WO2012137994A1

WO2012137994A1 - Image recognition device and image-monitoring method therefor

Info

Publication number: WO2012137994A1
Application number: PCT/KR2011/002341
Authority: WO
Inventors: 정의석; 김영석
Original assignee: (주)아이티엑스시큐리티
Priority date: 2011-04-04
Filing date: 2011-04-05
Publication date: 2012-10-11
Also published as: KR20120113014A; KR101212082B1

Abstract

The present invention relates to an image recognition device and to an image-monitoring method therefor. The image recognition device of the present invention performs an image-monitoring operation using a plurality of sensors and a plurality of cameras after dividing a monitoring area into a plurality of logical monitoring areas, and provides a monitoring result to an external system while tracing and monitoring the movement of an object recognized in a monitoring target area through an image interpretation of an image signal.

Description

Image recognition device and video surveillance method

The present invention relates to an image recognition device for dividing a surveillance zone into a plurality of logical unit surveillance zones to perform video surveillance with a plurality of sensors and a plurality of cameras, and a video surveillance method thereof.

An image recognition device is a device that recognizes and analyzes objects in an image by using various image processing techniques beyond simply storing an image signal input through a camera.

However, video surveillance apparatus using a conventional camera includes a CCTV system using a video cassette recorder (VCR) and a digital video recorder (DVR) for storing images on a hard disk. Such a device was merely a device for storing an image.

Researches on analyzing images and extracting and recognizing regions of interest or objects within images have been around for a long time. Applicant has already applied for a patent application 2010-0040469 on a method for monitoring object tracking based on image recognition.

In the present invention, the applicant has proposed an algorithm for tracking the movement of the object in a logical way beyond simply recognizing the object captured in the camera image of the surveillance zone.

SUMMARY OF THE INVENTION An object of the present invention is to provide a video recognition apparatus for performing video surveillance with a plurality of sensors and a plurality of cameras by dividing a surveillance zone into a plurality of logical unit surveillance zones and a video surveillance method thereof.

In order to achieve the above object, the image recognition device according to the present invention includes a plurality of cameras respectively photographing a plurality of unit surveillance zones divided into an entire surveillance zone, a plurality of sensors installed on an outer surface of the entire surveillance zone, and It includes an image controller that receives images from the plurality of cameras and is connected to an external system.

The image controller includes an image receiver, an event generator, an image processor, a monitor, and an external device linking unit.

An image monitoring method using an image controller includes: establishing a logical connection relationship between the plurality of cameras based on a moving path of a moving object according to spatial connection of the plurality of unit monitoring zones; Sensing one of the plurality of sensors installed outside the entire monitoring zone and generating a monitoring event; Recognizing and tracking a moving object in an image input from the primary tracking camera by using a camera logically connected to a sensor that detects the moving object among the plurality of cameras as a primary tracking camera according to the monitoring event; And providing an image to the external system that recognizes the moving object at the time when the moving object is recognized.

According to an embodiment of the present disclosure, if the object is not recognized during the tracking, the tracking is performed through an object recognition process for an image input from a secondary tracking camera that is logically connected to the primary tracking camera among the plurality of cameras. Tracking the object again; And providing the image of the secondary tracking camera to the external system when the object is re-recognized by the secondary tracking camera.

According to another embodiment of the present invention, when the object is not recognized during the tracking, when there are a plurality of cameras logically connected to the primary tracking camera among the plurality of cameras, the video surveillance method of the present invention is logically connected. The second tracking camera is selected by recognizing a moving object with respect to an image input from a plurality of cameras.

The image recognition apparatus of the present invention may provide a monitoring result to an external system while monitoring a video by dividing the monitoring zone into a plurality of unit monitoring zones using a plurality of sensors and cameras.

Since the camera and the sensor are logically connected to each other, the image recognition device may have the same effect as if the manager selects and records the monitored camera according to the moving path of the object once the monitored object is captured.

In addition, since such image analysis and tracking do not need to be performed on the entire video signal input from the entire installed camera, there is no problem that the overall system speed decreases according to the image processing.

