KR20050056092A - Intelligent surveillance and sentry robot system - Google Patents

Abstract

 The present invention relates to an intelligent boundary robot system for recognizing the shape and movement of an object, automatically tracking the movement of the object, and performing surveillance and boundary tasks capable of automatic firing of the object. Intelligent boundary robot system according to the present invention, at least one master camera; Active camera; A drive unit; And a control unit. The master camera is installed to be fixed to a frame, and recognizes a target from an input image. The active camera is installed to be able to rotate up, down, left and right with respect to the frame, and can track the target. The driving unit rotates the active camera vertically and horizontally with respect to the frame. The controller receives an image from the master camera and the active camera, recognizes the image, and controls the master camera, the active camera, and the driver.

Description

Intelligent surveillance and sentry robot system

The present invention relates to a boundary robot system, and more particularly, an intelligent boundary robot for recognizing the shape and movement of an object, automatically tracking the movement of the object, and performing surveillance and boundary tasks capable of automatic firing of the object. It's about the system.

Intelligent robot technology is a next-generation new technology that will lead industrial competitiveness and military science technology in the 21st century according to the trend of the development of artificial intelligence (AI) concept. In particular, the surveillance boundary system is a system in which various technologies such as ultra low light camera technology, image recognition technology, image processing and storage technology, voice recognition technology, servo technology, image tracking technology, and system control technology are applied in combination.

The intelligent surveillance boundary robot system is growing in demand in national infrastructure such as airports, ports, and nuclear power plants due to the rapid growth of the security industry. In particular, in the military field, the necessity is raised to improve the efficiency of the daily alert mission. In order to increase the safety of soldiers and the efficiency of combat, various unmanned equipments capable of performing 3D (Dangerous, Dirty, Dull) missions are being developed and deployed. In addition to replacement, it can greatly contribute to military combat power.

In particular, surveillance and alert robotic systems can play a key role in the advancement of military tactics, and can prevent human fatigue that may occur during the soldier's alert activity in terms of alert surveillance, and reduced concentration by simple repetition. Can be. In addition, by cultivating precise tracking and countermeasure ability in conjunction with firearms, the display will have a high-speed accurate hitting capability.

U.S. Patent 5,379,676, "Fire Control System," provides a fire control system for manual aiming rifles. In the US patent, the target is tracked by a video tracker and laser of an electro-optical device (EOD), and the distance and orientation of the target are calculated. The image of this target is sent to the user's video monitor, which allows the user to match the gun to the target and perform shooting through the video monitor.

However, such a fire control system tracks targets using electro-optical equipment, calculates distance and azimuth, and assists shooting by a user through a video monitor of the user. It is not doing.

In addition, the conventional surveillance boundary system is a system that monitors and alerts by using a thermal camera or a general surveillance camera, and is not a system capable of intelligently recognizing and automatically tracking an image, but a manual or novice level of automation. It was a system that introduced the concept.

In particular, when a thermal camera used in a conventional surveillance boundary system is used, an expensive and night-only camera has a problem that an unnatural image is input during the day. There is a problem that is difficult to utilize. In addition, there is a problem that the tracking performance of the object is lowered by not automatically tracking the movement of the object.

The present invention has been made to solve the above problems, an intelligent boundary robot system for recognizing the shape and movement of an object, automatically tracking the movement of the object, and performing the monitoring and boundary tasks capable of automatic firing of the object. The purpose is to provide.

Intelligent boundary robot system according to the present invention for achieving the above object, at least one master camera; Active camera; A drive unit; And a control unit.

The master camera is installed to be fixed to a frame, and recognizes a target from an input image. The active camera is installed to be able to rotate up, down, left and right with respect to the frame, and can track the target. The driving unit rotates the active camera vertically and horizontally with respect to the frame. The controller receives an image from the master camera and the active camera, recognizes the image, and controls the master camera, the active camera, and the driver.

In addition, it is preferable to further include a shooting device fixed to the active camera so that the direction in which the barrel is directed coincides with the central axis facing the target of the active camera.

