JP2007312271A - Surveillance system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a surveillance system which is adaptive to a surveillant's state. <P>SOLUTION: The surveillance system comprises a screen image acquiring part for acquiring an objective screen image, a concerned person feature information storing part for storing concerned person feature information which is information created based on a concerned person's screen image which is previously photographed, a photographed person feature information extracting part which extracts a photographed person area indicating a person from the screen image acquired by the screen image acquiring part and extracts photographed person feature information from the photographed person area, a screen image analyzing part for creating surveillance information indicating whether or not an unconcerned person is included in the photographed person area based on the concerned person feature information and photographed person feature information, and an output part for outputting in accordance with the surveillance information. When the unconcerned person is included in the photographed area, the output part displays the objective screen image acquired by the screen image acquiring part more largely. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a monitoring system that performs monitoring using a camera.

  Video monitoring systems that simultaneously handle video signals input from a plurality of cameras are known. In recent years, video surveillance systems in which video is transmitted via a network and video surveillance systems in which video is transmitted wirelessly in order to omit wiring have been sold.

In Patent Document 1, in a wide-area monitoring apparatus using a large number of video inputs, a moving body imaged for each predetermined number of cameras is tracked, the identity thereof is determined, and transmission of information is limited. An invention for reducing the amount of communication is disclosed.
JP-A-11-15981

  Usually, it is painful for the observer to constantly monitor the monitoring terminal. In addition, there may be situations where the watcher's condition (periodic patrols, breaks, etc.) prevents the terminal from being constantly watched, and anomalies that occur in the monitoring area (area taken by the monitoring camera) may be overlooked. Sex exists. When the monitoring work is performed by a plurality of monitoring personnel, the number of persons in charge who are watching the monitoring image varies depending on the situation for the same reason. Such fluctuations are undesirable for stable monitoring.

  The surveillance camera also captures the persons who are waiting in the surveillance area and the surveillance personnel who are traveling around. These are redundant images for normal monitoring operations. If the supervisor monitors such redundant video for a long time, the attention to the object to be monitored is reduced. Furthermore, such redundant video is recorded on a limited recording medium, thereby preventing long-time video recording.

  When monitoring work is performed while switching a large number of camera images, the burden on the monitoring worker is great. When a large number of persons are imaged simultaneously on a large number of videos, it is difficult for a monitoring worker to track a specific monitoring target.

An object of the present invention is to provide a monitoring system that can be adapted to the condition of the observer.
Another object of the present invention is to provide a monitoring system that enables a monitoring operation to be performed by excluding the parties in the monitoring area.
Still another object of the present invention is to provide a surveillance system that reduces video selection work and redundant video display when a large number of surveillance cameras are installed.
Still another object of the present invention is to provide a monitoring system that displays an image limited for each target and can efficiently monitor a plurality of captured images with a small number of monitoring personnel.

  In the following, means for solving the problem will be described using the numbers used in [Best Mode for Carrying Out the Invention] in parentheses. These numbers are added to clarify the correspondence between the description of [Claims] and [Best Mode for Carrying Out the Invention]. However, these numbers should not be used to interpret the technical scope of the invention described in [Claims].

  The monitoring system according to the present invention includes a video acquisition unit (1) that acquires a target video, and a party feature information storage that stores party feature information that is information generated based on a video of a party captured in advance. A person-to-be-photographed person area (35) indicating a person from the video acquired by the video acquisition unit (9), and a person-to-be-photographed person characteristic information extraction unit ( 21), a video analysis unit (22) for generating monitoring information indicating whether or not a non-related person is included in the photographed person area based on the participant characteristic information and the photographed person characteristic information, and the monitoring information And an output unit (50) for performing a corresponding output.

  In the monitoring system according to the present invention, the output unit (50) displays a larger target video acquired by the video acquisition unit (1) when a non-related person is included in the subject area (35). .