1 is a block diagram of an image recognition device according to an embodiment of the present invention;

2 is a diagram illustrating an example in which a sensor and a camera are installed in a surveillance area;

3 is a flowchart provided to explain an image monitoring method according to an embodiment of the present invention; and

4 shows an example of an image provided as a result of monitoring.

Hereinafter, the present invention will be described in more detail with reference to the accompanying drawings.

Referring to FIG. 1, the image recognition device 100 of the present invention is connected to an external system 30 through a network 10.

The external system 30 is a device capable of receiving a video surveillance result provided in the form of an image from the image recognition device 100 and displaying the result to the user. The external system 30 may be a mobile phone, a smart phone, or a separately designed security system.

The network 10 may be configured in various ways according to the type of the external system 30. For example, when the external system 30 is a portable terminal or a security system, the network 10 may correspond to a mobile communication network, a wired or wireless Internet, and the like.

The image recognition apparatus 100 may provide the external system 30 with an image surveillance result through image processing on an analog or digital image signal provided by the plurality of cameras 101a to 101e installed in the surveillance space.

The image recognition device 100 includes a plurality of cameras 101a to 101e installed in a surveillance space, a plurality of sensors 103a to 103c, and an image controller 110 for processing images captured by the cameras 101a to 101e. It includes.

The cameras 101a to 101e are not only analog cameras for generating analog video signals having a predetermined frame rate, digital cameras for generating digital video signals, but also IP (Internet Protocol) cameras that can be connected to the Internet.

2, the cameras 101a to 101e are installed to divide the surveillance zone to photograph each divided space, and each camera 101a to 101e is mutually logical with at least one other camera as described below. Is connected.

Sensors 103a to 103c are installed along the outer periphery of the surveillance zone as shown in FIG. 2 to detect an intruder invading into the surveillance zone, and correspond to an infrared detector and a magnet detector. The sensors 103a through 103c are also logically interconnected with at least one of the cameras 101a through 101e.

The image controller 110 stores and manages images captured by the cameras 101a to 101e. If necessary, the image controller 110 performs image processing on the images of the cameras 101a to 101e and provides the information to the external system 30 when it corresponds to a predetermined condition. To this end, the image controller 110 includes an image receiver 111, a storage medium 113, a device interface 115, and a controller 130.

The image receiving unit 111 receives and decodes the analog image signals provided by the cameras 101a to 101e and then displays them through a display unit (not shown), or converts them into digital image signals and stores them in the storage medium 113. Provided at 130.

The storage medium 113 stores the digital video signal provided by the image receiver 111.

The device interface unit 115 is a wired or wireless interface for connecting to the network 10. The device interface unit 115 is connected to the external system 30 through the network 10, and transmits various monitoring results provided by the controller 130 to the external system 30.

The controller 130 controls the overall operation of the image controller 110. However, although the image controller 110 of the present invention may perform various functions, for example, storing / searching each camera image, such a function is not a characteristic function of the present invention, and thus the description thereof is omitted and the present invention is presented. It will be described based on the interlocking function with the external system 30.

The controller 130 is an event generating unit 131, an image processing unit 133, a monitoring unit 135, a recording and playback processing unit 137 and an external device interlocking unit 139 for the characteristic intelligent image tracking function of the present invention. It includes.

The event generator 131 reads the sensing result of detecting the intruder from the sensors 103a to 103c to generate a monitoring event, and provides the generated monitoring event to the monitoring unit 135.

The image processor 133 recognizes a moving object by performing object recognition on the image input from the camera that the monitoring unit 135 processes from among the cameras in operation. When the moving object is detected, the image processor 133 provides the monitoring unit 135 with detection information of the corresponding object. Here, the detection information includes at least information on a time when a new object appeared in the image of the requested camera and a time when the new object disappeared.

Object recognition and tracking of the image processor 133 may use various known methods. For example, general object recognition may be obtained from a subtraction operation of an image of a previous frame and an image of a new frame, and tracking of the detected object may also use a known algorithm.