In addition, it is preferable to further include a sound device including a sound input unit for receiving sound information from the outside, the control unit recognizes the target by processing the sound information input from the sound input unit. In this case, it is preferable that the sound device further includes a sound output unit for outputting a warning signal to the outside.

Preferably, the master camera has a larger viewing angle than the active camera.

The master camera preferably recognizes a target by detecting a change in an input image.

Preferably, the active camera is provided with a zoom function to enlarge an image of a target.

The master camera and the active camera are ultra low light cameras having an infrared cut filter for blocking the image input of the infrared region, and the master camera and the active camera turn on the infrared cut filter during the day and input a color image. It is desirable to turn off the infrared cut filter at night and receive a black and white image.

Intelligent boundary robot system according to another aspect of the present invention, at least one master camera; Active camera; And a control unit.

The master camera recognizes a target from an input image. The active camera can track the target. The controller receives an image from a master camera and an active camera, recognizes the image, and controls the master camera and the active camera.

The controller may include an image processor configured to recognize image targets by processing image signals input from a master camera and an active camera, and to recognize a target movement; And a tracking control unit configured to control the vertical movement of the active camera, the vertical rotation, the horizontal rotation, and the zoom so that the active camera receives the target motion information from the image processing unit.

The image processor receives an image signal from a master camera to recognize a target, and recognizes the shape of the target and the movement of the target by at least one of the master camera and the active camera.

The master camera classifies the input image into frame units, compares the image signals of two or more frames, and recognizes a target from the difference of the image signals.

The firing apparatus is further provided with a shooting device which is rotated together by the rotation of the active camera so that the direction in which the barrel is directed coincides with the target axis of the active camera. The control unit further includes a shooting control unit for outputting a shooting signal to the shooting device. It is preferable to provide.

A sound device further includes a sound input unit for receiving sound information from the outside and a sound output unit for outputting a warning signal to the outside, wherein the control unit recognizes a target by processing sound information input from the sound input unit and outputs the sound output unit. It is preferable to further include a sound control unit for outputting a warning signal.

A boundary system according to another aspect of the present invention includes at least one or more of the intelligent boundary robot system; And a situation controller configured to receive boundary information from the intelligent boundary robot system, recognize a situation, and output a control command to the intelligent boundary robot system.

Intelligent boundary robot system according to another aspect of the present invention, at least one master camera; Active camera; And a control unit.

The master camera recognizes a target from an input image. The active camera can track the target. The controller receives an image from a master camera and an active camera, recognizes the image, and controls the master camera and the active camera.

The controller may include an image processor configured to recognize image targets by processing image signals input from a master camera and an active camera, and to recognize a target movement; And a tracking control unit configured to control the vertical movement of the active camera, the vertical rotation, the horizontal rotation, and the zoom so that the active camera receives the target motion information from the image processing unit.

The image processor receives an image signal from a master camera to recognize a target, and recognizes the shape of the target and the movement of the target by at least one of the master camera and the active camera.

The master camera classifies the input image into frame units, compares the image signals of two or more frames, and recognizes a target from the difference of the image signals.

The firing apparatus is further provided with a shooting device which is rotated together by the rotation of the active camera so that the direction in which the barrel is directed coincides with the target axis of the active camera. The control unit further includes a shooting control unit for outputting a shooting signal to the shooting device. It is preferable to provide.

A sound device further includes a sound input unit for receiving sound information from the outside and a sound output unit for outputting a warning signal to the outside, wherein the control unit recognizes a target by processing sound information input from the sound input unit and outputs the sound output unit. It is preferable to further include a sound control unit for outputting a warning signal.

It is preferable that the boundary information includes at least one of recognition information of the target, movement information of the target, and sound information.

Preferably, the control command includes at least one of a tracking signal of a target, a control signal of a shooting apparatus, and a warning signal.

According to the present invention, it is possible to image the shape and movement of the object, to automatically track the movement of the object, automatic foaming on the object.

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

1 is a perspective view schematically showing an intelligent robot system as one preferred embodiment according to the present invention.

Referring to the drawings, the intelligent perimeter robot system 1 includes at least one master camera 11; Active camera 12; A driving unit 13; And a control unit 14. The intelligent perimeter robot system 1 according to the present embodiment is operated by two cameras of the master camera 11 and the active camera 12, and receives the movement information of the target from each of them and performs tracking for monitoring and boundary. It is configured to increase the rate and improve the recognition rate.

