JP2017028561A - Image monitoring system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image monitoring system for efficiently monitoring a delivery action at a place where unspecified many persons gather.SOLUTION: An image monitoring system includes person detecting means for detecting a person existing in a monitoring area from an image, possessed object detecting means for identifying the type of an article possessed by the person from the image, delivery detecting means for detecting article delivery actions by plural persons, identifying occurrence positions of the delivery actions, and referring to an identification result of the possessed object detecting means to determine the types of the articles delivered through the delivery actions, a storage unit for storing monitoring reference information associated with a monitoring importance degree representing a degree to be noted as a monitoring target for each combination of the position within a monitoring area and the type of an article, monitoring importance degree calculating means for referring to the monitoring reference information and determining the monitoring importance level corresponding to the combination of the occurrence position and the type of the delivered article for each delivery action, and output control means for outputting a detection result of the delivery action according to the monitoring importance degree of the delivery action.SELECTED DRAWING: Figure 11

Description

  The present invention relates to an image monitoring system that detects a monitoring target from an image obtained by imaging a monitoring location, and more particularly to an image monitoring system that detects an article delivery action performed between persons as a monitoring target.

  Places where a large number of unspecified people gather, such as airports, train stations, and event venues, are likely to become crime targets. Therefore, in such a place, it is necessary to increase caution against a suspicious person or a suspicious object.

  Conventionally, in order to ensure the safety of a place where a large number of unspecified persons gather, the video is taken with the video camera as a monitoring place, and the video is monitored by a supervisor or analyzed, and a suspicious person or Surveillance was being monitored. For example, Patent Document 1 discloses a technique for detecting an object left in a monitoring place and leaving it by image analysis and outputting an alarm as a suspicious object. According to such a technique, it is possible to detect that a dangerous substance such as a bomb has been left in the monitoring place, and it is possible to prevent danger in advance.

JP-A-1-245395

  By the way, dangerous materials such as bombs may be handed over between suspicious individuals before being left on site. If the dangerous goods can be monitored with this delivery action as an opportunity, the dangerous goods can be dealt with earlier than before.

  However, there are a large number of people and their belongings at airports, stations, and event venues. And these persons deliver the possessed goods at various places and at various timings. For this reason, it is extremely difficult for the monitoring person to visually monitor the image of the monitoring place and to monitor all delivery actions.

  In addition, there are many actions with low suspiciousness among the delivery actions performed in such places. For this reason, the surveillance staff needs to visually check the video while distinguishing between a high suspicious degree delivery action and a low suspicious degree delivery action.

  In other words, there is a problem that a large burden is placed on the supervisor in order to monitor the delivery action in a place where a large number of unspecified persons gather.

  The present invention has been made in view of these problems, and an object of the present invention is to realize an image monitoring system capable of efficiently monitoring a delivery action even in a place where a large number of unspecified persons gather.

  In order to achieve such an object, the present invention provides an image monitoring system that monitors an action of a person existing in a monitoring area using an image obtained by photographing the monitoring area, and the person existing in the monitoring area from the image A person detecting means for detecting the object, a belonging detecting means for identifying the type of the article possessed by the person from the image, a delivery action of the article by a plurality of the persons, and the monitoring area for the delivery action A delivery detecting means for obtaining a position where the delivery act is generated in the document and referring to an identification result of the possession detecting means for obtaining a type of the article delivered by the delivery act, and a position and the article in the monitoring area. A storage unit that stores monitoring reference information that associates a monitoring importance level that represents a degree of attention as the monitoring target for each combination with the type of The monitoring importance calculation means for obtaining the monitoring importance corresponding to the combination of the generation position and the type of the delivered article for each delivery action with reference to the visual criteria information, and the monitoring importance of the delivery action According to the present invention, there is provided an image monitoring system comprising output control means for outputting a detection result of the delivery action.

  In addition, it is preferable that the monitoring importance calculation unit calculates the congestion level of the person around the position where the delivery action occurs, and changes the monitoring importance according to the congestion level.

  Further, the storage unit further stores, for each delivery action, the occurrence position of the delivery action detected within the past predetermined time in the monitoring area, and the monitoring importance degree calculation unit is configured to determine the delivery action occurrence position. It is preferable to change the monitoring importance according to the total number of delivery actions detected in the past predetermined time around.

  In addition, the storage unit further stores the position of the important area set in the monitoring area, and the monitoring importance calculation means calculates a distance from the position where the delivery action occurs to the position of the important area. It is preferable to change the monitoring importance according to the distance.