The monitoring unit 135 performs a monitoring algorithm for a monitoring target area with information on a logical connection relationship between each sensor and a camera (or a logical connection relationship between each monitoring area) to be described below.

The monitoring unit 135 requests the image processing unit 133 to recognize the intruder (object) through the analysis of the image input from the specific camera based on the monitoring algorithm, and based on the recognition result provided from the image processing unit 133. Based on the monitoring results are generated in a log file and stored in the storage medium (113).

The recording / playback processing unit 137 controls an operation for recording and reproducing a basic video signal provided by the image recognition device 100, and requests an image input from a specific camera according to the request of the monitoring unit 135. From this time, it may be managed to be stored as a separate image file or a separate tag or metadata may be assigned.

The external device interlocking unit 139 provides the monitoring result of the monitoring unit 135 to the external system 30 through the device interface unit 115. In this case, the monitoring result provided by the external device interworking unit 139 may include a camera still image at the moment of recognition of the object, corresponding camera position information, and time information of detecting the object in the image.

According to an embodiment, when the external device interlocking unit 139 provides the monitoring result to a portable terminal lacking system resources, the monitoring result is recorded by recording the camera position information and the detection time information as part of the image in the camera still image. It may be provided in the form of an image.

Hereinafter, prior to explaining the operation of the image recognition device 100 of the present invention, a description will be given of the configuration of the security surveillance camera according to the path analysis of the moving object according to the spatial connection of the surveillance zone, its logical connection and monitoring algorithm.

2 shows the entire surveillance zone to be monitored, the first unit surveillance zone L1 by the first camera 101a, the second unit surveillance zone L2 by the second camera 101b, and the third. The third unit monitoring zone L3 by the camera 101c, the fourth unit monitoring zone L4 by the fourth camera 101d, and the fifth unit monitoring zone L5 by the fifth camera 101e Is shown. In other words, each unit monitoring zone L1 to L5 is divided by cameras 101a to 101e.

In addition, the first to third sensors 103a to 103c are disposed along the outer periphery of the surveillance zone, and the first to third sensors 103a to 103c are also at least one of the first to fifth cameras 101a to 101e. Logically connected to

2 shows that the first sensor 103a is a magnet detector installed in the window of the first unit monitoring zone L1, the second sensor 103b is a magnet detector installed in the window of the second unit monitoring zone L2, The third sensor 103c is an example of an infrared detector installed in a window of the third unit monitoring zone L3.

Each surveillance zone, each camera or sensor is in a logical connection with each other by moving path analysis based on the purpose of the surveillance, the spatial characteristics of the surveillance zone and the characteristics of the object to be monitored (eg, intruder). For example, when the object to be monitored is a human, an intruder who invades the first unit monitoring area L1 must pass through the third unit monitoring area L3, and the intruder who appears in the first unit monitoring area L1 has a fourth moment. It does not occur in the unit monitoring area L4. Therefore, the first unit monitoring zone L1 is logically connected to the second unit monitoring zone L2 and the third unit monitoring zone L3.

In addition, according to the same movement path analysis, the first sensor 103a is logically connected to the first camera 101a, and the second sensor 103b is logically connected to the second camera 101b. In addition, the third sensor 103c is logically connected to the third camera 101c and the fourth camera 101d.

Accordingly, the logical connection relations of the cameras 101a to 101e and the sensors 103a to 103c of FIG. 2, that is, the logical connection relations of the unit monitoring zones, are shown in Table 1 below.

Table 1

Configuration	Logical connection camera
First camera	2nd, 3 camera
Second camera	1st, 3 camera
Third camera	1st, 2nd, 5th camera
4th camera	5th camera
5th camera	3rd, 4th camera
First sensor	First camera
Second sensor	Second camera
Third sensor	3rd, 4th camera

According to Table 1, the fifth unit monitoring zone (L5) → the third unit monitoring zone (L3) → the first unit monitoring zone (L1) is a possible connection, but the fifth unit monitoring zone (L5) → the fourth unit monitoring zone (L5) (L4) → The first unit monitoring area L1 is impossible because there is no connection between the fourth unit monitoring area L4 and the first unit monitoring area L1.