The master camera 11 is installed to be fixed to the frame 19 to recognize the target from the input image. The active camera 12 is installed to be rotatable in a tilting and panning direction with respect to the frame 19, so that the target can be tracked. The driving unit 13 rotates the active camera 12 vertically and horizontally with respect to the frame 19. The controller 14 receives an image from the master camera 11 and the active camera 12, recognizes the image, and controls the master camera 11, the active camera 12, and the driver 13. At this time, the controller 14 may be installed in the inside of the frame 19 shown.

In addition, a mobile simple controller 16 that can directly control the boundary robot system 1 can be used, and a separate operating personnel can directly operate the boundary robot system using the mobile simple controller 16 as necessary. There will be.

It is preferable to provide a shooting device 15 which is fixedly mounted to the active camera so that the direction in which the barrel is directed coincides with the central axis facing the target of the active camera. The apparatus may further include a sound device including a sound input unit for receiving sound information from the outside and a sound output unit for outputting a warning signal to the outside.

Preferably, the master camera 11 and the active camera 12 are ultra low light cameras having an infrared cut filter for blocking the image input of the infrared region. In this case, the master camera 11 and the active camera 12 turn on the infrared cut filter during the day and receive a color image, and turn off the infrared cut filter during the night and receive a black and white image. The ultra low light camera can be used to receive images during the day as well as at night.

Preferably, the master camera 11 has a larger viewing angle than the active camera 12. That is, the master camera 11 performs a function of detecting the overall movement in the field of view having a large viewing angle. At this time, the master camera 11 is also equipped with a zoom function, it is preferable to adjust the magnification and set according to the use conditions such as the observation distance and range. The master camera 11 acquires an image from a wide area of the field of view and recognizes the target and detects the overall movement of the target.

The active camera 12 is controlled to always maintain the center of the movement on the optical axis of the camera by using the motion information of the target recognized by the master camera 11. In addition, the active camera 12 detects information such as the speed, displacement, and size of the target moving more precisely while maintaining a higher resolution than the master camera 11.

To this end, the active camera 12 has functions of zooming, panning, and tilting up and down. By such left and right rotation and up and down rotation functions of the active camera 12, the optical axis of the active camera 12 is always driven so as to be maintained toward the center of the target. In addition, by zooming the image of the target by the zoom function of the active camera 12, it will be possible to observe the target in more detail.

Accordingly, since the direction in which the barrel of the shooting device 15 is fixed to the active camera 12 is directed to coincide with the central axis facing the target of the active camera 12, the barrel of the shooting device 15 is the target. To face.

In addition, a conventional tracking algorithm such as a centroid method or a correlation method may be used for the active camera 12 and the shooting device 15 to track the target.

The master camera 11 may recognize a target by sensing a change in an input image. For this purpose, the master camera 11 divides the input image into frame units, compares image signals in two or more frames, and compares the image signals. The target can be recognized from the difference of. For example, by comparing the image of the previous frame with the current frame, if an object that was not in the previous frame is in the current frame, it can be made a new target. In this case, if a moving object is detected in the input image, it may be targeted.

In this case, the input image divided in units of frames may be divided into images in smaller units than frames. For example, when an image is driven and displayed by a multi-scanning method, one frame is divided into field units, which are each scanning unit, and the target is determined from the difference of the image signal by comparing the images in each field unit. You might recognize it. In this case, recognition of the target, motion detection of the target, etc. may be performed from an image input from fewer image data.

Here, two or more cameras may be used for the master camera 11 and the active camera 12 for efficient operation of the intelligent perimeter robot system. That is, one master camera 11 may monitor a large area, and several active cameras 12 may simultaneously track a plurality of moving targets. In addition, two or more master cameras 11 may be used to efficiently monitor the entire area or to measure distance to the target.

In addition, the monitoring and alerting work is performed by the cooperative work of two or more cameras of the master camera 11 and the active camera 12, so that the monitoring and alerting service is continuously performed even when a new event occurs after one event occurs. This is possible.

