted by the same reference numerals as much as possible.

The terms or words used in the present specification and claims described below should not be construed as being limited to conventional or dictionary meanings, and the inventor shall appropriately define his or her invention in terms of the best way to describe it. Based on the principle that it can be done, it should be interpreted as meaning and concept consistent with the technical idea of the present invention.

Therefore, the embodiments described in this specification and the configurations shown in the drawings are only the most preferred embodiment of the present invention, and do not represent all the technical ideas of the present invention, so at the time of the present application, various It should be understood that there may be equivalents and variations.

In the accompanying drawings, some components are exaggerated, omitted, or schematically illustrated, and the size of each component does not fully reflect the actual size. The present invention is not limited by the relative size or spacing drawn in the accompanying drawings.

In the entire specification, when a part "includes" a certain component, it means that other components may be further included, rather than excluding other components, unless otherwise stated. In addition, when a part is said to be "connected" with another part, it includes not only the case where it is "directly connected" but also the case where it is "electrically connected" with another element interposed therebetween.

The singular expression includes the plural expression unless the context clearly dictates otherwise. Terms such as “comprise” or “have” are intended to designate that a feature, number, step, action, component, part, or combination thereof described in the specification is present, and includes one or more other features, numbers, or steps. , it should be understood that it does not preclude in advance the possibility of the existence or addition of an operation, component, part, or combination thereof.

Also, as used herein, the term “unit” refers to a hardware component such as software, FPGA, or ASIC, and “unit” performs certain roles. However, "part" is not meant to be limited to software or hardware. A “unit” may be configured to reside on an addressable storage medium and may be configured to refresh one or more processors. Thus, by way of example, “part” includes components such as software components, object-oriented software components, class components, and task components, processes, functions, properties, procedures, subroutines, segments of program code, drivers, firmware, microcode, circuitry, data, databases, data structures, tables, arrays and variables. The functionality provided within components and “parts” may be combined into a smaller number of components and “parts” or further divided into additional components and “parts”.

Hereinafter, with reference to the accompanying drawings, embodiments of the present invention will be described in detail so that those of ordinary skill in the art can easily carry out the embodiments of the present invention. However, the present invention may be embodied in many different forms and is not limited to the embodiments described herein. And in order to clearly explain the present invention in the drawings, parts irrelevant to the description are omitted, and similar reference numerals are attached to similar parts throughout the specification.

Hereinafter, preferred embodiments of the present invention will be described with reference to the accompanying drawings.

1 is a view showing a gate companion pass monitoring device according to an embodiment of the present invention, FIG. 2 is a view showing a gate image, and FIG. 3 is a view showing an image classification category.

As shown in Figure 1, the gate accompanying pass monitoring apparatus 100 according to an embodiment of the present invention implements a deep learning-based learning model for the image taken near the gate 10 installed in the access control area, and , it is possible to block the case of attempting a companion passage without physical restrictions and cost burden on the existing door control system by monitoring the accompanying passage situation of the gate 10 using the same.

Here, the access control zone may be a zone that controls access within a specific security zone, such as a building or business place, or an zone that controls free riding for a specific service, such as a subway exit or an entrance to a stadium.

In addition, the gate 10 may have various access control types, and is not limited to any one type. However, for convenience of description, the case of a speed gate will be described as an example. The gate 10 may include a sliding panel 12 that is opened and closed across a gate body 11 forming an access passage and a gate body 11 facing each other.

The security manager of the access control area may bring the manager terminal 20 interlocking with the gate companion pass monitoring device 100 to receive a monitoring result on the accompanying passage situation from the gate companion pass monitoring device 100 .

Here, the manager terminal 20 is a wired/wireless device capable of transmitting and receiving data with the gate companion pass monitoring apparatus 100 , and may be a desktop computer, a notebook computer, a tablet PC, a smart phone, a PDA, or the like.

Hereinafter, the gate accompanying pass monitoring apparatus 100 will be described in detail with reference to FIG. 1 .