  The monitoring system according to the present invention includes a second video acquisition unit (5) that acquires a video of a supervisor area (32) in which a monitor (50) on which a video acquired by the video acquisition unit (1) is displayed is installed. A participant who extracts a supervisor region feature from a video of the supervisor region and generates state information indicating the state of the participant in the supervisor region by collating the supervisor region feature with the participant feature information. A state information generation unit (63). The output unit (50) changes the output according to the state information.

  The monitoring system according to the present invention further includes a recording unit (10) for recording a history of state information.

  The surveillance system according to the present invention is extracted from the first subject feature information extracted from the first area included in the first video acquired by the video acquisition unit (1) and the second area included in the second video. The same subject determination unit (22) for associating the first area with the second area when the second subject characteristic information is matched based on a predetermined criterion and matched, and the first image and the second image A video evaluation value determining unit (23) that generates video evaluation values and evaluates the magnitude relationship of the video evaluation values. The output unit (50) performs output according to the magnitude relationship.

  A monitoring server according to the present invention includes a video acquisition unit (1) that acquires a target video, and a party feature information storage that stores party feature information that is information generated based on a video of a party captured in advance. A person-to-be-photographed person area (35) indicating a person from the video acquired by the part (9) and the video acquisition part (1), and the person-to-be-photographed person characteristic information for extracting the person-to-be-photographed person characteristic information An extraction unit (21), a video analysis unit (22) that generates monitoring information indicating whether or not a non-related person is included in the photographed person area based on the participant characteristic information and the photographed person characteristic information; And an output unit (50) that performs output in accordance with the monitoring information.

ADVANTAGE OF THE INVENTION According to this invention, the monitoring system which can be adapted to the condition of the observer is provided.
Furthermore, according to the present invention, there is provided a monitoring system that makes it possible to perform monitoring work by excluding persons concerned in the monitoring area.
Furthermore, according to the present invention, there is provided a monitoring system that reduces video selection work and redundant video display when a large number of surveillance cameras are installed.
Furthermore, according to the present invention, there is provided a monitoring system capable of displaying an image limited for each object and efficiently monitoring a plurality of objects to be imaged by a small number of observers.

[First Embodiment]
The best mode for carrying out the present invention will be described below with reference to the drawings. FIG. 1 shows a facility to which the monitoring system according to the first embodiment is applied. The monitoring system is used for a monitoring person (monitoring person) to monitor the monitoring area 31. The monitoring area 31 includes indoors, outdoors, or both.

  In order to monitor the monitoring area 31, at least one camera 101a to 101d is used. An area photographed by one camera and an area photographed by another camera may overlap. The video imaged by each camera is a still image or a moving image. In the following description, the cameras 101a to 101d capture color moving images.

  A monitor who monitors the monitoring area is located in the monitoring room 32. The monitoring room 32 is located at a part of the facility to be monitored or at a location away from the facility. A display terminal 50 is installed inside the monitoring room 32. The display terminal 50 displays an image of the monitoring area captured by the cameras 101a to 101d. A camera 105 is installed in the monitoring room 32. The camera 105 captures an area inside the monitoring room 32 including the position of the supervisor who monitors the video displayed on the display terminal 50.

  FIG. 2 shows the configuration of the monitoring system in the first embodiment. The monitoring system includes a video photographing means 1, a video analyzing means 2, a video selection display means 3, a monitoring room video photographing means 5, a supervisor status grasping means 6, a system operating condition determining means 7, a warning means 8, and a supervisor characteristic information recording. Means 9 are provided.

  The supervisor feature information recording unit 9 stores supervisor feature information created on the basis of the video of the supervisor (related person) taken in advance. The video photographing means 1 includes cameras 101a to 101d, and transmits monitoring area image information including images of the monitoring area 31 photographed by the cameras 101a to 101d to the video analysis means 2. The video analysis means 2 analyzes the received monitoring area image information, extracts the target area of the monitoring target (typically a person) photographed in the monitoring area, and previously sets the characteristics of the target area as the supervisor characteristics. By comparing with the supervisor characteristic information recorded in the information recording unit 9, a determination result for determining whether the target is a related person is generated and transmitted to the video selection display unit 3. The video selection display unit 3 generates video selection information based on the received determination result, generates a display screen based on the video selection information, and displays it on the display terminal 50. The video selection display unit 3 further controls the warning unit 8 depending on the received analysis result, and generates a warning for the supervisor.

