KR20230169497A - Access image storage apparatus for analyzing access trace of unauthorized visitor - Google Patents

Abstract

An access video storage device capable of analyzing the entry and exit movements of unauthorized visitors is disclosed. The access video storage device according to the present invention includes a camera installed in a surveillance area to photograph the surveillance area and output the captured video; An access control device including an access recognition device that obtains and outputs access authentication information from the visitor, a door lock device that locks or unlocks the door, and opens the door when a door open control signal is input; and an access control video storage control device that stores video captured through a camera and controls entry and exit of visitors through an access control device.

Description

Access image storage device that can analyze the entry and exit movement of unauthorized persons {ACCESS IMAGE STORAGE APPARATUS FOR ANALYZING ACCESS TRACE OF UNAUTHORIZED VISITOR}

The present invention relates to an access video storage device, and more specifically, to a device equipped with a digital video recording (DVR) function that stores video in conjunction with perimeter surveillance and entrance control, and to record video based on walking. Security and safety are achieved by performing more thorough external and internal access control through object detection through gender and age identification, visual change, and learning functions, and by clearly analyzing the entry and exit paths of unauthorized persons using captured images. It relates to an access video storage device that can improve.

In general, various types of security systems are installed in spatial areas such as apartments, buildings, parks, roads, and government offices to protect individual lives and the assets of individuals, companies, and countries.

These security systems include Digital Video Record (DVR) and Entrance Control Equipment.

Digital video processing devices (DVR) are installed in certain areas such as houses, buildings, government offices, parks, and roads where the assets of individuals, companies, and the nation are located (hereinafter referred to as “surveillance areas”) to protect the assets of individuals, companies, and the nation. A number of cameras, i.e. CCTVs, are installed to film the surveillance area in each sub-region unit, and each video captured through the cameras is displayed and stored. The video captured in this way is displayed on a display device in real time, allowing security managers to monitor the surveillance area, allowing early detection of crimes, and is used as evidence for crimes that have occurred.

Figure 1 is a diagram showing the configuration of a digital image storage system to which a general digital image storage device is applied.

Referring to FIG. 1, the image storage system 1 includes a plurality of cameras (CCTV: 30) installed in a certain sub-area unit in a target area to be monitored, that is, a surveillance area, and a display that displays images captured by the cameras 30. Digital image storage that stores images captured through the device 20 and the cameras 30 and displays or divides at least one image selected by a security manager among the captured images on the display device 20 It includes device 10.

The digital image storage device 10 typically has a function for storing image data captured in a surveillance area, as well as a fire detection function through analysis of the captured image data.

In addition, access control devices are installed at the entrances of buildings such as houses, apartments, buildings, and government offices, and at each entrance within the building, allowing only authorized persons to enter through the entrance and preventing unauthorized third parties from entering the entrance. Prevent tangible and intangible property from being stolen by third parties.

Figure 2 is a diagram showing the configuration of an access control system to which a general access control device is applied.

Referring to FIG. 2, the access control system 100 is a door lock device that includes a door control unit 122 that is installed on the door of the entrance and locks or unlocks the door, and an access recognition unit 121 that recognizes and outputs visitor information. (120) and the visitor information for approved visitors is registered and stored in advance, and the visitor information is input from the access recognition unit 121 and the pre-stored visitor information is checked to see whether the same visitor information registered in advance exists. It includes an access control device 110 that controls the door control unit 122 to unlock the door.

As described above, the conventional digital image storage system and access control system are separate systems, so there was a problem in that both the digital image system and the access control system had to be installed in order to monitor and control access to the security area.

Because both a digital imaging system and an access control system had to be installed, there was a problem in that the cost of building a security system was high.

In addition, since both the digital imaging system and the access control system must be installed, the number of wires connecting the devices of each system increases, making installation difficult, which may increase installation costs.

In addition, since both the digital imaging system and the access control system had to be installed, there was a problem of having to allocate a lot of space for the arrangement of the digital imaging device and the access control device.

Registered Patent Publication No. 10-1387628 (Registration Date: 2014.04.15)

The present invention was created to solve the above-described problems, and has a digital video recording (DVR) function that stores images in conjunction with perimeter surveillance and access control, and provides a walking-based External and internal access control is performed more thoroughly through object detection through gender and age identification, visual change, and learning functions, and security and safety are improved by clearly analyzing the entry and exit movements of unauthorized persons using captured images. The purpose is to provide an access video storage device that can be improved.

An access video storage device according to an aspect of the present invention for achieving the above-described object includes a camera installed in a surveillance area, photographing the surveillance area, and outputting the captured image; An access control device including an access recognition device that obtains and outputs access authentication information from the visitor, a door lock device that locks or unlocks the door, and opens the door when a door open control signal is input; and an access control image storage control device that stores images captured through the camera and controls entry and exit of visitors through the access control device, wherein the access control image storage control device includes images captured through the camera. Includes a video DB that stores data by mapping it to the camera information, a visitor authentication information DB that stores visitor authentication information for authorized persons for each security target entrance in the surveillance area, and walking information according to gender and age. a storage unit including a walking information DB; A communication unit connected to the access control device to receive and output visitor authentication information from the access recognition device, transmit an input control signal to the door lock device, and receive and output an image from the camera; And an image storage module that controls the camera through the communication unit, receives images captured by the cameras and stores them in the image DB, and acquires images directly from the camera and the access control device or through an image storage module, Determine whether there is an object in each frame of the captured image, and if there is an object, detect the object with the most reliable movement using the location and median of the object, and perform face recognition if the detected object is a person. , a video object tracking module that performs authentication by comparing it with a face pre-registered in the visitor authentication information DB and outputs the result, and a visitor authentication information of an arbitrary visitor from the access recognition device for which an access request has been generated through the communication unit. The user is an authorized user by receiving input and comparing the user authentication information of those authorized for access to the entrance with the access recognition device installed in advance registered in the user authentication information DB, and the authentication result information by facial recognition from the video object tracking module indicates that authentication is successful. It is characterized in that it includes a control unit including an access control module that opens the door by outputting a door open control signal to the door lock device through the communication unit.