The gate companion pass monitoring device 100 includes an image collection unit 110 , a panel open detection unit 120 , an image feature extraction unit 130 , an image classification unit 140 , a notification unit 150 , and a database 160 . include

First, the image collection unit 110 collects a gate image in the vicinity of the gate 10 photographed through the CCTV 30 installed in the access control area.

Here, since the CCTV 30 is a normal camera installed in the access control area, additional installation is not required if it is already installed. Such CCTV 30 may be mounted on the wall or ceiling of the access control area.

The image collection unit 110 corresponds to the location information of the CCTV 30 to the photographing location of the gate image. The location information of the CCTV 30 corresponds to the location information of the access control area. This is to identify the shooting location of the gate image collected from the plurality of access control areas.

That is, the photographing location of the gate image is stored in the database 160 by matching the location information of the CCTV 30, the location information of the access control area, and the security manager information (manager terminal information).

Next, the panel open detection unit 120 opens and closes between both sides of the gate body 11 facing each other through image processing on the gate image transmitted through the image collection unit 110 in the open state of the sliding panel 12 . to detect

That is, the panel open detection unit 120 extracts an edge image through edge detection with respect to the gate image transmitted through the image collection unit 110 , and then extracts an edge image from the edge image of the gate body 11 . ) The open state of the sliding panel 12 is detected by checking whether an edge of the sliding panel 12 formed across both sides exists.

Referring to FIG. 2 , in the corresponding edge image EI, the edge GE of the gate body 11 and the edge SE of the sliding panel 12 are orthogonal to each other. The panel open detection unit 120 checks whether the edge SE of the sliding panel 12 is present.

Here, the panel open detection unit 120 detects that the sliding panel 12 is closed when the edge feature of the sliding panel 12 is present, and when the edge feature of the sliding panel 12 does not exist, the sliding panel 12 is not present. ) is detected as open.

In this case, the panel open detection unit 120 may apply a line fitting algorithm to determine the edge feature of the sliding panel 12 .

In another embodiment, the panel open detection unit 120 detects the color of the LED lamp 13 identified on the upper surface of the gate body 11 through the gate image transmitted through the image collection unit 110 to the sliding panel. (12) detect the open state.

For example, the color of the LED lamp 13 formed on the upper surface of the sliding panel 12 is red or green when the sliding panel 12 is opened, and the sliding panel 12 is This closed case may be displayed in white. Accordingly, when the color of the LED lamp 13 is red or green, the panel open detection unit 120 detects that the sliding panel 12 is open.

Next, when the image feature extraction unit 130 detects the open state of the sliding panel 12 by the panel open detection unit 120 , the image feature of the gate image transmitted through the image collection unit 110 . to extract

The image feature extraction unit 130 is configured to classify the gate image transmitted through the image collection unit 110 when the sliding panel 12 is detected in an open state by the panel open detection unit 120 . The process of extracting image features applied in the deep learning model is performed.

Here, the deep learning model includes element technologies using machine learning, and a convolutional neural network (CNN) model can be applied.

The image feature extracting unit 130 pre-defines a filter region F for extracting image features from the gate image transmitted through the image collecting unit 110 .

The filter area (F) is an area in which a stride is performed while a filter (or kernel) for extracting image features is swept while performing a convolutional layer by applying a convolutional neural network model.

Referring to FIG. 2 , the filter area F is a rectangular area passing through intersections Pa and Pb of the gate body 11 and the sliding panel 12 forming a passage while facing each other. The two end points located in the diagonal direction may be located at the end points Pc and Pd located in the diagonal direction of the gate body 11 .

As such, the gate image is captured to include the filter area F that can detect the open state of the sliding panel 12 and can count the number of people passing through the gate 10 .

For example, assuming that the CCTV 30 sets the center of the passage of the gate 10 (that is, the point where the sliding panel abuts) as a reference point, and forms a hemispherical coordinate system around the reference point, a predetermined angle or more from the ground It is preferable to form the shooting angle θ with

The shooting angle θ of the CCTV 30 in the hemispherical coordinate system may be between 0°≤θ≤180°, preferably 45°≤θ≤135°. In the extreme, if the angle of capture of the CCTV 30 is 0°, the filter area F is not included in the gate image because it is in a state parallel to the ground.