  The monitoring room video photographing means 5 includes a camera 105, and transmits monitoring room image information including an image of the monitoring room 32 photographed by the camera 105 to the supervisor situation grasping means 6. Based on the received monitoring room image information and the supervisor feature information recorded in advance in the storage medium of the feature information storage means 9, the supervisor situation grasping means 6 is in a state for each supervisor (near the display terminal 50. Whether or not the user is in the monitoring room) is generated and transmitted to the system operating condition determining means 7. The system operation condition determination unit 7 generates system operation condition information based on the received supervisor status information and transmits it to the video selection display unit 3. The video selection display means 3 generates video selection information based on the received system operating condition information, generates a display screen based on the video selection information, and displays it on the display terminal 50. The video selection display means 3 further controls the warning means 8 depending on the received system operating condition information, and generates a warning for the supervisor.

  Such a monitoring system is realized using a computer. For example, the supervisor characteristic information recording unit 9 is realized using a computer-readable storage medium exemplified by a hard disk. The video analysis means 2, the video selection display means 3, the supervisor status grasping means 6 and the system operating condition determination means 7 are each realized by a program that causes a computer to execute a predetermined procedure by being stored in a storage medium and read by the computer. It is possible. The video photographing means 1, the monitoring room video photographing means 5 and the warning means 8 exchange information with a computer via an interface.

  FIG. 3 is a flowchart showing the operation of the monitoring system. The video photographing means 1 transmits a photographed image obtained by photographing the monitoring area 31 to the video analyzing means 2. The video analysis means 2 performs image recognition on the received captured image and extracts a target area where the monitoring target is shown. The video analysis means 2 further extracts feature information from each target area by a predetermined method (step A1).

  FIG. 5 shows the captured image 33 and the monitoring target 34 in the captured image 33. The video analysis means 2 identifies, for example, a monitoring target 34 that is a region indicating a captured human form in the captured image 33, extracts the region as a target region 35 as shown in FIG. Feature information is extracted from 35.

  The target region 35 is extracted by a conventional method exemplified by the background difference method and the region division method. From one video, 0, 1 or a plurality of target areas 35 are extracted. Information on each target area 35 is added with an identifier for identifying a video including the target area 35 and information on the time when the video was shot.

  The video analysis means 2 compares the extracted feature information with the supervisor feature information recorded in the supervisor feature information recording means 9, so that each supervisor 34 included in the photographed image 33 is registered in advance. A supervisor feature comparison process for determining whether or not the person is (related party) is performed (step A2).

  The video selection display unit 3 displays the captured image 34 based on the determination result obtained by the supervisor feature comparison process and indicating whether each monitoring target 34 in the captured image 33 is a supervisor (related party). A video selection display process for selecting and displaying to the user is performed (step A3).

  The video selection display means 3 selects the captured image 34 using the following method. For example, when a person other than the related person is shown in the captured image 33 in the monitoring area 31 where it is assumed that there is no person other than the related person, the captured image is displayed as a large image to be noted or with a warning display. It is reflected. FIG. 7 shows the display screen 40 of the display terminal 50. The captured image 34 captured by the cameras 101a to 101d is displayed in a small size in the camera image display area 41. A large image to be noted is displayed in the enlarged image display area 42. A warning message is displayed in the message display area 43.

  The cycle from step A1 to step A3 is continuously repeated at a fixed period. In parallel with this cycle, the process shown in FIG. 4 is executed.

  The monitoring room image photographing unit 5 transmits the monitoring room image obtained by photographing the monitoring room 32 to the supervisor situation grasping unit 6. The supervisor situation grasping means 6 performs image recognition on the received monitoring room video and extracts a target area where a target (typically a person) is shown. The supervisor situation grasping means 6 further extracts feature information from each target area by a predetermined method (step B1).