The driving unit 13 rotates the active camera 12 vertically and horizontally with respect to the frame 19. The driving unit 13 receives a driving signal generated from the movement information of the target of the control unit and the position and angle information of the optical axis of the active camera 12, and rotates the active camera 12 up and down and left and right, and the active camera. 12 is driven to track the target.

The controller 14 receives images from the master camera 11 and the active camera 12 to recognize the image, detects the target, recognizes the shape of the target, detects the movement of the target, and accordingly, the master camera. (11), the active camera 12, and the drive unit 13 are controlled.

The sound apparatus includes a sound input unit and a sound output unit. The sound input unit may receive a sound signal from the outside, process the same in the controller, and supplement the image signal to obtain a type, a position, and an azimuth angle of the target. That is, the active camera 12 can be driven according to the information about the target obtained by the sound detection.

Tracking of the active camera 12 and the shooting device 15 according to the image input may be performed by the following method.

First, a target is detected by detecting the overall movement in the field of view from the video signal obtained from the master camera 11. In addition, the initial movement of the target is detected from the master camera 11 and the active camera 12, the shape is recognized by comparison with a reference image stored or learned in advance, and the movement of the target is tracked. In order to obtain the velocity and position information of the moving target for tracking, the position and velocity of the target are detected by comparing the patterns in the moving regions between the images.

In addition, the target recognition receives the target image passed from the detection unit and determines the type and state of the target. The control unit 14 receives the image tracking information, sound and audio information to control the operation of the driving unit, and the image data is sent to the mobile simple controller 26 or the situation control unit 27 shown in FIG. It can be recorded by the video recording device of the situation control unit.

In addition, the tracking and shooting of the target according to the detection of the image may be automatically performed by the control of the control unit 15, but may be made by an offline operation personnel. That is, when the situation is judged by the operating personnel, the operating personnel give a firing command, and the firing command is input to the control unit 14 through a communication means such as RS-485, so that the firing signal is received from the control unit 14. The shot is input to the shooting device 15 and shooting can be performed toward the target. At this time, since the barrel of the shooting device 15 is always coincident with the central axis of the active camera, shooting can be performed while tracking the image.

There are many methods for tracking algorithms, but the target tracking method and the correlation method are widely used in the conventional tracker. In the target center method, each time-varying image is separated into a target and a background and binarized, the center of the target is extracted, and the target movement information is estimated from the position of the center. This method has the advantage of being relatively simple to implement and capable of tracking moving targets with large displacements, whereas it is difficult to segment an image to separate the target and the background when the brightness of the target and the background is similar or complicated.

Therefore, in the present embodiment, a correlation tracking method is used as a tracking image signal processing method, which is an optical flow method that has been widely studied as a motion estimation method in the field of image processing and computer vision. It's the same way. This method is impossible to track moving targets with a large displacement compared to the target center method, but it is possible to estimate the movement information relatively stably in the tracking of small displacements, and it is not necessary to separate the target and the background separately. Estimation of is relatively easy.

As an example of a method for tracking a moving object by the correlation tracking method of the present embodiment, a method of detecting an object movement, tracking an object movement, tracking while zooming, tracking an object obscured by an obstacle, etc. Tracking can be done through a method such as tracking.

In addition, in the intelligent perimeter robot system according to the present invention, it is preferable to recognize the shape of the target object recognized by the master camera 11 and the active camera 12 so that appropriate measures can be taken according to the type of the shape.

To this end, the shape recognition part of the target tracking system receives an image of the detected target and determines the type of the target. At this time, the types of targets to be distinguished are varied, the size, shape or brightness distribution is also varied, and there may be a lot of noise due to various environmental changes. In order to solve this problem, in the present embodiment, it is difficult to accurately classify a single feature of each target, and thus, by combining various features of the same object, a lot of data for each object is learned in advance and classified for the input object. Neural network (NN) based technology or Root Mean Square Error (RMS) based technology can be applied.

Figure 2 is another preferred embodiment according to the present invention, which is a boundary system having at least one intelligent boundary robot system.