Next, the image classification unit 140 applies the image features extracted from the image feature extraction unit 130 to the learning model stored in advance in the database 160 and gates the gate image transmitted through the image collection unit 110 . 10) is classified into categories divided according to the number of people passing through.

Referring to FIG. 3 , categories are classified according to the number of people passing through the gate 10 , for example, 'non-human' (a), which is a case where no person passes through the gate 10, and the gate 10 ) can be divided into 'single-human' (b) when there is one person passing through, and 'multi-human' (c) when there are two or more people passing through the gate 10 . Here, the case of 'multi-human' (c) is a case where it is determined that the entry and exit are accompanied.

Then, the image classification unit 140 classifies the gate image into a category, and then counts the number of people passing through the gate 10 to check the count index, so that the accompanying pass state of the gate 10 is determined by the security manager's manager. It can be notified to the terminal 20 .

At this time, the image classification unit 140 generates a count index indicating the classification result of the gate image and transmits it to the notification unit 150 . Here, the count index may be assigned to '0' for non-human, '1' for single-human, and '2' for multi-human.

Accordingly, the notification unit 150 receives the count index through the image classification unit 140 , and when the count index exceeds the threshold value, the notification unit 150 notifies the occurrence of the accompanying entry/exit state to the manager terminal 20 of the security manager. Here, the threshold value of the count index is 1.

At this time, the notification unit 150 can confirm the security manager information stored in the database 160 by checking the location information of the access control area corresponding to the gate image, so that the occurrence of the accompanying access state to the manager terminal 20 of the security manager can tell you

Next, the database 160 stores a learning model for performing image classification based on a deep learning model on the gate image.

This learning model is generated by performing machine learning based on a sufficient amount of learning data to classify the gate image by category according to the number of people passing through the gate 10 . Here, the process of generating the learning model is the same as the process performed by the image feature extracting unit 130 and the image classifying unit 140 described above, and supervised learning using the number of people passing through the gate 10 as a target variable. Machine learning can be performed based on (supervised learning).

In addition, the database 160 may store and manage the location information of the CCTV 30, the location information of the access control area, and the security manager information (manager terminal information) in a table format by correspondence with each other as described above.

Accordingly, the gate companion pass monitoring apparatus 100 may monitor and block the accompanying pass state in any access control area even if the accompanying pass monitoring is performed for a plurality of access control areas instead of a single access control area.

On the other hand, as another embodiment, the gate companion pass monitoring apparatus 100 may perform the gate companion pass monitoring method according to the embodiment of the present invention when executing computer readable instructions stored in the memory by at least one or more processors. .

Here, the processor is at least one processor, and may also be referred to as a controller, a microcontroller, a microprocessor, a microcomputer, or the like. In addition, the processor may be implemented by hardware or firmware, software, or a combination thereof.

Also, a memory may be a single storage device, or may be a collective term for a plurality of storage elements. The computer readable instructions stored in the memory may be executable program code or parameters, data, and the like. In addition, the memory may include random access memory (RAM) or non-volatile memory (NVRAM) such as magnetic disk storage or flash memory.

4 is a view showing a gate companion pass monitoring method according to an embodiment of the present invention.

First, the gate companion pass monitoring device 100 collects a gate image in the vicinity of the gate 10 photographed through the CCTV 30 installed in the access control area (S201).

Thereafter, the gate accompanying pass monitoring device 100 checks whether the sliding panel 12 of the gate 10 is in an open state, that is, the open state of the gate 10 ( S202 ). At this time, the gate accompanying pass monitoring device 100 extracts an edge image through edge detection for the gate image, and then the edge of the sliding panel 12 formed across both sides of the gate body 11 in the corresponding edge image is present. The open state of the sliding panel 12 is detected by checking whether the

Then, when the sliding panel is in an open state (S202), the gate companion pass monitoring apparatus 100 extracts image features for the gate image to classify the image into an appropriate category for the gate image (S203).