  The supervisor situation grasping means 6 compares the extracted feature information with the supervisor feature information recorded in the supervisor feature information recording means 9, so that each subject included in the monitoring room video is registered in advance. A supervisor feature comparison process for determining whether or not the person is (related party) is performed (step B2).

  The system operation condition determination means 7 determines the state of the supervisor based on the determination result obtained as a result of the supervisor feature comparison process and indicating whether each object of the monitoring room video is a supervisor (related party). recognize. The state of the monitor is determined by image recognition, sitting in front of the display terminal 50 in the monitor room 32, being in the monitor room 32 but away from the display terminal 50, in the break room, or in the monitor room 32. Based on the result of the determination and a predetermined standard, the four values of “in monitoring work”, “outside monitoring work and in the monitoring room”, “resting”, “away from seat” It is classified and recorded as either. The system operation condition determination unit 7 performs a system operation condition determination process for changing the operation condition of the video selection display unit 3 based on the recognized state of the supervisor.

  In order to improve the accuracy of the state to be determined, even if it is determined that the state of the monitor is in circulation only when the time when the monitor is not detected in the monitoring room or break room is longer than the predetermined period Good. In addition, when the supervisor area is limited, the status of the supervisor is in circulation only when the supervisor is not extracted after the supervisor is extracted near the entrance / exit in the video. You may determine that there is.

  The operating condition is changed as follows. When the video analysis means 2 shows that a person other than the related person is shown in the monitoring video in the monitoring area 31 and the supervisor status grasping means 6 recognizes that the monitor is in front of the display terminal 50, the display terminal 50 Is displayed on the display. The video analysis means 2 recognizes a person other than the related person in the monitoring video of the monitoring area 31 and the supervisor status grasping means 6 recognizes that the supervisor is not in front of the display terminal 50 but is in the monitoring room 32. When this is done, the display terminal 50 issues a warning to the supervisor by voice. Further, a warning is notified to the supervisor who is taking a break by radio.

  A plurality of monitoring targets are detected in the monitoring area 31 when it is determined that the status of a plurality of supervisors is simultaneously being monitored in the monitoring room 32 and each of them is watching a different monitoring terminal. In some cases, the monitoring system displays the representative video of each monitoring target separately on each monitoring terminal.

  According to the monitoring system according to the present embodiment, the monitoring target in the monitoring video of the monitoring area 31 is compared with the characteristics of the related parties recorded in advance, so that the monitoring work is performed by automatically excluding the related parties. Therefore, monitoring work sensitive to the situation where people other than those involved are photographed is realized.

  Furthermore, according to the monitoring system according to the present embodiment, the situation of the supervisor is grasped from the video taken of the surveillance room 32, and the operation of the surveillance system is automatically changed according to the situation, and the situation of the supervisor in the event of an abnormality. It is configured to be warned in response to Therefore, an efficient monitoring operation that suppresses oversight of an abnormality is realized without manually changing the setting in response to a change in the situation on the side of the supervisor.

  The monitoring system in the present embodiment can be suitably applied to the facility shown in FIG. 9 instead of the facility shown in FIG. In order to monitor the monitoring area 31 on the local side, at least one camera 101a to 101d is used. A supervisor who monitors the monitoring area 31 is located in the monitoring room 32. A display terminal 50 is installed inside the monitoring room 32. The display terminal 50 displays an image of the monitoring area captured by the cameras 101a to 101d. A camera 105 is installed in the monitoring room 32. The camera 105 captures an area inside the monitoring room 32 including the position of the supervisor who monitors the video displayed on the display terminal 50. Images taken by the cameras 101a to 101d and 105 are subjected to the same processing as in the case of the facility of FIG.