Referring to the drawings, the boundary system 2 comprises at least one of the intelligent boundary robot systems 28A, 28B, etc; And a situation controller 27 which receives boundary information from the intelligent boundary robot system and recognizes a situation and outputs a control command to the intelligent boundary robot system. In this case, the present embodiment includes the intelligent boundary robot system of FIG. 1, and similar reference numerals are used for the same components having the same functions as those of FIG. 1, and detailed description thereof will be omitted.

Intelligent boundary robot system according to the present embodiment, the master camera 21; Active camera 22; Acoustic device 29; Shooting device 25; And a control unit. In this case, the controller 24 may include an image processor 241; Tracking control unit 242; An acoustic controller 243; Fire control unit 244; And a main control unit 245.

The image processor 241 recognizes a target by processing image signals input from the master camera 21 and the active camera 22, and recognizes a target movement. The tracking controller 242 receives the target motion information from the image processor 241 and controls the vertical, horizontal, left and right rotation and zoom of the active camera 22 so that the active camera 22 can track the target. The sound controller 243 processes the sound information input from the sound input unit to recognize the target and outputs a warning signal to the sound output unit. The fire control unit 244 outputs a fire signal to the fire device 25.

The main controller 245 controls and interconnects the image processor 241, the tracking controller 242, the sound controller 243, and the fire controller 244 as a whole. The main controller 245 receives a target signal from the image processor 241, receives sound information from the sound controller 243, and receives various control command information from the situation controller 27. The target movement data is output to the tracking control unit 242, the warning control signal is output to the sound control unit 243, the fire control signal is output to the fire control unit 244, and various boundary information is output to the situation control unit 27. Output

At this time, the boundary information output from the main control unit 245 to the situation control unit 27 includes at least one of recognition information of the target, movement information of the target, and sound information. The control command input to 245 preferably includes at least one of a tracking signal of a target, a control signal of a shooting apparatus, and a warning signal.

The control command input from the situation control unit 27 to the main control unit 245 may be automatically performed. However, depending on the situation, a separate operation agent exists, and the operation agent monitors the video and audio information, and executes the control command. You might get off.

In addition, the boundary system 2 may further include a mobile simple controller 26 that can directly control the boundary robot system, and if necessary, a separate operating personnel uses the mobile simple controller 26 to control the boundary robot system. Allows direct manipulation of the system.

3 is a graph schematically illustrating the operation principle of the ultra low light camera of the intelligent perimeter robot system of FIGS. 1 and 2.

Referring to the drawings, the master camera (11 of FIG. 1) and the active camera (12 of FIG. 1), an ultra-low light camera having an infrared cut filter for blocking the image input of the infrared region may be used. In this case, it is preferable that the ultra low light camera has an ultra low light function capable of monitoring day and night at the same camera, and a WDR (Wide Dynamic Range) function which improves the subject's discrimination ability even under a backlight condition. It also has functions such as Auto White Balance (AWB), Auto Exposure (AE), and Auto Focus (AF).

The camera of the present embodiment may use a method of turning on the infrared cut filter during the day and receiving a color image, and turning off the infrared cut filter at night and receiving a black and white image. The ultra low light camera can be used to receive images during the day as well as at night.

More preferably, it is possible to obtain a color image not only during the day but also at night by using a color ultra-low camera that is equally applicable to the master camera 11 and the active camera 12 day and night. Surveillance will be more precise.

4 and 5a to 5c schematically illustrate the principle of distance measurement to a target by a master camera and an active camera of the intelligent boundary robot system of FIGS. 1 and 2.

Referring to the drawings, the boundary robot system according to the present invention can measure the position and azimuth angle by recognizing an object using two or more cameras, thereby measuring the distance without using an additional distance sensor.

4 illustrates the recognition of an object, that is, a person by two cameras of the master camera 31 and the active camera 32 in the present embodiment. At this time, the distance between the master camera 31 and the active camera 32 is w, the distance from these cameras to the object is D.

FIG. 5A illustrates a target recognized by the master camera 31, which is located at the right side of the image input to the master camera 31. 5B illustrates a target recognized by the active camera 32, which is located on the left side of the image input to the active camera 32. FIG. 5C is a composite image of the image input to the master camera 31 (FIG. 5A) and the image input to the active camera 32 (FIG. 5B), the left side being a target recognized by the active camera 32, and the right side The target recognized by the master camera 31.