At this time, the gate companion pass monitoring apparatus 100 applies a learning model to the image feature and classifies it into an appropriate category for the gate image (S204). Here, the category may be divided according to the number of people passing through the gate 10 .

Then, the gate companion pass monitoring device 100 classifies the gate image into categories (S204), and then counts the number of people passing through the gate 10 to check the count index, so the accompanying pass state of the gate 10 can be informed to the manager terminal 20 of the security manager (S205 to S207).

At this time, the gate companion pass monitoring apparatus 100 checks the count index according to the image classification, and when the count index exceeds a predetermined threshold value (S205, S206), it may inform the manager terminal 20 of the accompanying passage state. (S207).

The method according to some embodiments may be implemented in the form of program instructions that can be executed through various computer means and recorded in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc. alone or in combination. The program instructions recorded on the medium may be specially designed and configured for the present invention, or may be known and available to those skilled in the art of computer software. Examples of the computer-readable recording medium include magnetic media such as hard disks, floppy disks and magnetic tapes, optical media such as CDROMs and DVDs, and magneto-optical disks such as floppy disks. hardware devices specially configured to store and execute program instructions, such as magneto-optical media, and ROM, RAM, flash memory, and the like. Examples of program instructions include not only machine language codes such as those generated by a compiler, but also high-level language codes that can be executed by a computer using an interpreter or the like.

Although the foregoing description has focused on novel features of the invention as applied to various embodiments, those skilled in the art will recognize the apparatus and method described above without departing from the scope of the invention. It will be understood that various deletions, substitutions, and changes are possible in the form and details of Accordingly, the scope of the invention is defined by the appended claims rather than by the description above. All modifications within the scope of equivalents of the claims are included in the scope of the present invention.

10 ; gate 11 ; gate body
12 ; sliding panel 13 ; LED lamp
20 ; manager terminal 30 ; CCTV
10 ; image collecting unit 120 ; panel open sensor
130 ; image feature extraction unit 140 ; image classification unit
150 ; notification unit 160 ; database

Claims (23)