  A remote monitoring center 52 is installed on the remote side away from the monitoring area 31. A display terminal 54 is installed in the remote monitoring center 52, and a monitor who watches the screen of the display terminal 54 is arranged. The images of the cameras 101a to 101d and the image displayed on the display terminal 50 are transmitted to the remote monitoring center 52 via the network 56 which is an information communication line network exemplified by the Internet, and displayed on the display screen of the display terminal 54. Is displayed.

  When a warning is issued from the display terminal 50 by the monitoring system when the state of the monitor in the monitoring room 32 is resting, the monitoring system transmits a warning signal to the server that controls the display terminal 54 of the remote monitoring center 52. To do. When the server receives the warning signal, the display 54 issues a warning to the monitor at the terminal remote monitoring center 52.

  When a plurality of monitoring targets are included in the video, a certain target may be displayed on the display terminal 50 of the monitoring room 32 and another target may be displayed on the display terminal 54 of the remote monitoring center 52.

  The system operating condition determination means 7 displays an image on the display terminal 50 near the supervisor when the supervisor is in the monitoring room. The system operating condition determination means 7 is configured to display a representative video of a certain target to the monitoring room supervisor when the state of the monitoring person is being monitored in the monitoring room and a plurality of monitoring objects are detected in the monitoring area. At the same time, a representative image of another object is transmitted to the display terminal 54 of the supervisor in the remote monitoring center 52.

  When all the monitoring personnel are not in the monitoring room and are resting or traveling, the remote monitoring mode is set. In this case, the monitoring system transmits the image of the monitoring area 31 to the display terminal 54 of the monitor in the remote monitoring center 52. The video of the monitoring area 31 is displayed on the display terminal 54 and is monitored by a monitor of the remote monitoring center 52.

  In the remote monitoring mode, when the monitoring system recognizes that there is an intruder in the monitoring area 31, or recognizes the intrusion into the preset entry prohibition area, the display terminal 50 issues a warning to the monitoring person by voice. . The monitoring system further transmits an abnormality signal that notifies the server of the remote monitoring center 52 of the abnormality. When the server of the remote monitoring center 52 receives the abnormal signal, it issues a warning to the monitor of the remote monitoring center 52 from the display terminal 54 by video and audio. The supervisor who recognizes the warning contacts the local supervisor and confirms the situation.

  According to such a monitoring system, the monitoring video is transmitted to the remote monitoring center 52 only when assistance is required according to the state of the monitoring person in the local monitoring room 32. By monitoring the local side monitor and the remote side monitor in cooperation, it is possible to perform an effective monitoring operation with few oversights by a small number of supervisors. Since the monitoring video is transmitted only when necessary, the network bandwidth is used efficiently.

  A plurality of local monitoring systems having the configuration shown in FIG. 9 can be connected to one remote monitoring center 52 via the network 56. According to such a configuration, the remote monitoring center 52 can support the work of a local supervisor who monitors each of the plurality of monitoring areas 31. Therefore, more efficient and flexible monitoring work is realized as compared with the case where all the supervisors are arranged on the local site.

[Second Embodiment]
The monitoring system in the second embodiment is applied to the facility having the configuration shown in FIG. 1 as in the first embodiment. FIG. 10 shows the configuration of the monitoring system in the second embodiment. The monitoring system includes a video photographing means 1, a video analyzing means 2, a video selection display means 3, a monitoring room video photographing means 5, a supervisor status grasping means 6, a system operating condition determining means 7, a warning means 8, a supervisor characteristic information recording. Means 9 and video selection condition recording means 10 are provided. Such a monitoring system can be realized using a computer as in the first embodiment. In this monitoring system, the target given the same number as in FIG. 1 has the same function as in the first embodiment.

  The video analysis unit 2 includes a target area extraction unit 21, a same target determination unit 22, and a video evaluation value determination unit 23. The supervisor situation grasping means 6 includes a supervisor area extracting means 61, a feature matching means 62, and a situation determining means 63.

  The video selection condition recording unit 10 records the operation condition of the video selection display unit 3 determined by the system operation condition determination unit 7. The video selection display means 3 displays the video selected by the video analysis means 2 to the user according to the operating conditions recorded in the video selection condition recording means 10.