The image object tracking module includes an image acquisition unit that acquires an image directly from the camera or acquires an image through the image storage module; Determine whether there is an object in each frame of the captured image, and if there is an object, detect and track the object with the most reliable movement using the location and median of the object, and determine whether the detected object is a person. an object tracking unit that detects a face if it is a person; And a situation determination unit that performs facial recognition on the detected face, performs authentication by comparing it with a face pre-registered in the visitor authentication information DB, and outputs the result to the access control module.

The object tracking unit includes an object detection unit that determines whether an object exists in each frame of a captured image and, if an object exists, detects and tracks the object with the most reliable movement using the location and median of the object; and a face detection unit that detects and outputs a face if the detected object is a person.

The storage unit further includes a gait information DB that stores body characteristic information and gait characteristic information by gender and age, and the object tracking unit determines whether the detected object is a person from the object image, and if it is a person, the gait information It further includes a star detection unit that estimates the gender and age of the object based on the walking image of the object with reference to the DB and outputs object gender and age information, wherein the situation determination unit includes object gender and age information input from the gender detection unit. Authentication is performed by further comparing the gender and age information of the authorized person pre-registered in the visitor authentication information DB.

The object detection unit includes a learning unit that analyzes the input image using learned learning information and outputs image analysis data for the object and background; a detection unit that scans the image analysis data in units of scanning windows to determine whether an object exists in the image and outputs the result and a bounding box containing the object; a tracking unit that represents an object by a bounding box and estimates the movement of the object between consecutive frames based on an extended median flow tracking method; and a data integration unit that combines and outputs bounding boxes output from the detection unit and the tracking unit.

The learning unit includes a classification unit that classifies data not confirmed by the tracking unit into objects and backgrounds and then modifies classification information of a previous frame for the input object; an expert execution unit that receives classification information for unconfirmed data from the unconfirmed data and classification unit, and executes P-expert and N-expert to generate expert analysis information for objects and backgrounds; Analysis settings that receive confirmed data from the detection unit, train the input confirmed data, store training examples, and update the training set by receiving information detected by the expert execution unit. wealth; and a training unit that receives the training set output from the analysis setting unit, trains it, and then generates learning information (classifier parameters).

The video object tracking module acquires voice through a microphone included in the access control device or a microphone installed in pairs with the camera, and determines whether the acquired voice is a general voice or an event voice according to a specific situation. It further includes a voice acquisition unit that outputs the result, wherein the situation determination unit determines the situation of the location where the voice was acquired according to the voice determination result output from the voice acquisition unit and provides situation determination information including the result information. It is characterized by output.

The control unit designates a flame candidate area using various color space model values, receives at least one of classification information and object state information output from the object detection unit, and determines whether the object in the flame candidate area is a flame. It is characterized in that it further includes a fire monitoring module.

The camera may include an outer camera installed on the outskirts of the surveillance area; and an internal camera installed inside a building in the surveillance area.

The device further includes: an exterior intrusion detection sensor installed on the outskirts of the surveillance area to detect an intruder entering the surveillance area from the outside, and the control unit operates on the outskirts of the surveillance area through the exterior intrusion detection sensor. It further includes an outer surveillance module that detects an intruder and stores the video captured through the outer camera in the video DB, wherein the video storage module of the control unit stores the video captured through the inner camera in the video DB. It is characterized by saving.

The device further includes: a fire detection sensor installed in the surveillance area to detect a fire occurring in the surveillance area and output a fire detection signal, and the storage unit includes the exterior camera, the interior camera, and the fire detection sensor. It further includes a detector information DB that stores detector information including installed location information, wherein the control unit determines the location of the fire detection sensor by referring to the detector information DB when detecting a fire through an arbitrary fire detection sensor, and It may further include a fire monitoring module that transmits the identified location information to at least one of the image storage module and the perimeter monitoring module to photograph the area where the fire occurred.

The device further includes: an RFID reader installed at the parking lot entrance and parking lot floor movement path of the surveillance area, which is attached to any vehicle and reads and outputs the vehicle information from an RFID tag storing the vehicle information of the vehicle, The storage unit further includes a detector information DB that stores sensor information including location information where the RFID reader is installed, and the control unit detects a vehicle entering or exiting the parking lot by obtaining vehicle information of an arbitrary vehicle through the RFID reader. It is characterized in that it further includes a parking control module that controls parking by determining the entrance and exit of the vehicle.

The storage unit further includes a movement tracking DB that stores visitor authentication information and information and images of the access recognition device through which the visitor attempted to enter, and the control unit stores visitor authentication information recognized through an arbitrary access recognition device. If the person entering is determined to be an unauthorized person or an intruder for the access recognition device more than a certain number of times, the access recognition device information of the two access recognition devices through which the unauthorized person and the intruder attempted to enter is provided to the video storage module to prevent the unauthorized person and the intruder. Receives a video captured by a camera near the person who attempted to enter or a video of the time period passing the two access recognition devices through at least one camera between the two access recognition devices and stores it in the movement trace DB, and unauthorized person and a movement tracking module that sequentially stores location information of entry/exit recognition devices used by the intruder and tracks the movement path of the unauthorized person and the intruder.

When the fire monitoring module detects a fire through the fire detection sensor, it transmits a fire notification signal including location information of the fire to the access control module, and the access control module generates a fire when a fire notification signal is input from the fire monitoring module. It is characterized by disabling the access control mode of the access recognition device installed at the low-level entrance that can escape from the location to the outside.