  The image monitoring system further includes monitoring person position calculating means for obtaining a position of the monitoring person satisfying a predetermined criterion in the monitoring area, and the monitoring importance degree calculating means is configured based on the position where the delivery act is generated. It is preferable to calculate a distance to the position of the monitoring person and change the monitoring importance according to the distance.

  Further, it is preferable that the delivery detection means further obtains the occurrence time of the delivery action, and the monitoring importance calculation means further obtains the monitoring importance using the occurrence time of the delivery action.

  In addition, the image monitoring system includes a display unit that displays a detection result of the delivery action, and the output control unit causes the display unit to display the delivery action with higher monitoring importance so that the delivery action is more noticeable. Is preferred.

  The output control means outputs the detection result of the delivery action for a delivery action with the monitoring importance level equal to or higher than a predetermined value, and detects the delivery action for the delivery action with the monitoring importance level less than the predetermined value. It is preferable not to output the result.

  Further, the present invention is an image monitoring system that monitors an action of a person existing in a monitoring area using an image obtained by photographing the monitoring area, and identifies the type of article existing in the monitoring area from the image. The possession detection means detects a pattern composed of a plurality of persons existing in the monitoring area from the image and an article delivered between the persons from the image to detect a delivery action between the persons. Determining the occurrence position of the delivery action in the monitoring area, and referring to the identification result of the belonging detection means, the delivery detection means for obtaining the type of the article delivered in the delivery action, and in the monitoring area A storage unit that stores monitoring reference information that associates a monitoring importance level that represents a degree of interest as a monitoring target for each combination of a position and an article type; With reference to the monitoring standard information, for each delivery action, monitoring importance calculation means for obtaining the monitoring importance corresponding to the combination of the generation position and the type of the delivered article, and monitoring importance of the delivery action According to the present invention, there is provided an image monitoring system comprising output control means for outputting a detection result of the delivery action.

  According to the present invention, it is possible to efficiently monitor a delivery action at a place where an unspecified number of people gather. For this reason, it becomes possible to deal with an abnormal situation occurring in the monitoring place at an early stage triggered by the delivery action.

It is the figure which showed the whole structure of the image monitoring system in this Embodiment. It is a functional block diagram of the image monitoring apparatus in this Embodiment. It is the figure which showed the article information in this Embodiment. It is the figure which showed the example of a setting of the block and area of a monitoring area | region. It is the figure which showed the example of a setting of the block and area of a monitoring area | region. It is the figure which showed the example of a setting of the block and area of a monitoring area | region. It is the figure which showed the example of a setting of the block and area of a monitoring area | region. It is the figure which showed the area information in this Embodiment. It is the figure which showed the monitoring reference | standard information in this Embodiment. It is the figure which showed typically the example of a detection of the delivery action in this embodiment. It is the figure which showed the mode of the monitoring area | region in this Embodiment. It is the figure which showed the example of a display of the display part in this embodiment. It is the figure which showed the example of a display of the display part in this embodiment. It is the figure which showed the example of a display of the display part in this embodiment. It is a flowchart of the image monitoring process in this Embodiment. It is a flowchart of the delivery detection process in this Embodiment. It is the figure which showed the example of a setting of the block of a monitoring area | region, and a frequency calculation area | region. It is the figure which showed typically the generation | occurrence | production frequency area | region and the generation | occurrence | production position of the delivery action. It is the figure which showed typically the position of the pursuit person and the generation | occurrence | production position of the delivery action. It is the figure which showed typically the generation | occurrence | production position of a delivery act, and the distance to an important area. It is the figure which showed typically the generation | occurrence | production position of the delivery act, and the position of the monitoring person.

  Embodiments of an image monitoring system according to the present invention will be described below. In this embodiment, an example of an image monitoring system that detects an event venue where an unspecified number of persons come and go and a surrounding area of the event site as a monitoring location and detects a delivery of an article carried in the monitoring location will be described. In this embodiment, the monitoring center's monitor confirms in detail whether or not the delivery action detected by the image monitoring system is suspicious based on the video, etc. Appropriate measures are taken, such as giving instructions to employees.

  FIG. 1 is a diagram showing an overall configuration of an image monitoring system according to the present embodiment. As shown in FIG. 1, an image monitoring system 1 includes a plurality of imaging devices 2 installed at various locations of a monitoring location, an image monitoring device 3 that analyzes an image of the monitoring location, and a monitoring center device installed at a monitoring center. 4 is configured. Each of these devices is connected by wired communication or wireless communication, and inputs and outputs various types of information.