Further, the third sensor 103c → fourth camera 101d → fifth camera 101e are possible connections, but the third sensor 103c → second camera 101b → third camera 101c is a third connection. Since there is no connection between the sensor 103c and the second camera 101b, this is impossible.

Hereinafter, referring to FIG. 3, an image monitoring method of the image recognition device 100 of the present invention centering on the monitoring unit 135 and the image processing unit 133 will be described.

The event generator 131 reads the first to third sensors 103a to 103c and generates a monitoring event when an intruder is detected (S301).

When the surveillance event is generated by the event generator 131, the monitor 135 requests the image processor 133 to recognize and track an object of an image of a camera logically connected to the surveillance event (S303).

Hereinafter, for convenience of description, it is assumed that the monitoring event has occurred by the first sensor 103a. After the event occurs, it is assumed that the object p enters the first unit monitoring area L1 and moves to the third unit monitoring area L3. Accordingly, the event generator 131 generates a surveillance event by the first sensor 103a, and the monitor 135 requests the image processor 133 to recognize and track the object of the image of the first camera 101a. Done.

At the request of the monitoring unit 135, the image processing unit 133 starts object recognition and tracking of an image input from a camera requested to be monitored (in the example of FIG. 2, the first camera). Object recognition, extraction, and tracking of the image processor 133 may be processed by a method known in the image processing field as described above.

When a moving object is extracted during image analysis, the image processor 133 provides object recognition information to the external device interlocking unit 139 and the monitoring unit 135 and starts tracking the extracted object. The object recognition information may include object recognition time and object recognized camera information, and if necessary, may include a corresponding still image.

The external device interlocking unit 139, which receives the object recognition information from the image processing unit 133, provides the image processed by the image processing unit 133 to the external system 30. At this time, the camera location information and time information of detecting the intruder are also provided. In the example of FIG. 2, since the intruder will be recognized in the image of the first camera 101a, the monitoring result is obtained from one image captured by the first camera 101a, the first camera position, and the image of the first camera 101a. Time information identifying the intruder is included.

4 shows an example of one image provided to the cellular phone which is one of the external systems 30 as a result of the monitoring. The position of the first camera and the recognition time are recorded in the image provided to the external system 30.

When the object disappears from the image of the camera requested to be monitored after a certain time due to the movement of the object and the escape of the unit monitoring zone, the image processor 133 ends the tracking of the object with the end time information to the monitoring unit 135. to provide.

In the example of FIG. 3, when the intruder enters the first unit monitoring area L1, the image processing unit 133 starts object recognition and tracking, and the intruder leaves the third unit monitoring area L3 which is a corridor. At this time, the object disappears from the image input from the first camera 101a and the image processor 133 ends the tracking of the object.

When the monitoring unit 135 is notified of the end of the object tracking for the requested camera image from the image processing unit 133, the camera to continue tracking among the cameras (or unit monitoring zones) logically connected to the camera being tracked (unit monitoring zones) Select. To this end, the monitoring unit 135 requests the image processing unit 133 to track all of the logically connected cameras for a predetermined time. In the example, since the logically connected cameras of the first camera 101a are the second camera 101b and the third camera 101c, the monitoring unit 135 is configured for the second camera 101b and the third camera 101c. The image processor 133 requests the monitoring of the predetermined time. Here, the predetermined time may be determined in consideration of the moving capability of the object, the spatial characteristics of the surveillance zone.

The image processor 133 monitors whether there is a newly displayed object with respect to the images of the cameras which are requested to be tracked for a set time at the request of the monitor 135, and provides the monitoring result to the monitor 135. As the object p moves from the first unit surveillance zone L1 to the third unit surveillance zone L3, the image processor 133 notifies the monitoring unit 135 that a new object has been recognized in the third surveillance zone. do.

In some cases, the second unit monitoring zone L2 or the third unit monitoring zone L3 may have a previously moving object. Even in this case, the image processor 133 extracts only the newly appearing object from the image of the unit monitoring zone and provides the result to the monitoring unit 135.