  The target area extraction means 21 in the video analysis means 2 distinguishes the area in which the target in the video acquired by the video shooting means 1 is shown from the background using image processing methods exemplified by the background difference method and the area division method. Extract.

  The same target determination unit 22 in the video analysis unit 2 analyzes the image feature amount including the color distribution and texture inside each target region extracted by the target region extraction unit 21. The same object determining means 22 identifies whether or not the same person is included in a plurality of videos. That is, the same target determining unit 22 compares the image feature amount of the first target region included in the first video with the image feature amount of the second target region included in the second video, and the same target (example) : Same person) is determined. As a result, when the same target is shown in a plurality of videos, an identifier indicating that the same target is shown in the first target area and the second target area is given.

  More specifically, the same object determining means 22 determines the identity of the object shown by comparing the characteristics obtained by a conventional image processing method for each target area. Examples of the region characteristics used at this time include a color distribution histogram, texture, size and shape in the image.

  The case where a color distribution histogram is used as the region characteristic will be described in more detail. For all pixels included in each target region, each pixel value obtained in RGB is converted into an HSI color space, a distribution histogram of the obtained hue H is obtained for each predetermined section, and each frequency is calculated as the total number of pixels in the region. The normalized histogram obtained by dividing by is set as the feature amount of the area. For all the regions included in each video, the similarity of the histogram is evaluated by the sum of the differences between the frequencies, regions below a predetermined threshold are regarded as the same target, and the same label is assigned to each region. As a result, a plurality of areas in which the same object is shown are associated. When characteristics other than the color distribution are used together, the identity of the object is evaluated using the weighted sum of the similarity based on the histogram and other similarities.

  The identity of the object is determined using only the region characteristics in the image. Therefore, it is not necessary to collect information on the angle of view or the enlargement ratio of the video photographing means 1 in order to obtain the position or shape of the target in real space.

  Further, the same target determining unit 22 compares the feature amount obtained for each target region with the feature amount of the related party recorded in the supervisor feature information recording unit 9, and whether the related party has moved to each target region. Determine whether or not. This determination can be performed on a plurality of target areas identified as having the same target by using feature values extracted from these target areas in common, thereby suppressing the matching process. .

  The video evaluation value determination unit 23 of the video analysis unit 2 evaluates each video based on the determination result obtained by the same object determination unit 22. Specifically, a high evaluation value is assigned to a video including a target area where a target other than the related person is captured. When a plurality of targets other than those related to the subject are photographed, a higher evaluation value is assigned to the target region having a higher degree of clarity calculated by a predetermined method for the target region in which the target is reflected.

Specifically, the video evaluation value determination means generates an evaluation value by processing information generated from the video using a conventional image processing method based on a predetermined standard. The following criteria are exemplified:
It is recognized that a face is imaged in the processing target area;
The area of the target area is a moderate area in the image;
The region of interest is near the center of the image;
Less hiding of objects by other objects;
The surrounding lighting conditions are good and the contrast of the target area is appropriate;
The target area is in focus and there is little motion blur; and it is close to a predetermined area in each image.
Based on these evaluation criteria, evaluation values obtained by existing image processing methods are obtained. That is, an evaluation value for each evaluation criterion is determined by analysis processing, and an evaluation value set for the target region is obtained. The weighted sum of the evaluation value set is set as the representative evaluation value.

  As an example, an evaluation value calculation process based on face recognition of a processing target area will be described. A conventional face area recognition method is applied to each image to perform face recognition in the video. When an area determined to have a face by the face recognition is included in the detected target area, 1 is added to the evaluation value of the target area. The evaluation value of the area that is not detected is set to zero.

  The video selection display unit 3 selects and displays an appropriate video from a plurality of videos of the same target according to the conditions and parameters recorded in the video selection recording unit 10 according to the evaluation value set. Further, when it is detected that the state of the supervisor is resting and it is considered that the supervisor is not gazing at the image, the warning means 8 is used to issue a warning to the supervisor or the monitoring target.