  Specifically, the images of the monitoring locations sequentially captured by the imaging device 2 are sequentially acquired by the image input unit 31 of the image monitoring device 3 and input to the storage unit 32 and the image processing unit 33 of the image monitoring device 3. The The image processing unit 33 processes the input image and detects a delivery action from the image. The image monitoring device 3 outputs information about the detected delivery action (hereinafter, delivery information) to the monitoring center device 4. The monitoring center device 4 displays the delivery information input from the image monitoring device 3 on the display unit 41. Here, the delivery information includes an image showing the state of delivery, the time of occurrence of the delivery action, the type of the place where the delivery action occurs, the type of the delivered article, the image of the delivered article, and the like. Further, the monitoring center device 4 stores the transfer information input from the image monitoring device 3 in the recording unit 42. The delivery information recorded in the recording unit 42 is used when re-verifying whether the delivery act is suspicious, or used as evidence information when a criminal act occurs.

  In addition, the image monitoring system 1 calculates a monitoring importance level that represents a degree to which the monitor should pay attention as a monitoring target for each detected delivery action. The monitoring importance is handled as one of the above-described delivery information. Then, the image monitoring system 1 changes the display method on the display unit 41 in accordance with the monitoring importance so that the monitoring staff at the monitoring center can preferentially confirm the delivery action with high monitoring importance. And control whether or not the delivery information is output. As a result, even if there are a large number of unspecified persons such as event venues, and monitoring a place where a large number of delivery actions occur, the monitoring center's observers can perform a highly suspicious delivery action, It is possible to efficiently monitor the delivery action while easily distinguishing the delivery action with a low suspicious degree.

  Next, the imaging device 2, the image monitoring device 3, and the monitoring center device 4 constituting the image monitoring system 1 of the present embodiment will be described in detail.

  The imaging device 2 is a monitoring camera configured to include an imaging element such as a CCD element or a C-MOS element, an optical system component, an analog / digital converter, and the like. The imaging device 2 sequentially captures a predetermined monitoring area at a monitoring location at predetermined time intervals, and inputs the captured digital image to the image monitoring device 3. In the present embodiment, the imaging device 2 captures the monitoring area at 5 fps and generates a visible color image. Here, the imaging device 2 acquires the current time at the time of shooting, and adds information indicating the shooting time to the shot image. Further, in the present embodiment, since it is necessary to photograph a wide range of monitoring places such as the event venue and the vicinity of the event venue, a plurality of imaging devices 2 are installed at various locations of the monitoring location. For this reason, the imaging device 2 is managed by a camera ID for identifying each, and a camera ID is also added to an image captured by the imaging device 2. The imaging device 2 is installed at an installation position, an imaging direction, an installation height, and a depression angle that are suitable for imaging an image feature of a person existing in the monitoring area and an article possessed by the person. Information such as the installation height and depression angle of the imaging device 2 is stored in the storage unit 32.

  In the present embodiment, a plurality of imaging devices 2 are used. However, the present invention is not limited to this, and the image monitoring system 1 may be realized by a single imaging device 2. Further, the imaging device 2 is not limited to being fixedly installed in the monitoring place, and the monitoring area may be photographed using a mobile imaging device such as an airship or a robot having a photographing function.

  Next, the image monitoring apparatus 3 will be described in detail with reference to FIG. FIG. 2 is a functional block diagram of the image monitoring apparatus 3 in the present embodiment. As shown in FIG. 2, the image monitoring device 3 includes an image input unit 31, a storage unit 32, and an image processing unit 33.

  The image input unit 31 is an interface that sequentially acquires images taken by the imaging device 2 and sequentially outputs the images to the storage unit 32 and the image processing unit 33. Hereinafter, the image acquired by the image input unit 31 is referred to as an input image. The image input unit 31 does not directly output the input image acquired from the photographing apparatus 2 to the image processing unit 33, but reads the input image from the storage unit 32 in accordance with the processing timing of the image processing unit 33, and You may make it output to the process part 33. FIG.

  The storage unit 32 includes a semiconductor memory such as a ROM (Read Only Memory) and a RAM (Random Access Memory), and a memory device such as a hard disk. The storage unit 32 is accessible from the image input unit 31, the image processing unit 33, and the like. Main information stored in the storage unit 32 is article information 321, area information 322, and monitoring reference information 323. Although not shown, the storage unit 32 stores the input image input from the image input unit 31. In addition to these pieces of information, the storage unit 32 stores programs and various data necessary for executing each process of the image monitoring apparatus 3.