The access control device is installed on at least one type of door, revolving door, elevator, and escalator.

According to the present invention, since it is equipped with both exterior and interior access control, unauthorized person's movement tracking function, and video surveillance function, it is possible to more accurately control unauthorized person's entry and intrusion from the outside and inside of the surveillance area. By tracking the person's movements and securing the location and video, security managers can quickly respond to intrusions by unauthorized persons. In other words, the present invention has the effect of increasing the security of the surveillance area.

Figure 1 is a diagram showing the configuration of a digital image storage system to which a general digital image storage device is applied.
Figure 2 is a diagram showing the configuration of a digital image storage system to which a general access control device is applied.
Figure 3 is a diagram showing the configuration of an access control image storage device according to the present invention.
Figure 4 is a diagram showing the configuration of an access control video storage control device according to the present invention.
Figure 5 is a diagram showing the configuration of the video object tracking module of the access control video storage and control device according to the present invention.
Figure 6 is a diagram showing the configuration of the object detection unit of the video object tracking module according to the present invention.
Figure 7 is a diagram showing the configuration of the learning unit of the object detection unit according to the present invention.
Figure 8 is a flowchart showing a control method of an access control video storage device according to the present invention.
Figure 9 is a flowchart showing a method of tracking an intruder's movements in an access control video storage device according to the present invention.
Figure 10 is a diagram for explaining a method for analyzing gender and age based on a person's gait according to the present invention.
Figure 11 is a diagram showing the characteristics of a scream and a general voice according to the present invention.

Hereinafter, some embodiments of the present invention will be described through exemplary drawings. When assigning reference numerals to components in each drawing, identical components are indicated with the same reference numerals as much as possible, even if they are shown in different drawings. Additionally, when describing embodiments of the present invention, if detailed descriptions of related known configurations or functions are judged to impede understanding of the embodiments of the present invention, the detailed descriptions will be omitted.

Additionally, when describing the components of an embodiment of the present invention, terms such as first, second, A, B, (a), and (b) may be used. These terms are only used to distinguish the component from other components, and the nature, sequence, or order of the component is not limited by the term. When a component is described as being "connected," "coupled," or "connected" to another component, that component may be directly connected, coupled, or connected to that other component, but may not be connected to that component or that other component. It should be understood that another component may be “connected,” “coupled,” or “connected” between elements.

Figure 3 is a diagram showing the configuration of an access control image storage device according to the present invention.

Referring to FIG. 3, the access control image storage device according to the present invention includes a display device 300, an exterior camera 310, an interior camera 311, an access recognition device 321 including a camera, and a door lock device 322. An access control device 320 including an access recognition device 331 without a camera and a plurality of access control devices 330 including a door lock device 332, an exterior intrusion detection sensor 340, and a fire detection sensor ( 350) Includes an RFID reader 360 and an administrator terminal 370 that read information from an RFID tag.

The access control video storage control device 200 includes a display device 300, an external camera 310, an internal camera 311, an access recognition device 321 equipped with a camera, and a door lock device 322. A device 320, a plurality of access control devices 330 including an access recognition device 331 and a door lock device 332, an exterior intrusion detection sensor 340, and a fire detection sensor 350 read information from RFID tags. is connected to the RFID reader 360 and the manager terminal 370.

The outer camera 310 and the inner camera 311 transmit analog video captured through analog communication to the access control video storage control device 200 if it is an analog video camera, and if it is an Internet Protocol (IP) camera, Digital image data is transmitted to the access control video storage control device 200 through TCP/IP communication.

The access control devices 320 and 330 include access recognition devices 321 and 331 and door lock devices 322 and 332, as described above.

The access control devices 320 and 330 may be installed at each entrance, such as a door, a revolving door, an escalator, and an elevator. The access control devices 320 and 330 support at least one of serial communication and TCP/IP communication protocols, and are connected to the access control video storage control device 200 through either serial communication or TCP/IP communication to perform data communication. do.

However, even if the access control devices (320, 330) support only serial communication, they are equipped with a serial communication-TCP/IP communication conversion repeater (not shown) to provide access control video storage control device (200). ) and data communication using the TCP/IP communication protocol.

The access recognition devices 321 and 331 obtain authentication information about the visitor and provide it to the access control video storage control device 200. Then, the access control video storage control device 200 compares the input authentication information with the authentication information of the access authorized person pre-registered for the entrance where the access recognition devices 321 and 331 are installed, and if there is matching information, It is determined that the person has a normal authorization for the entrance, and the door is opened by controlling the access recognition device (321, 331), or the door lock device (322, 332) is unlocked so that the escalator and elevator operate.

At this time, the access control image storage control device 200 detects the object by separating the background and the object from the acquired image, changes the perspective or corrects the curvature of the detected object, and performs gender, age, and face identification in the case of a person. It can be configured to control access. The method for distinguishing between men and women includes physical structures such as height, head, hairstyle, chest, back, waist, hips, and legs of men and women through repeated acquisition using Gait analysis, an object tracking algorithm. It could be configured to collect data on walking, such as stride speed, stride distance, etc., and distinguish between men and women through program learning to restrict or permit human access control functions.

The outer intrusion detection sensor 340 is installed on the outskirts of the surveillance area, detects intruders entering the surveillance area, and provides an intrusion detection signal to the access control video storage control device 200. It would be desirable for the exterior intrusion detection sensor 340 to be installed at a location corresponding to the shooting direction of the exterior camera 310.

The fire detection sensor 350 is installed inside and outside the building in the surveillance area to monitor whether a fire occurs, and when a fire is detected, it outputs a fire detection signal to the access control video storage control device 200.

The RFID reader 360 is installed at the parking lot entrance and each floor entrance to read vehicle information from vehicles attached with RFID tags that store vehicle information and output it to the access control image storage and control device 200.