In this case, the target recognized by the master camera 31 and the target recognized by the active camera 32 are separated by a predetermined distance d in FIG. 5C. In this case the distance D from the cameras to the object is inversely proportional to the distance d between the targets recognized by each camera. As is well known in the ordinary case, using this relationship, the distance D from the cameras to the object can be determined by the distance w between the cameras, the distance d between the recognized targets, the angle from the camera, and the like. Can be obtained by considering.

FIG. 6 is a diagram schematically illustrating an acoustic detection system of the intelligent perimeter system of FIG. 2.

Referring to the drawings, in the intelligent perimeter system, the perimeter for the intrusion and detection of the enemy, and the position and direction of the enemy using the acoustic sensor for the defense technology can effectively operate the perimeter system. It can be more useful in situations where it is difficult to capture images such as invading the enemy's camouflage in a short distance. That is, by using at least one acoustic sensor, such as a microphone, it is made in the form of an array as shown to detect the movement of the target when it occurs so that the boundary task can be efficiently performed. In the present embodiment, it can be mainly used as a function to assist recognition of a target by image recognition.

To this end, acoustic sensors of multiple channels (MICs) are used, and multiple signal classification (MUSIC) and coherent signal subspace (CSS) are used as multiple signal classification techniques for this. Acquire the signal through the acoustic sensor, process the acquired signal, remove the signal enhancement and noise signal, detect the delay and direction (DOA), analyze the signal and detect the target. The target detection information by this series of operations is transmitted to the main control unit.

In addition to the methods described above, various acoustic signal analysis and processing methods commonly used may be used for such an acoustic detection system.

7 is a flowchart schematically illustrating an operating procedure of the boundary system of FIG. 2.

Referring to the drawings, the operation of the boundary system may be performed by separating the boundary task performance 100 in the boundary robot and the task performance 200 in the situation control unit. It can be done simultaneously. In the case of the illustrated embodiment, the boundary task performance of one of the two or more boundary robots and the performance of the primary control station controller as the situation control unit are illustrated.

First, the boundary task performance of the boundary robot is as follows. While monitoring the sector determined by the master camera (S101), the moving object is measured (S102). Next, it is determined whether the target is recognized from the moving object (S103). If the target is recognized, the speed is measured by the master camera and the active camera, and the target is tracked by the active camera (S104). At this time, if the target is not recognized, the process returns to the step S101 of determining the sector boundary again.

In addition, it is determined whether the current mode is the combat mode (S105), and in the case of the combat mode, automatic shooting is performed through the shooting apparatus (S106). At this time, if the mode is not in the battle mode, the process returns to the step S101 of determining the sector boundary again. Subsequently, video surveillance of the situation area is continued (S107), and video and audio information is stored in a storage device to secure evidence data (S108).

In this case, the sound may be separately detected (S109), and the situation may be identified through the password recognition recognition (S110) by the password exchange by the sound apparatus.

Next, the flow of work performed by the primary control officer as the situation control department is as follows. The situation is monitored through the object information by the moving object measurement (S102) in the boundary robot (S201), the situation is determined according to whether the target is recognized (S103) (S202), the pia is identified (S203), and the boundary robot is Fire command can be issued (S204). At this time, the situation may be reported to a separate command and control command to obtain control orders.

In addition, the situation is identified from the sound detection information (S109, S110) and the image information (S107) of the situation area (S205), and the result of the combat situation is reported to the command control center (S206).

According to the intelligent boundary robot system according to the present invention, it is possible to recognize the shape and movement of an object, to automatically track the movement of the object, and to automatically fire the object.

Although the present invention has been described with reference to one embodiment shown in the accompanying drawings, it is merely an example, and those skilled in the art may realize various modifications and equivalent other embodiments therefrom. I can understand. Accordingly, the true scope of protection of the invention should be defined only by the appended claims.

1 is a perspective view schematically showing an intelligent robot system as one preferred embodiment according to the present invention.