When the target area is small, an enlarged image is displayed by the display terminal 50. When multiple objects are included in the video, the following display method is applied:
Target areas with higher representative evaluation values are displayed;
The area of the screen is divided and each object is displayed in the divided area;
Images containing each object are displayed alternately in time-division;
A reduced image of each target area is presented to the user, and the reduced image selected by the user is enlarged and displayed.

  FIG. 8 shows an example of the display method. A plurality of reduced images are displayed in the reduced image display area 44. The image selected by the user with the mouse among the plurality of reduced images is enlarged and displayed in the enlarged image display area 42.

  The monitoring system can divide a plurality of videos showing a single target selected by the user from among a plurality of captured images, and can divide the behavior of a specific target from a plurality of viewpoints. Like that.

  The supervisor area extraction means 61 in the supervisor situation grasping means 6 distinguishes the area where the object in the video of the monitoring room 32 where the monitoring worker acquired by the monitoring room video photographing means 5 performs the monitoring work is shown from the background. The process to extract is performed. This process is executed by a conventional method exemplified by the background subtraction method or the region division method, similarly to the target region extraction means 21.

  The feature collating unit 62 in the supervisor situation grasping unit 6 uses, for each target region obtained by the supervisor region extracting unit 61, the image feature amount including the color distribution and texture in the region as in the same target determining unit 22. Analysis is performed and the obtained feature quantity is compared with the feature quantity of the related party recorded in the supervisor feature information recording unit 9 to determine whether each target area corresponds to the related party.

  The situation determination means 63 in the supervisor situation grasping means 6 judges the situation of the supervisor from the state including the presence / absence, number and position of monitoring workers in the monitoring room 32 based on the judgment result obtained by the feature matching means 62. To the system operating condition determining means 7.

  FIG. 11 is a flowchart showing the operation of the monitoring system. The video photographing means 1 transmits a photographed image obtained by photographing the monitoring area 31 to the video analyzing means 2. The target area extraction means 21 performs a target area extraction process for extracting each target area in the video from the received captured image (step C1).

  The same target determining unit 22 analyzes the image feature amount including the color distribution and texture in the region for each target region extracted by the target region extraction process, and obtains the obtained feature information in each target region belonging to a plurality of videos. In comparison, for similar target areas, a process of assigning a common area number indicating that the same target is photographed is performed (step C2).

  The same target determining unit 22 includes an image feature amount including an intra-region color distribution and texture for each target region extracted by the target region extraction process, and feature information (monitoring relationship) recorded in advance in the supervisor feature information recording unit 9. A supervisory feature comparison process for generating a determination result for determining whether or not the target in the image is a registered monitoring party (step C3).

  The video evaluation value determination means 23 performs a video evaluation value determination process for determining an evaluation value for each video based on the determination result obtained as a result of the supervisor feature comparison process and the feature amount of each target area (step) C4).

  The video selection / display unit 3 selects a video based on the evaluation value of each video obtained as a result of the video evaluation value determination process and the rules pre-recorded in the video selection condition recording unit 10. A video selection display process to be displayed to the (supervisor) is performed (step C5).

  The cycle from the above steps C1 to C5 is continuously repeated at a fixed period. In parallel with this cycle, the process shown in FIG. 12 is executed.

  The monitoring room image photographing unit 5 transmits the monitoring room image obtained by photographing the monitoring room 32 to the supervisor situation grasping unit 6. The supervisor area extraction means 61 performs a supervisor area extraction process for extracting each target area in the video for the received monitoring room video (step D1).

  The feature matching unit 62 compares the feature of each target extracted by the supervisor region extraction process with the feature information stored in the supervisor feature information storage unit 9, and the monitoring relationship in which the target in the image is registered in advance. A supervisor feature comparison process is performed to determine whether the person is a person and generate a determination result (step D2).

  The situation determination unit 63 performs a situation determination process for determining the state of the monitoring worker based on information including the determination result obtained as a result of the supervisor feature comparison process and the target position (step D3).