The access control video storage control device 200 according to the present invention is equipped with an access control function, a video storage function, a perimeter surveillance function, a fire surveillance function, a parking control function, and an unauthorized person's movement line tracking function.

The access control video storage control device 200 displays at least one video selected by a security manager, etc. among the images received from a plurality of external cameras 310 and internal cameras 311 on the display device 300 or displays them on the administrator terminal. Provided as (370). The images will be stored in the access control image storage control device 200. In addition, the access control video storage control device 200 selectively selects the access recognition device 331 of the access control device 330, operates the camera of the selected access recognition device 331 to perform photography, and performs filming. It may be configured to provide and store images to the display device 300 and the administrator terminal 370.

The access control video storage control device 200 includes an external camera 310, an internal camera 311, an access recognition device 321, 331, a fire detection sensor 350, and an RFID reader 360 (hereinafter referred to as camera, access recognition device). , “detector” is used as a term that includes fire detection sensors and RFID readers), stores information on the installed location of each detector, and is approved through the access recognition devices 321 and 331 at entrances and exits at random locations. The person who requested the access is an unauthorized person, and the unauthorized person's movement path is monitored through the access recognition devices (321, 331) of the entrance that the accessor attempted more than a certain number of times through the access recognition device and the cameras (310, 311) between them. Track.

The configuration and operation of the access control video storage control device 200 will be described in detail with reference to FIG. 4 below.

Figure 4 is a diagram showing the configuration of an access control video storage control device according to the present invention.

The access control video storage control device 200 of the present invention includes a control unit 210, a storage unit 220, an interface unit 230, an image processing unit 240, and a communication unit 260.

The control unit 210 controls the overall operation of the access control video storage control device 200 according to the present invention. The detailed configuration and operation of the control unit 210 will be described later.

The storage unit 220 includes an image DB 221, visitor authentication information DB 222, detector information DB 223, access record DB 224, movement line tracking DB 225, and walking information DB 226. .

The image DB 221 stores image information including captured image data captured by the outer camera 310 and the inner camera 311. The video information includes information about the camera that captured the video data. The camera information may be a unique identification number of the camera.

The visitor authentication information DB 222 stores authentication information for visitors. The authentication information includes personal information of the visitor and authentication information of the visitor. The personal information includes the name of the person entering, identification information such as resident registration number and employee number, gender and age, and the authentication information of the person entering may be a password for each entrance, biometric information of the person entering, and a password for each person entering.

The detector information DB 223 includes detectors (exterior camera 310, interior camera 311, entry/exit recognition devices 321, 331, exterior intrusion detection sensor 340, fire detection sensor 350, RFID reader ( 360) Each identification information and location information are stored. In addition, the detector information DB 223 may further store level information of the entrance where the access recognition devices 321 and 331 are installed.

The access record DB 234 stores access records for each access recognition device 321 and 331, including visitor information, whether access is authorized, and attempted access time information for each person who attempted access.

The movement line tracking DB 225 stores the identification information of the access recognition devices 321 and 331 through which an unauthorized person attempted entry among the visitors who attempted to enter through the entrance where the access recognition devices 321 and 331 were installed, according to the order of entry and exit attempted by the unauthorized person. Stores video data between the access recognition devices 321 and 331 where unauthorized persons enter.

The gait information DB 226 stores physical characteristic information and gait characteristic information by gender and age. The physical characteristic information may include height, head, hairstyle, chest, back, waist, hips, and leg shape information. In addition, the walking characteristic information may include walking form information such as stride length, body tilt when walking, walking speed, and movement of the hips, legs, and chest while walking.

The interface unit 230 is connected to input devices such as a keyboard and mouse and outputs signals input from the keyboard and mouse to the control unit 210.

The image processing unit 240 is provided with a plurality of image output ports so that it can be connected to at least one display device 300, and processes input images into a preset image format to control the control unit 210 and/or the display device 300. Output as

The communication unit 260 includes a video communication unit 261 and a data communication unit 262, which are analog communication units.

The video communication unit 261 has a plurality of composite input ports and is connected to an analog camera through the composite input ports to receive analog video signals and provide them to the video processing unit 240.

The data communication unit 262 includes a serial communication unit 263 that performs serial data communication, a short-range communication unit 264 that connects to a local area network (LAN) and performs data communication with other devices connected to the local area network, and It includes a wireless communication unit 265 that wirelessly connects to a wireless network such as a WiFi network or a personal area network (PAN) and performs data communication with other devices connected to the wireless network.

The serial communication unit 263 may be a communication device that performs serial communication according to the RS485 protocol, and is connected to an RFID reader 360 that supports serial communication to read vehicle information from an RFID tag attached to a vehicle entering or leaving the parking lot. performs data communication to receive, is connected to the access control devices (320, 330), receives authentication information of the entrant, and performs data communication such as transmission of locking and opening control data to the door lock devices (322, 332).

The short-range communication unit 254 and the wireless communication unit 265 are connected to the outer camera 310 and the inner camera 311 that support IP communication through a short-range communication network and transmit pan/tilt/zoom control data of the camera. Receives video data captured from sources. In addition, the short-range communication unit 254 is connected to access control devices 320 and 330 that support IP communication through a short-range communication network and transmits and receives access locking and unlocking control data and visitor authentication information.

In addition, the short-range communication unit 254 and the wireless communication unit 265 perform data communication with the exterior intrusion detection sensor 340 and the fire detection sensor 350 that support IP communication, such as exterior intrusion detection data and fire detection data. will receive.

If the wireless communication unit 254 is a Zigbee wireless communication device, devices such as cameras (310, 311), access control devices (220, 230), perimeter surveillance sensors (240), and fire detection sensors (350) that support Zigbee wireless communication. It can be configured to establish a PAN with other devices and perform data communication with other devices through the PAN to transmit and receive video data, camera control, door control, and visitor authentication information.