an image collection unit for collecting the gate image taken through CCTV;
a panel open detection unit for detecting an open state of the sliding panel through image processing on the gate image;
an image feature extracting unit for extracting image features for image classification of the gate image according to detecting an open state of the sliding panel;
an image classification unit for classifying the gate image into categories divided according to the number of people passing through the gate by applying the image feature to a pre-stored deep learning-based learning model; and
a notification unit for receiving the classification result of the gate image from the image classification unit and notifying the administrator terminal of the occurrence of an accompanying entry/exit state;
Gate companion pass monitoring device comprising a.
The method of claim 1,
The panel open detection unit,
After extracting an edge image through edge detection with respect to the gate image, it is checked whether an edge of the sliding panel formed across both sides of the gate body in the corresponding edge image exists to determine the open state of the sliding panel. companion pass monitoring device.
3. The method of claim 2,
The panel open detection unit,
If there is an edge feature of the sliding panel, it is detected that the sliding panel is closed,
If the edge feature of the sliding panel does not exist, it is sensed that the sliding panel is in an open state.
4. The method of claim 3,
The panel open detection unit,
A gate companion pass monitoring device that applies a line fitting algorithm to determine the edge characteristics of the sliding panel.
The method of claim 1,
The panel open detection unit,
A gate companion pass monitoring device that detects the open state of the sliding panel by detecting the color of the LED lamp identified on the upper surface of the gate body through the gate image.
6. The method of claim 5,
The panel open detection unit,
If the color of the LED lamp is red or green, it is detected that the sliding panel is open,
If the color of the LED lamp is white, the gate accompanying pass monitoring device detects that the sliding panel is closed.
The method of claim 1,
The image feature extraction unit,
A gate companion pass monitoring device that performs a stride by defining a filter region for extracting image features for the gate image in advance.
8. The method of claim 7,
The filter area is
A rectangular area passing through the intersection of a gate body and a sliding panel forming a passage while facing each other, wherein two end points located in a diagonal direction in the rectangular area are positioned in a diagonal direction of the gate body companion pass monitoring device.
9. The method of claim 8,
The gate image is
Gate companion pass monitoring device that is photographed so as to include the filter area.
The method of claim 1,
The category is
Case 1 with no person passing through the gate (non-human), Case 2 with one person passing through the gate (single-human), Case 3 with 2 or more people passing through the gate (multi-human) Gate companion pass monitoring device that is distinguished by .
11. The method of claim 10,
The image classification unit,
A count index indicating the classification result of the gate image is generated and delivered to the notification unit,
The count index is,
and '0' is assigned to the first case, '1' is assigned to the second case, and '2' is assigned to the third case.
12. The method of claim 11,
The notification unit,
When the count index transmitted from the image classification unit exceeds a threshold value, the gate companion pass monitoring device that notifies the occurrence of a companion access state to the manager terminal.
13. The method of claim 12,
The threshold value is 1 gate companion pass monitoring device.
14. The method of claim 13,
A database for storing and managing the photographing location of the gate image in a table format by correspondence with the location information of the CCTV, location information of the access control area, and information on the manager terminal with each other;
The notification unit,
The gate companion pass monitoring device for checking the manager terminal information through the database corresponding to the gate image and notifying the manager terminal of the occurrence of the accompanying entry/exit state.
15. The method of claim 14,
The database is
A gate companion pass monitoring device that stores the learning model generated through machine learning based on the amount of learning data for classifying the gate image by category according to the number of people passing through the gate.
(a) collecting the gate image taken through CCTV;
(b) detecting an open state of the sliding panel through image processing on the gate image;
(c) extracting image features for image classification of the gate image according to detecting the open state of the sliding panel;
(d) classifying the gate image into categories divided according to the number of people passing through the gate by applying the image feature to a pre-stored deep learning-based learning model; and
(e) notifying a manager terminal of occurrence of an accompanying entry/exit state according to the classification result of the gate image;
A gate companion pass monitoring method comprising a.
17. The method of claim 16,
Step (b) is,
After extracting an edge image through edge detection with respect to the gate image, it is checked whether an edge of the sliding panel formed across both sides of the gate body in the corresponding edge image exists to determine the open state of the sliding panel. How to monitor the pass-through.
17. The method of claim 16,
The step (c) is,
The gate accompanying pass monitoring method, wherein the filter performs a stride by defining a filter region for extracting image features for the gate image in advance.
19. The method of claim 18,
The filter area is
A rectangular area passing through the intersection of a gate body and a sliding panel forming a passage while facing each other, wherein two end points located in a diagonal direction in the rectangular area are positioned in a diagonal direction of the gate body How to monitor the pass-through.
17. The method of claim 16,
The category is
Case 1 with no person passing through the gate (non-human), Case 2 with one person passing through the gate (single-human), Case 3 with 2 or more people passing through the gate (multi-human) A gate companion pass monitoring method that is distinguished by .
21. The method of claim 20,
Step (e) is,
Generates a count index indicating the classification result of the gate image,
The count index is,
and '0' is assigned to the first case, '1' is assigned to the second case, and '2' is assigned to the third case.
22. The method of claim 21,
Step (e) is,
When the count index exceeds a threshold value, the gate companion pass monitoring method of notifying the occurrence of an accompanying entrance condition to the manager terminal.
A gate companion pass monitoring device comprising:
at least one processor; and
a memory for storing computer readable instructions;
The instructions, when executed by the at least one processor, cause the gate companion pass monitoring device to:
Gather the gate images filmed through CCTV,
It detects the open state of the sliding panel through image processing on the gate image,
Extracting image features for image classification of the gate image according to detecting the open state of the sliding panel,
By applying the image feature to a pre-stored deep learning-based learning model, the gate image is classified into categories divided according to the number of people passing through the gate,
The gate companion pass monitoring device to notify the manager terminal of the occurrence of the accompanying entrance condition according to the classification result of the gate image.

Images