Figure 2 is another preferred embodiment according to the present invention, which is a boundary system having at least one intelligent boundary robot system.

3 is a graph schematically illustrating the operation principle of the ultra low light camera of the intelligent perimeter robot system of FIGS. 1 and 2.

4 and 5a to 5c schematically illustrate the principle of distance measurement to a target by a master camera and an active camera of the intelligent boundary robot system of FIGS. 1 and 2.

FIG. 6 is a diagram schematically illustrating an acoustic detection system of the intelligent perimeter system of FIG. 2.

7 is a flowchart schematically illustrating an operating procedure of the boundary system of FIG. 2.

<Explanation of symbols for the main parts of the drawings>

1: intelligent perimeter robot system, 2: perimeter system,

11, 21: master camera, 12, 22: active camera,

13: drive part, 14, 24: robot control part,

15, 25: fire gear, 16, 26: mobile simple controller,

241: image processing unit, 242: tracking control unit,

243: sound control unit, 244: fire control unit,

245: main control, 27: situation control,

29: Acoustic device.

Claims (17)

At least one master camera installed to be fixed to a frame and recognizing a target from an input image; An active camera installed to be rotatable vertically and horizontally with respect to the frame, the active camera capable of tracking the target; A driver configured to rotate the active camera vertically and horizontally with respect to the frame; And And a controller configured to receive the image from the master camera and the active camera, recognize the image, and control the master camera, the active camera, and the driving unit. The method of claim 1, And a shooting device fixed to the active camera so that the direction in which the barrel is directed coincides with the central axis of the active camera toward the target. The method of claim 1, Further comprising a sound device including a sound input unit for receiving sound information from the outside, The intelligent boundary robot system to recognize the target by processing the sound information input from the sound input unit in the control unit. The method of claim 3, And the sound device further comprises a sound output unit for outputting a warning signal to the outside. The method of claim 1, Intelligent boundary robot system, wherein the master camera has a larger viewing angle than the active camera. The method of claim 1, Intelligent boundary robot system for the master camera to recognize the target by detecting a change in the input image. The method of claim 1, Intelligent boundary robot system that is provided with a zoom function in the active camera to enlarge the image for the target. The method of claim 1, The master camera and the active camera is an ultra low light camera having an infrared cut filter for blocking the image input of the infrared region, And the master camera and the active camera turn on the infrared cut filter and receive a color image during the day, and turn off the infrared cut filter and receive the black and white image at night. At least one master camera recognizing a target from an input image; An active camera capable of tracking the target; And And a controller configured to receive the image from the master camera and the active camera, recognize the image, and control the master camera and the active camera. The method of claim 9, The control unit, An image processor which processes the image signals input from the master camera and the active camera to recognize the target and recognizes the movement of the target; Intelligent tracking robot system comprising a tracking control unit for receiving the movement information of the target from the image processing unit to control the vertical, horizontal, left and right rotation and zoom of the active camera to track the target. The method of claim 10, And the image processor receives an image signal from the master camera to recognize the target, and recognizes the shape of the target and the movement of the target by at least one of the master camera and the active camera. The method of claim 11, And the master camera classifies the input image into frame units and compares image signals of two or more frames, respectively, and recognizes a target from the difference of the image signals. The method of claim 10, It is further provided with a shooting device which is rotated together by the rotation of the active camera, the direction in which the barrel is directed coincides with the central axis toward the target of the active camera, And the control unit further comprises a fire control unit for outputting the fire signal to the fire apparatus. The method of claim 10, A sound device further comprises a sound input unit for receiving sound information from the outside and a sound output unit for outputting a warning signal to the outside, And a sound controller configured to process the sound information input from the sound input unit to recognize the target and output the warning signal to the sound output unit. The intelligent perimeter robot system of any one or more of claims 1 to 14; And And a situation control unit that receives boundary information from the intelligent boundary robot system, recognizes a situation, and outputs a control command to the intelligent boundary robot system. The method of claim 15, And the boundary information includes at least one of recognition information of the target, movement information of the target, and sound information. The method of claim 15, And the control command comprises at least one of a tracking signal of the target, a control signal of the shooting device, and the warning signal.