Hereinafter, the configuration of the control unit 210 will be described in detail.

The control unit 210 includes an access control module 211, an image storage module 212, a perimeter surveillance module 213, a fire monitoring module 214, a parking control module 215, a movement tracking module 216, and a record inquiry module. (217) and an image object tracking module (218).

The access control module 211 is connected to the access control devices 320 and 330 by at least one of the serial communication unit 263, the short-distance communication unit 264 and the wireless communication unit 265 to control the access control devices 320 and 330. If the visitor authentication information is received from the recognition devices 321 and 331 and the visitor is authorized by referring to the visitor authentication information of the access authorized person pre-registered in the visitor authentication information DB 222, the access recognition device 321 is sent through the corresponding communication unit. The door is opened by outputting a door open control signal to the door lock devices 322 and 332, which are paired with , 331). In addition, the access control module 211 uniquely identifies the access recognition device 321, 331 where the access event occurred when an access attempt (or “access event”) occurs through each access control device 320, 330. Information, location information, access time, etc. are stored and managed in the access record DB (224). In particular, the access control module 221 is an accessor registered in the accessor authentication information DB 222, but an unauthorized person attempts to enter the doorway attempted to enter more than a certain number of times, or an unauthorized person not registered in the accessor authentication information DB 222 When an entry attempt is made, the unauthorized person's entry event is stored in the access record DB 224, and then the movement path tracking module 216 is requested to track the unauthorized person's movement path.

And the access control module 211 receives image data captured from the access recognition device 320 equipped with a camera, detects the person's face and extracts facial features from the image data, and creates the visitor authentication information DB 222. It may further include a face recognition means (not shown) that performs face recognition by comparing the facial feature information of the authorized person pre-registered in . At this time, the captured image data will be provided to the image storage module 212.

In addition, the access control module 211 releases the access control of the access recognition devices 321 and 331 installed at the entrance classified as low level when a fire detection signal is input from the fire monitoring module 215, so that people inside can quickly escape in the event of a fire. Allow for evacuation.

If there is an analog camera among the internal cameras 311, the image storage module 212 is connected to the internal camera 311, which is an analog camera, through the video communication unit 261 of the communication unit 260, and receives an analog video signal to image the image. After performing digital image processing through the processing unit 240, the image is stored in the image DB 221 and provided to the display device 300 and the administrator terminal 370.

In addition, the video storage module 212 receives camera information including video data captured by the outer camera 310 and identification information of the shooting camera through the outer surveillance module 213 and stores it in the video DB 221, and stores it in the video DB 221 at a short distance. It is connected to digital cameras and internal cameras 311 capable of TCP/IP communication through the communication unit 264 and wireless communication unit 265, and performs data communication by the TCP/IP communication protocol, thereby performing fan/interval of the internal camera 311. Tilt/zoom is controlled, and the image captured through the internal camera 311 is processed in a predetermined format through the image processing unit 240, and is stored in the image DB 221 along with camera information including identification information of the camera. It is stored or provided to the display device 300 or the administrator terminal 370. However, video data captured by the outer camera 310 may be configured to be stored in the video DB 221 by the outer surveillance module 213.

In addition, the video storage module 212 stores video data selected by the manager through the interface unit 230 and the manager through the manager terminal 370, or captured through the camera of the access recognition device 321 that has generated an access request. It is input from the recognition device 320 and stored in the image DB 221 along with access recognition device information including unique identification information of the access recognition device 320.

In addition, the image storage module 212 is connected to the image processing unit 240 when a plurality of display devices 300 are connected to the image processing unit 240 or when a request to display images captured by multiple cameras occurs on one display device. ) is controlled to display two or more images simultaneously on one or more display devices 300.

In addition, the image storage module 212 receives an image request signal containing the location information of the two access recognition devices 320 and 330 from the movement tracking module 216, and stores the information between the two access recognition devices 320 and 330. Images captured through cameras installed on the route are searched from the image DB 221 and provided to the movement tracking module 216.

The image storage module 212 also has a face recognition means that extracts a person's face from images captured through the external camera 310, the internal camera 311, and the camera of the entry/exit recognition device 320 and performs face recognition. It may be configured to include (not shown).

The perimeter surveillance module 213 is connected to the exterior camera 310 and the exterior intrusion detection sensor 340 through the video communication unit 261 and the data communication unit 262, and is monitored from the outside by the exterior intrusion detection sensor 340. It monitors whether there is an intruder invading the area, receives analog video signals or digital video data from each camera, and stores them in the video DB 221 through the video storage module 212 or directly.

The fire monitoring module 214 is connected to the fire detection sensors 350 through the data communication unit 262 to monitor whether a fire occurs in the surveillance area, and generates an alarm when a fire detection signal is received from the fire detection sensor 350. The location of the fire detection sensor that detected the fire is determined by referring to the detector information DB (223), and the determined location information is transmitted to the image storage module and the perimeter surveillance module (213) to capture the information captured through a camera near the location. The video is acquired and stored in the video DB (221), and a fire detection signal is provided to the access control module (211).

When detecting a flame, the fire monitoring module 214 must distinguish between the background and the object flame. For background detection, a background modeling method is used, and for flame detection, flame candidate areas are designated using various color space model values, and object state information output from the image object detection unit 218, which will be described later, or The presence of a flame is determined using a classifier, which is learning information generated in the learning unit 451.

Here, F(x, y) is defined as the flame pixel,

τ is a constant and is determined by analysis of the receiver operating characteristics.

The parking control module 215 is connected to the RFID readers 360 installed at the entrance of the parking lot, each floor of the parking lot, and the parking lot through the data communication unit 260 to store vehicle information. Vehicles read from vehicles with attached RFID tags It receives information and controls the entry and parking of vehicles entering and leaving the parking lot.

