KR100779858B1 - picture monitoring control system by object identification and the method thereof - Google Patents

Abstract

The present invention relates to a video surveillance control system and method by object recognition, and to extract only the part of the object detected motion from the input image and to compare with the comparison sample image to determine the type of the detected object, The purpose is to prevent unnecessary recording and alarm by controlling the recording and alarm. The video surveillance control system according to the object recognition of the present invention configured to achieve this object is to perform the necessary control by analyzing the image input in real time through the image input unit, the image input unit for taking an image of the surveillance area in real time An image surveillance control system according to an object recognition comprising a surveillance control unit and an image recording unit for recording an image input through an image input unit according to an instruction of the surveillance control unit, wherein the object in the image input in real time through the surveillance control image input unit Motion detection module for detecting the movement of the; An OFI group module in which a plurality of OFIs, which are comparative sample image data for a monitored object, are stored; And an object compare inference (OCI) module for determining the type of the corresponding object by comparing the detected object with the OFI.

Surveillance, image, object recognition, category, motion detection, tag

Description

Image monitoring control system by object identification and the method

1 is a block diagram of a conventional video surveillance control system.

Figure 2 is an example screen for explaining the operation of the conventional video surveillance control system.

Figure 3 is a block diagram of a video surveillance control system according to the object recognition of the present invention.

4 is a flowchart for explaining a method for controlling video surveillance by object recognition according to the present invention.

5A to 5C are exemplary screens for explaining a method for controlling video surveillance by object recognition according to the present invention.

[Description of Symbols for Main Parts of Drawing]

10. Video input unit 20. Monitoring control unit

22. Operator setting module 24. Motion detection module

26. Recording Control Module 30. Alarm Output

40. Video Recorder

100. Video input unit 200. Monitoring control unit

210. Operator setting module 220. Motion detection module

230. OFI Group Module 240. OCI Module

250. Recording control module 260. Tag generation module

270. Video Encoding Module 300. Alarm Output

400. Video recording unit

The present invention relates to an image surveillance control system and method by object recognition, and more particularly, by combining a motion detection function and an object recognition function to selectively monitor only a desired monitoring target object to prevent unnecessary recording and alarming. Of course, the present invention relates to a video surveillance control system and method using object recognition to quickly search by type of recorded video.

1 is a block diagram of a conventional video surveillance control system.

As shown in FIG. 1, the video surveillance control system according to the related art analyzes an image inputted in real time through an image input unit 10 and an image input unit 10 to capture an image of a surveillance area in real time. The alarm controller 20 outputs an alarm according to the instructions of the video recording unit 40 and the monitoring controller 20 which record an image input through the video input unit 10 according to the instruction of the monitoring controller 20. It may be made to include an alarm output unit 30.

In the above-described configuration, the image input unit 10 may be implemented as a conventional surveillance video camera, and the alarm output unit 30 may be a sound function of a computer or a siren, a buzzer, a warning lamp, or the like separately installed. The video recording unit 40 may be implemented as a hard disk (HDD) which is a storage medium of the computer itself, a video server or a digital video recorder (DVR) through a network communication module, or the like. There will be. Of course, it could be implemented as an analog video cassette recorder (VCR) using magnetic tape.

On the other hand, the monitoring control unit 20 may be implemented by a computer and operating software mounted thereon, specifically, the operator setting module 22 and the image input unit 10 to support the operator to make various settings necessary for monitoring control The motion detection module 24 detects the motion using a difference image between images input in real time through a real time or uses a Kalman Filter method and operates the alarm output unit 30 when the motion is detected. And a recording control module 26 which controls the image recording unit 40 to record the image input through the image input unit 10 to the image recording unit 40 when the motion detection fact is received from the motion detection module 24. It can be made, including).

2 is an exemplary screen for explaining an operation of a conventional video surveillance control system.