  The system operating condition determining unit 7 performs a system operating condition determining process for changing the operating condition of the video selection display unit 3 based on the state of the monitoring worker obtained as a result of the situation determining process (step D4).

  The monitoring system according to the present embodiment associates target areas in which the same target is shown from a plurality of videos, and displays a limited video for each target. Therefore, the work of switching a large number of surveillance camera videos is reduced, and redundant video display is suppressed. Therefore, it is possible to efficiently monitor a plurality of captured objects with a small number of supervisors.

  The monitoring system according to the present embodiment does not need to set field angle information in the real space of each camera. For this reason, when the camera is installed, the work of obtaining information indicating the installation angle can be omitted. Furthermore, operation using a camera with a variable angle of view is easy.

FIG. 1 shows a facility to which a monitoring system is applied. FIG. 2 shows the configuration of the monitoring system. FIG. 3 is a flowchart showing the operation of the monitoring system. FIG. 4 is a flowchart showing the operation of the monitoring system. FIG. 5 shows a captured image and a monitoring target. FIG. 6 shows the target area extracted from the captured image. FIG. 7 shows a display screen. FIG. 8 shows an example of the display method. FIG. 9 shows a facility to which the monitoring system is applied. FIG. 10 shows the configuration of the monitoring system. FIG. 11 is a flowchart showing the operation of the monitoring system. FIG. 12 is a flowchart showing the operation of the monitoring system.

Explanation of symbols

33 ... Captured image 40 ... Display screen 41 ... Camera image display area 42 ... Enlarged image display area 43 ... Message display area 44 ... Reduced image display area 101 ... Camera 105 ... Camera

Claims (6)

A video acquisition unit for acquiring the target video;
A party feature information storage unit for storing party feature information, which is information generated based on a video of a party shot in advance;
A person-to-be-photographed person feature information extracting unit for extracting a person-to-be-photographed person area indicating a person from the video obtained by the video obtaining unit, and extracting person-to-be-photographed person characteristic information from the person to be photographed;
A video analysis unit that generates monitoring information indicating whether or not a non-related person is included in the subject person area based on the participant characteristic information and the subject characteristic information;
A monitoring system comprising: an output unit that performs output according to the monitoring information. A monitoring system according to claim 1, wherein
The output unit displays the target video acquired by the video acquisition unit larger when the non-related person is included in the subject area. A monitoring system according to claim 1 or 2, wherein
A second video acquisition unit that acquires a video of a supervisor area in which a monitor on which the video acquired by the video acquisition unit is displayed;
By extracting a supervisor area feature quantity from the video of the supervisor area and comparing the supervisor area feature quantity with the related party feature information, state information indicating the status of the related party in the monitor area is obtained. A related party state information generation unit to generate,
The output unit changes the output according to the state information. A monitoring system according to claim 3, wherein
Furthermore, a monitoring system comprising a recording unit for recording the history of the state information. The monitoring system according to claim 3, wherein:
Furthermore, the first subject feature information extracted from the first region included in the first image acquired by the image acquisition unit and the second subject feature extracted from the second region included in the second image. The same object determining unit associating the first area with the second area when information is matched based on a predetermined criterion and matched,
A video evaluation value determining unit that generates a video evaluation value for each of the first video and the second video and evaluates a magnitude relationship between the video evaluation values;
The said output part is a monitoring system which performs the output according to the said magnitude relationship. A video acquisition unit for acquiring the target video;
A party feature information storage unit for storing party feature information, which is information generated based on a video of a party shot in advance;
A person-to-be-photographed person feature information extracting unit for extracting a person-to-be-photographed person area indicating a person from the video obtained by the video obtaining unit, and extracting person-to-be-photographed person characteristic information from the person to be photographed;
A video analysis unit that generates monitoring information indicating whether or not a non-related person is included in the subject person area based on the participant characteristic information and the subject characteristic information;
A monitoring server comprising: an output unit that performs output according to the monitoring information.

Images