The movement tracking module 213 determines that the person whose entry and exit authentication information is recognized through the access recognition devices 321 and 331 is an unauthorized person a certain number of times (for example, 3 times) or is an intruder. If determined, the access recognition device information of the two access recognition devices 321 and 331 through which the unauthorized person and the intruder attempted to enter is provided to the image storage module 212 so that the unauthorized person and the intruder can access the two access recognition devices 321 and 331. ) is received and stored in the movement line tracking DB (225), and the location information of the access recognition devices (321, 331) used by unauthorized persons and intruders is sequentially stored to track the movement paths of unauthorized persons and intruders. It is then stored in the movement tracking DB 225 along with the video and provided to the display device 300 and the administrator terminal 370.

The record inquiry module 217 provides a record inquiry means to the administrator connected through the interface unit 230 and the administrator terminal 370, and each information stored in the storage unit 220 according to the search information input through the record inquiry means. The relevant information is searched from the DB and displayed on the display device 300 or provided to the administrator terminal 370.

The video object tracking module 218 acquires images directly from the access control devices 320, 330, the outer camera 310, and the inner camera 311 or through the image storage module 212, and detects objects in the captured images. It determines whether an object exists or not, and if there is an object, detects and tracks the object. The video object tracking module 218 provides learning information, such as classifier parameters or object status information, to the fire monitoring module 214 to detect fire more accurately.

In addition, the image object tracking module 218 detects the background and the object separately in the captured image, detects the bending that may occur depending on the viewing angle, that is, the viewing angle, and performs generalization in the preset unique perspective. Correct the curvature by comparing it with the object data. The object may be a person or a flame, and the generalized object data may be data obtained by capturing a generalized person and a flame from the unique viewpoint.

The detailed configuration and operation of the image object tracking module 218 will be described with reference to FIG. 5, which will be described later.

Figure 5 is a diagram showing the configuration of the video object tracking module of the access control video storage control device according to the present invention, and Figure 10 is a diagram illustrating the gender and age analysis method based on a person's gait according to the present invention. 11 is a diagram showing the characteristics of a scream and a general voice according to the present invention. Hereinafter, it will be described with reference to FIGS. 5 and 10 to 11.

Referring to FIG. 5, the image object tracking module 218 includes an image acquisition unit 410, an audio acquisition unit 420, an image correction unit 430, an object tracking unit 440, and a situation determination unit 480. .

The image acquisition unit 410 acquires images directly from the access control device 330, the exterior camera 310, and the internal camera 311 through at least one of the image communication unit 261 and the short-range communication unit 262, or stores the images. Images captured by the camera devices (access control device 330, exterior camera 310, and interior camera 331) are acquired and output through the module 212.

The voice acquisition unit 420 transmits voice collected through a microphone included in the access control device 330 or a microphone installed in pairs such as the outer camera 310 and the internal camera to the short-range communication unit 262 or a separate voice processing unit. (not shown), determines whether the acquired voice is a scream or a normal voice, and outputs the result. Whether the voice is a scream or a general voice can be determined by the waveform and spectrum specific information of the scream and general voice as shown in the waveform diagram (a) and spectrum diagram (b) of the scream and general voice in FIG. 11. will be.

The image correction unit 430 detects the background and the object separately from the image input from the image acquisition unit 410, detects the viewing angle of the detected object, that is, the bending that may occur depending on the angle, and sets the preset unique visual angle. The curvature is corrected and output by comparing it with the generalized object data in .

The object tracking unit 440 analyzes the image input from the image acquisition unit 410 or the image correction unit 430 to detect and track an object in the image. If the tracked object is a person, the body type according to gender and age The person's gender and age is detected by referring to the walking information DB (not shown) that stores information and walking information, and the result is output by comparing the face information with the visitor's face information in the visitor authentication information DB (222). do. To be more specific, the object tracking unit 440 includes an object detection unit 450, a gender detection unit 460, and a face detection unit 470.

The object detection unit 450 detects and tracks an object in an input image and outputs an object image and object status information for the detected object. As described above, the object detection unit 450 provides the object status information to the fire monitoring module 214 so that the fire monitoring module 214 can more easily determine whether the detected object is a flame or not.

The gender detection unit 460 determines whether the detected object is a person from the object image, and if it is a person, estimates the gender and age of the object based on the walking image of the object as shown in FIG. 10 and outputs the object gender and age information. do.

That is, the gender detection unit 460 substitutes preset physical factors to classify the appropriate gender and age group of the object. The physical elements may be hair, hairstyle, chest, waist, back, buttocks, legs, etc., as shown in (a) of FIG. 10. And, as shown in (b) of Figure 10, a gait information DB is created by configuring the person's gait pattern information according to gender and age, and based on this,

If the object output from the object detection unit 450 is a person, the face detection unit 470 detects the face, recognizes the detected face, and determines whether the person is an authorized person or an unauthorized person registered in the visitor authentication information DB 222. Print the results.

The situation determination unit 480 receives the voice judgment result information output from the voice acquisition unit 420, receives information on the detected object from the object tracking unit 440, and receives the location where the image was transmitted (access control device (access control device) The current situation regarding the installation position of the camera 330, the installation position of the outer camera 310, or the installation position of the inner camera 311) is determined and the resulting situation determination information is output. If the object in the video is a person, the situation determination information includes information on whether the object is an authorized person or an unauthorized person, and includes information on whether an unexpected situation has occurred (due to a scream, etc.) due to the object. The situation determination information output from the situation determination unit 480 may be provided and used by the access control module 211 and the movement tracking module 217.

Figure 6 is a diagram showing the configuration of the object detection unit of the video object tracking module according to the present invention.

Referring to FIG. 6, the object detection unit 450 includes a learning unit 451, a detection unit 452, a tracking unit 453, and a data integration unit 454.