As shown in FIG. 2, the motion monitoring function in the video surveillance control system according to the related art detects motion by comparing differences between front and rear images of blocks of a certain size in a simple grid shape, and thus detects motion (or events). Although recording and alarming can be performed according to occurrence, there is a problem that recording and alarming are frequently performed even on the movement of an object that is not a desired detection target because the detected movement is unknown.

Furthermore, there is a problem in that the tag or marking for each type is not made at all in performing the recording, and therefore, when recording for a long time, the operator must check all of these images in order to find a necessary image from the recording.

The present invention was devised to solve the above-mentioned problems of the prior art, and it is possible to grasp the type of the detected object by extracting only the part of the object whose motion is detected from the input image and comparing it with the comparison sample image. It is an object of the present invention to provide a video surveillance control system and method using object recognition to prevent the unnecessary recording and alarm by controlling the recording and alarm based on the received information.

In addition, another object of the technique according to the present invention is to extract only the part of the object where the motion is detected from the input image and to compare it with the comparison sample image to determine the type of the detected object, and in performing the recording, By automatically recording and recording, it is possible to quickly search by type of recorded video.

In accordance with an aspect of the present invention, an image surveillance control system according to an object recognition apparatus includes a video input unit for photographing an image of a surveillance region in real time, and a surveillance control unit for performing necessary control by analyzing an image input in real time through the image input unit. In the video surveillance control system by the object recognition comprising a video recording unit for recording the video input through the video input unit according to the instructions of the monitoring control unit, the monitoring control unit of the object in the image input in real time through the video input unit A motion detection module detecting a motion; An OFI group module in which a plurality of OFIs, which are comparative sample image data for a monitored object, are stored; And an object compare inference (OCI) module for determining the type of the object by comparing the detected object with the OFI.

Meanwhile, according to another aspect of the present invention, a method for controlling video surveillance by object recognition includes (a) selecting an OFI for a monitoring target object from an OFI group module in which a plurality of OFIs, which are sample data for comparison, are stored. Receiving step; detecting movement of the input image by subframe image subtraction (difference image) when arbitrary frames of the input image are input; (c) adjusting the size of the input image based on the center position of the area labeling in which the motion is detected; (d) determining whether a match ratio with the selected OFI is equal to or greater than a predetermined reference ratio REF while rotating the outline of the adjusted input image in the horizontal and vertical directions; And (e) determining that the monitoring target object is equal to or greater than the reference ratio REF to perform necessary control.

Hereinafter, with reference to the accompanying drawings will be described in detail with respect to the video surveillance control system and method by the object recognition according to an embodiment of the present invention.

3 is a block diagram of a video surveillance control system by object recognition according to the present invention.

As shown in FIG. 3, the video surveillance control system based on object recognition according to an embodiment of the present invention provides a video input unit 100 and a video input unit 100 that capture images of a surveillance area in real time. In accordance with the instructions of the monitoring control unit 200 and the monitoring control unit 400 to record an image input through the video input unit 100 according to the instructions of the monitoring control unit 200 to perform the necessary control by analyzing the monitoring control unit 400. It may be made by including an alarm output unit 300 for outputting the alarm.

In the above-described configuration, the image input unit 100 may be implemented as a conventional surveillance video camera, and the alarm output unit 300 may be, for example, a sound function of a computer or a siren, a buzzer, a warning lamp, or the like separately installed. The video recording unit 400 may be implemented as a hard disk (HDD) which is a storage medium of a computer itself, a video server or a DVR through a network communication module, or the like.

On the other hand, the monitoring control unit 200 according to the present invention may be implemented by a computer and the operating software mounted thereon, specifically, the operator from the various settings necessary for monitoring control, for example, from the video camera for surveillance video input unit 100 Operator setting module 210 to set the display, recognition pattern setting and recording or alarm generation of the image to utilize the input image, motion detection module 220 for detecting the movement of the object in the input image, that is, movement OFI group module that stores a large number of comparison sample image data (hereinafter referred to as "OFI", which stands for Object Frame Image) to recognize what kind of object an object is detected in the input image. 230), a configuration including an object compare inference (OCI) module 240 for comparing and determining an object from which motion is detected with OFI, which is sample image data for comparison. And to the type classification of jikim the sensing object can be detected only monitoring the object aimed by the.