If there is only one unit monitoring zone logically connected to the tracking initiated unit monitoring zone disclosed in step S301, step S311 may not be performed.

The monitoring unit 135 determines whether there is a camera in which the object to be tracked is searched based on the result provided from the image processing unit 133 in step S311.

If there is a detected camera (the third camera in the example of FIG. 2), the monitoring unit 135 selects the unit monitoring zone (the third unit monitoring zone in the example of FIG. 2) to continue tracking, and the corresponding unit monitoring zone. While requesting the monitoring result notification of S307 to the external device interlocking unit 139, the image processing unit 133 requests the same object recognition and tracking as in step S309.

Since the object p has moved from the first unit monitoring zone L1 to the third unit monitoring zone L3, the external device interlocking unit 139 is an image of the third camera 101c from which the object p was first captured. To the external system 30 as a monitoring result.

In the example of FIG. 2, the first camera 101a becomes the primary tracking camera and the third camera 101c becomes the secondary tracking camera. By repeating this process, the external system 30 can check in real time the information on which path the intruder who invaded the surveillance area is invading.

Steps S307 to S313 are repeated until the connection monitoring zone to continue tracking is not selected because the moving object is not recognized.

Although the above has been illustrated and described with respect to preferred embodiments of the present invention, the present invention is not limited to the above-described specific embodiments, it is usually in the technical field to which the invention belongs without departing from the spirit of the invention claimed in the claims. Various modifications can be made by those skilled in the art, and these modifications should not be individually understood from the technical spirit or the prospect of the present invention.

Claims

In the video surveillance method of an image recognition device having a plurality of cameras for respectively photographing a plurality of unit surveillance zones divided the entire surveillance zone,

Establishing a logical connection relationship between the plurality of cameras based on a moving path of the moving object according to the spatial connection of the plurality of unit monitoring zones;

Sensing one of the plurality of sensors installed outside the entire monitoring zone and generating a monitoring event;

Recognizing and tracking a moving object in an image input from the primary tracking camera by using a camera logically connected to a sensor that detects the moving object among the plurality of cameras as a primary tracking camera according to the monitoring event; And

And providing an image to the external system that recognizes the moving object at the time when the moving object is recognized.
The method of claim 1,

If the object is not recognized during the tracking, the object being tracked is again tracked through an object recognition process for an image input from a secondary tracking camera that is logically connected to the primary tracking camera among the plurality of cameras. step; And

And providing the image of the secondary tracking camera to the external system when the object is recognized again by the secondary tracking camera.
The method of claim 2,

When the object is not recognized during the tracking, when there are a plurality of cameras logically connected to the primary tracking camera among the plurality of cameras, an object moving with respect to an image input from all of the logically connected cameras is selected. And recognizing and selecting the secondary tracking camera.
A plurality of cameras respectively photographing a plurality of unit surveillance zones in which the entire surveillance zone is divided;

A plurality of sensors installed on the outer surface of the entire monitoring area; And

An image controller configured to receive images from the plurality of cameras and to be connected to an external system;

The image controller,

An image receiver configured to receive images captured by the plurality of cameras;

An event generator for detecting a moving object invading the entire surveillance zone by using the sensor and generating a surveillance event;

An image processor configured to recognize and track a moving object by processing an image input from the image processor;

Set a logical connection relationship between the plurality of cameras based on a moving path of the moving object according to the spatial connection of the plurality of unit monitoring zones, and logically connected to a sensor that detects the moving object among the plurality of cameras according to the monitoring event. A monitoring unit requesting the image processing unit to track an image of the primary tracking camera using the camera as a primary tracking camera; And

And an external device interlocking unit for providing the external system with an image recognizing the moving object at the time when the moving object is recognized.
The method of claim 4, wherein

The monitoring unit, when the object is not recognized during the tracking, requests the image processing unit to continuously track the camera that is in a logical connection with the primary tracking camera,

And the external device interlocking unit provides an image of the secondary tracking camera to the external system when the object is re-recognized by the secondary tracking camera.