The learning unit 451 analyzes the input image using the learned learning information and collects data about the object and its surroundings, that is, the background (referred to as “video analysis data”), to the detection unit 452 and the tracking unit 453. It includes an object modeling unit 455 that outputs , and updates learning information by performing PN learning on the object and background detected by the detection unit 452.

The detection unit 452 scans the image analysis data for the input image from the object modeling unit 455 of the learning unit 451 in units of scanning windows, determines the presence or absence of an object in the image, and provides the result. Print out.

The tracking unit 453 represents the object by a bounding box and estimates the movement of the object between consecutive frames based on the extended median flow tracking method. That is, the tracking unit 453 estimates the movement of multiple points within the bounding box of the object and selects the object with the most reliable movement within the bounding box using the average. Here, the bounding box refers to a box containing an object, meaning an object.

The data integration unit 454 combines the bounding boxes output from the detection unit 452 and the tracking unit 453 and outputs them as object state information. If there is no bounding box in one of the detection unit 452 and the tracking unit 453, the data integration unit 454 determines that there is no object.

Figure 7 is a diagram showing the configuration of the learning unit of the object detection unit according to the present invention.

Referring to FIG. 7, the learning unit 451 includes an analysis setting unit 4511, a training unit 4512, a classification unit 4513, and an expert execution unit 4514.

The classification unit 4513 classifies data not confirmed by the tracking unit 453, that is, data for which object tracking is not possible, into object and background, and then classifies the previous frame of the object input from the training unit 4512 (Classifier). ).

The classification information is a function of the family parameterized by θ. The family can be fixed and implemented by training, and the training is the estimated response of the parameter θ. Since θ is a well-known technology in this technical field, its description is omitted.

The expert execution unit 4514 receives classification information (or “classifier”) for unconfirmed data from the unconfirmed data and classification unit 4513, and executes the P-expert and N-expert to classify the object. and generate expert analysis information on the background.

Specifically, the P-expert utilizes the temporal structure of the video, assumes that the object moves along an inter-frame trajectory, remembers the position of the object in the previous frame, and uses the inter-frame tracking unit 453 to track the current frame. Estimate the location of the object in .

The N-expert utilizes the spatial structure of the video, assumes that an object can be displayed at a single location, and analyzes all responses generated by the detection unit 452 and all responses from the tracking unit 453 for the current frame. Choose the most obvious one.

The analysis setting unit 4511 receives confirmed data, that is, object tracking data, from the detection unit 452, trains the input confirmed data, stores the training examples, and expert execution unit 4514 ) receives the information detected in and updates the training set.

The learning tracking unit 4512 receives the training set output from the analysis setting unit 4511, trains it, and generates learning information (classifier parameters).

Figure 8 is a flowchart showing a control method of an access control video storage device according to the present invention.

Referring to FIG. 8, first, the control unit 210 controls the image storage mode through the image storage module 212, the fire monitoring mode through the fire monitoring module 214, the perimeter surveillance mode and parking mode through the perimeter surveillance module 213. Operates in parking control mode through the control module 215 (S511)

During operation in the image storage mode, the access control module 211 of the control unit 210 checks whether an access request event occurs from the access control devices 320 and 330 (S513). The access request event includes visitor authentication information. It may occur when an access request signal is received. It would be desirable for the access request signal to include device-specific identification information as well as the user authentication information. If the access control devices 320 and 330 are IP devices, the identification information may be a MAC address or an IP (Internet Protocol) address.

When the access request event occurs, the access control module 211 of the control unit 210 determines whether a fire has been detected through the fire monitoring module 214 (S514). If a fire has occurred as a result of the determination, the fire monitoring module 214 After generating an alarm, a fire communication signal containing the location information of the fire detection sensor 350 that detected the fire will be transmitted to the access control module 211 to notify that a fire has occurred. The access control module 211, which has received a fire detection notification, releases the access control mode of the access recognition devices 321 and 331 at the entrance classified as low level (S516). At this time, the access control module 211 is included in the fire notification signal. Based on the location information, only the access recognition devices 321 and 331 installed at the low-level entrance outside the location of the location information may be configured to release the access control mode.

On the other hand, for access recognition devices (321, 331) where no fire has been detected or installed at the low-level entrance, the access control module 211 receives access control device information and Enter the visitor authentication information and determine the location of the access request by referring to the detector information DB (223) (S519)

When the access request location is determined, the access control module 211 determines whether the access recognition device at that location is the access recognition device 321 equipped with a camera (S520)

If the access recognition device 321 is equipped with a camera, the access control module 211 drives the camera of the access recognition device 321 (S521) and starts filming the entering person (S523). At this time, the captured video signal is an image. It may be stored in the video DB 221 through the storage module 212, or directly through the access control module 211.

When the shooting starts, the access control module 211 analyzes the captured image directly or through the image storage module 212, receives the result, and determines whether the face of the visitor is being photographed (S525)

As a result of the determination, if the face of the person entering the room is not being filmed, it is determined whether there is another camera nearby by referring to the detector information DB 223 (S527), and if there is another camera nearby, the corresponding camera information is provided to the video storage module 212 to store the video. The corresponding image is acquired through the module 212 and stored in the access record DB 224 as an access event for the access recognition device 321 (S529).

On the other hand, if there is no adjacent camera or after S529, the access control module 211 determines whether the visitor whose authentication information is input from the access recognition devices 321 and 331 is an authorized person who can enter through the corresponding entrance (S531)

As a result of the determination, if it is an authorized person, the access control module 211 controls the door lock devices 322 and 332 corresponding to the access recognition devices 321 and 331 to open the door (S533).