In addition, the monitoring control unit 200 additionally generates the type of the object (for example, the ID data) identified in the OCI module 240 and the time stamp data including the date and the time when the video is recorded (hereinafter, integrated as described above). A tag generation module 260 for generating the generated data " tag data " and an image encoding module 270 for integrating the tag data generated in the tag generation module 260 with the image data to be recorded. It makes it possible to quickly search by object type or time in the image. In the drawing, reference numeral 250 instructs the tag generation module 260 to generate tag data under the control of the OCI module and transmits the input image to the image encoding module 270 to generate the tag in the tag generation module 260. Represents a recording control module for encoding the tagged data into an input video signal.

On the other hand, the motion detection module 220 detects motion by using the inter-frame difference of each frame of the image input through the image input unit 100. Specifically, if any difference occurs between the frames of the image, The motion is detected and tracked according to the result calculated by applying the image subtraction (difference image) formula between frames. In addition, the motion detection module 220 performs the tracking function of the monitored object by extracting the absolute value of the difference image, performing region labeling, and displaying the center position of the region.

In addition, the OFI group module 230 includes each object (for example, in a state in which characteristic data such as speed, size, 3D information, or color is included in image data in the form of a vector map representing an outline characteristic of each object. , Files such as cars, airplanes, people, dogs, etc.) are stored by giving them unique IDs.

Hereinafter, the operation of the video surveillance control system by object recognition according to the present invention will be described in detail with the method.

4 is a flowchart illustrating a method for controlling video surveillance by object recognition according to the present invention, and FIGS. 5A to 5C are exemplary screens for explaining a method for controlling video surveillance using object recognition according to the present invention. Unless otherwise stated, the monitoring control unit 200 may be performed as a subject.

As shown in FIG. 4, first, at step S10, at least one or more OFIs to be monitored or selected from the OFI group module 230 to be monitored by an operator through the operator setting module 210 are received and thus selected at step S12. The size, change, or speed of each OFI is selected or set. The maximum and minimum sizes of the set OFIs can be used to determine whether recognition of the monitored object shown in the image is valid.

Furthermore, the change and speed of the OFI are determined by the unit time of the monitored target object in the image, for example, the rate of change of the environment per second, 1 minute or 1 hour, and the unit time of the monitored object, eg movement by 1 second or 1 minute. Speed, which can also be used to determine whether the object being monitored is valid. In this step S12, it is also possible to be set whether to alarm and recording.

Next, in step S14, it is determined whether shooting has started. If it is not started, step S14 is repeated, whereas in the case of start, the process proceeds to step S16. Determine if the frame is input.

If no frame is input as a result of the determination in step S16, step S16 is repeatedly performed, whereas if it is input, the process proceeds to step S18 and any difference value between the frames of the image using the inter-frame difference is obtained. If the difference occurs, the motion is detected according to the calculation result by applying the image subtraction (difference image) formula between the two frames, and the absolute value of the difference image is extracted again to perform the area labeling and the area. It will track the object to be monitored by displaying its center position.

Step S18 may be performed by the motion detection module 220. FIG. 5A shows a result of extracting only the vehicle part from the entire input image in this manner.

On the other hand, since the image input through the surveillance video camera, the image input unit 100 is photographed at various angles and positions, the size and angle of the monitoring object included in the input image may be different, which is OFI In order to compare with the OFI (FIG. 5B) stored in the group module 230, in step S20, the size of the object whose motion is detected is based on the center position of the area labeling where the motion is detected, as shown in FIG. 5C. Adjust to the same size.

In steps S22 and S24 subsequent to step S20, as shown in FIG. 5C, the outline of the object whose movement is detected is rotated horizontally and vertically, and the OFI selected in step S10 (FIG. 5B) and a predetermined reference ratio REF or more are exceeded. It checks whether the coincidence value comes out to determine whether the object whose motion is detected is the target object to be monitored.