After opening the door, the access control module 211 stores a video entry event including video data input and stored from the video storage module 212 or a camera in the access record DB 224 (S534) and ends video storage. Do(S535)

On the other hand, if the visitor requesting access to the entrance is an unauthorized person or an intruder, the access control module 211 provides the visitor information to the movement trace module 216 and at the same time requests the entrance tracer to track the entrance. Then, the moving line tracking module 216 performs a moving line tracking process (S537). The moving line tracking method will be described in detail with reference to FIG. 6 below.

Figure 9 is a flowchart showing a method of tracking an intruder's movements in an access control video storage device according to the present invention.

When a request for movement line tracking is generated from the access control module 211, the movement line tracking module 216 determines that the person entering the visitor authentication information input from the access control module 211 is an unauthorized person for the entrance, but the person entering the person authentication information DB 222 is Determine whether you are a registered visitor (S611)

As a result of the determination, if the visitor is registered in the visitor authentication information DB 222, the movement line tracking module 216 checks whether the number of entry attempts of the visitor exceeds a certain standard number (S613). If the number of entry attempts exceeds the standard number If so, you are considered an intruder.

On the other hand, if the visitor is not registered in the access authentication information DB 222, he or she is judged to be an intruder.

If it is determined that the intruder is an intruder, the movement tracking module 216 sets the intruder as a movement tracking target and notifies the administrator through the administrator terminal 370, etc. (S615), and secures the entry and exit information of the unauthorized intruder (S617). If the intruder is a registered visitor in the visitor authentication information DB 222, the visitor information may be the information of the visitor stored in the visitor authentication information DB 222, and may be an visitor not registered in the visitor authentication information DB 222. If the rear access recognition devices 321 and 331 are biometric recognition devices, it can be biometric information, and facial images obtained through the cameras of the outer camera 310, the internal camera 311, and the access recognition device 321, etc. It could be.

When the information on the intruder is secured, the movement tracking module 216 starts monitoring the movement of the intruder based on the secured information on the intruder (S619).

When monitoring is initiated, the movement tracking module 216 checks whether unauthorized persons are detected through the access control module 211, the perimeter surveillance module 213, and the parking control module 215 (S621). When an unauthorized person is detected, the movement tracking module 216 determines the detection location (S623) and checks whether there is a camera near the identified location by referring to the detector information DB 223 (S625).

As a result of the inspection, if there is a camera in the vicinity, the movement tracking module 216 acquires and stores the image captured by the camera through the image storage module 212 and the outer surveillance module 213 (S627). Here, the nearby camera Although the image was acquired only through the movement tracking module 216, the location information of the entry recognition device of the two consecutive entrances through which the intruder attempted to enter and the entry and exit time information of each of the two entrances were provided to the image storage module 212 to determine the location. Images corresponding to the times and times between the two times may be acquired and stored in the movement tracking DB 225.

Although embodiments according to the present invention have been described above, they are merely illustrative, and those skilled in the art will understand that various modifications and equivalent scope of embodiments are possible therefrom. Accordingly, the scope of protection of the present invention should be determined not only by the following claims but also by their equivalents.

Claims (3)

A camera installed in the surveillance area to photograph the surveillance area and output the captured video;
An access control device including an access recognition device that obtains and outputs access authentication information from the visitor, a door lock device that locks or unlocks the door, and opens the door when a door open control signal is input; and
It includes an access control video storage control device that stores video captured through the camera and controls entry and exit of visitors through the access control device,
The access control video storage control device,
An image DB that stores video data captured by the camera by mapping them with the camera information, an accessor authentication information DB that stores visitor authentication information about the person authorized to enter each security target entrance of the surveillance area, and gender and a storage unit containing a gait information DB including gait information according to age;
A communication unit connected to the access control device to receive and output visitor authentication information from the access recognition device, transmit an input control signal to the door lock device, and receive and output an image from the camera; and
an image storage module that controls the cameras through the communication unit, receives images captured by the cameras, and stores them in the image DB;
Obtain images directly from the camera and access control device or through an image storage module, determine whether there is an object in each frame of the captured image, and if there is an object, use the location and median of the object to determine the most reliable A video object tracking module that detects a moving object, performs face recognition if the detected object is a person, performs authentication by comparing it with a face pre-registered in the visitor authentication information DB, and outputs the result;
Through the communication unit, the visitor authentication information of any visitor is input from the access recognition device for which an access request has been generated, and the visitor authentication information of those authorized for access to the entrance where the access recognition device is installed, which is pre-registered in the visitor authentication information DB, is compared. If the user is an authorized visitor and the authentication result information by facial recognition from the video object tracking module is successful, a control unit including an access control module that outputs a door open control signal to the door lock device through the communication unit to open the door. An access image storage device capable of analyzing the entry and exit movement lines of unauthorized visitors. According to paragraph 1,
The video object tracking module,
an image acquisition unit that acquires an image directly from the camera or acquires an image through the image storage module;
An object detection unit that determines whether there is an object in each frame of the captured image and, if there is an object, detects and tracks the object with the most reliable movement using the location and median of the object;
an object tracking unit including a face detection unit that detects and outputs a face if the detected object is a person; and
Characterized by comprising a situation determination unit that performs facial recognition on the detected face, performs authentication by comparing it with a face pre-registered in the visitor authentication information DB, and outputs the result to the access control module. An entry/exit video storage device that can analyze the entry/exit movement of visitors. According to paragraph 1,
The object detection unit,
A learning unit that analyzes the input image using learned learning information and outputs image analysis data for the object and background;
a detection unit that scans the image analysis data in units of scanning windows to determine whether an object exists in the image and outputs the result and a bounding box containing the object;
a tracking unit that represents an object by a bounding box and estimates the movement of the object between consecutive frames based on an extended median flow tracking method; and
An access image storage device capable of analyzing the entrance and exit movement lines of unauthorized persons, comprising a data integration unit that combines and outputs bounding boxes output from the detection unit and the tracking unit.