If the matching ratio is greater than or equal to the reference ratio REF as a result of the determination in step S24, it corresponds to the case where the object whose motion is detected is the target monitoring target, and the flow proceeds to step S26 with the unique ID code data and the shooting time for the corresponding OFI. Generate tag data consisting of a timestamp indicating that step S26 may be performed by the tag generation module 260. Next, in step S28, the tag data generated as described above is recorded in the video recording unit 400 while being encoded together with the input video signal.

In step S30 subsequent to step S28, it is determined whether the alarm function is set. If so, the alarm generation unit 300 is driven in step S32 to generate an alarm, and in step S34, it is determined whether the recording is completed. .

As a result of the determination in step S34, when the recording is not completed, the process returns to step S30, whereas when the recording is completed, the program ends. On the other hand, if the alarm function is not set in step S30, the routine proceeds directly to step S34. If the coincidence ratio does not reach the predetermined reference ratio REF in step S24, the process returns to step S16. The determination of the completion of recording in step S34 may be performed based on a time or condition previously set by the operator.

Image surveillance control system and method according to the object recognition of the present invention can be used for the construction of a practical system, such as monitoring and retrieval of intrusion surveillance in the restricted access area, transport or installation of goods in the restricted area, and in addition to the type of vehicle It is also applicable to the field of classification.

The video surveillance control system and method according to the object recognition of the present invention are not limited to the above-described embodiments and can be modified in various ways within the scope of the technical idea of the present invention. For example, in the embodiment of FIG. 3, each function module of the monitoring controller 200 may be arbitrarily divided, and may be appropriately integrated or separated.

As described above, according to the video surveillance control system and method according to the present invention, it is possible to accurately grasp the target object to be monitored from the input video, and to control the recording and alarm based on the information thus obtained, thereby unnecessary recording. And alarms can be prevented.

Furthermore, in performing the recording, the marked type information is automatically marked and recorded so that the search for the type of recorded video can be performed quickly.

Claims (5)

An image input unit for capturing an image of a surveillance area in real time, a surveillance control unit for performing necessary control by analyzing an image input in real time through the image input unit, and recording an image input through the image input unit according to the instruction of the surveillance control unit In the video surveillance control system by the object recognition made of a video recording unit, The monitoring control unit includes a motion detection module for detecting the movement of the object in the image input in real time through the image input unit; An OFI group module in which a plurality of OFIs, which are comparative sample image data for a monitored object, are stored; And It includes an object compare inference (OCI) module to determine the type of the object by comparing the object detected the movement with the OFI, The monitoring control unit may generate a tag generation module including tag data including the identified object type data and an image encoding module encoding the tag data generated by the tag generation module to an input image and outputting the same to an image recording unit. Video surveillance control system by object recognition, further comprising. delete The OFI group module of claim 1, wherein the OFI group module assigns a unique ID to each object in a state of including characteristic data such as speed, size, 3D information, or color in image data in a vector map form representing the outline characteristic of each object. Video surveillance control system by object recognition, characterized in that the data files are stored. (a) receiving an OFI selected for the monitoring target object from an OFI group module in which a plurality of OFIs, which are comparative sample image data for the monitored object, are stored; detecting movement of the input image by subframe image subtraction (difference image) when arbitrary frames of the input image are input; (c) adjusting the size of the input image based on the center position of the area labeling in which the motion is detected; (d) determining whether a matching ratio with the selected OFI is equal to or greater than a predetermined reference ratio REF while rotating the outline of the adjusted input image in the horizontal and vertical directions; And and (e) determining that the monitoring target object is the monitoring target object when the matching ratio is equal to or greater than the reference ratio REF, and performing necessary control. 5. The method of claim 4, further comprising, after step (e), generating tag data including unique ID code data of the OFI corresponding to the surveillance target object, and encoding and recording the input image. Video surveillance control method by object recognition.

Images