  Here, the article information 321, the area information 322, and the monitoring reference information 323 stored in the storage unit 32 will be described in detail.

  The article information 321 is information used to identify the type of article existing in the input image. FIG. 3 is a diagram showing article information 321 in the present embodiment. As shown in FIG. 3, in the present embodiment, an article name of an article to be detected, an article ID for identifying the article, and an article image (hereinafter, article image) are associated with each other and stored as article information 321. Here, for an article having various types and shapes such as “bag”, a plurality of article images are associated with one article name. In the present embodiment, an article image is stored as the article information 321. However, the present invention is not limited to this. For example, instead of storing the article image itself, a feature amount extracted from the article image in advance may be stored. In addition, the article information 321 to be stored may be appropriately determined in accordance with the identification method of the belongings detection unit 332 described later. In this embodiment, the article name and the article ID are stored as information indicating the type of the article, but either the article name or the article ID may be stored. Further, the article to be detected is not limited to the present embodiment, and may be appropriately determined according to the monitoring purpose.

  The area information 322 is information used to identify the type of place where the delivery action occurs. In the present embodiment, the monitoring place is divided into a plurality of areas in advance, and the area information 322 is used to determine in which area the detected delivery action has occurred. Specifically, in the present embodiment, the monitoring area captured by each imaging device 2 is divided into a plurality of blocks, and an area is set for each block. Then, the area name, the area ID for identifying the area, and the position of the block belonging to the area are associated with each camera ID and stored as area information 322. In the present embodiment, the area name and the area ID are stored as information indicating the type of the occurrence location, but either the area name or the area ID may be stored.

  Here, the area information 322 in the present embodiment will be described in detail with reference to FIGS. 4, 5, 6, 7, and 8. 4 to 7 are diagrams showing an example of setting of the block and area of the monitoring area in the present embodiment. As shown in FIGS. 4 to 7, the monitoring area captured by each imaging device 2 is divided into a plurality of blocks. In the present embodiment, the size of one block is set to an actual distance of 2.5 m × 2.5 m, and the monitoring area is divided into a plurality of blocks as shown in the figure. If the size of one monitoring area is 30 m × 30 m, the monitoring area is divided into 12 blocks in the X-axis direction and 12 blocks in the Y-axis direction, for a total of 144 blocks. Here, the position of each block in the area information 322 and the XY coordinates in the real space obtained from the image obtained by capturing the monitoring area with each imaging device 2 are associated with each other and stored in the storage unit 32 in advance. In the present embodiment, the XY coordinates in the real space are two-dimensional coordinates when the floor surface on the image is the XY plane. In this way, each block position in the area information 322 and the XY coordinates in the real space are stored in association with each other, so that the occurrence position of the delivery action detected from the input image (the XY coordinates in the real space). However, it is possible to determine which block of the area information 322 corresponds to. The area information 322 is not limited to the present embodiment. For example, in the present embodiment, the number of block divisions is the same regardless of the imaging device 2, but the number of block divisions may be changed for each imaging device 2. In the present embodiment, the blocks and areas in the monitoring area are set with two-dimensional information, but may be set with three-dimensional information.

  FIG. 4 is a diagram illustrating an area setting example of the imaging apparatus 2 to which the camera ID “100” is assigned. As shown in FIG. 4, the plain block is set to area ID “001”, the left hatched block is set to area ID “002”, and the shaded block is set to area ID “003”. FIG. 5 is a diagram illustrating an area setting example of the imaging device 2 to which the camera ID “200” is assigned. As shown in FIG. 5, it can be seen that all the monitoring areas of the imaging apparatus 2 to which the camera ID “200” is assigned are set to the area ID “001”. FIG. 6 is a diagram illustrating an area setting example of the imaging apparatus 2 to which the camera ID “300” is assigned. As shown in FIG. 6, the block with the right diagonal line is set to the area ID “004”, and all the monitoring areas of the imaging device 2 to which the camera ID “300” is assigned are set to the area ID “004”. I understand that. FIG. 7 is a diagram illustrating an area setting example of the imaging apparatus 2 to which the camera ID “400” is assigned. As shown in FIG. 7, the plain block is set to area ID “001”, and the right hatched block is set to area ID “004”.

  FIG. 8 is a diagram showing the area information 322 in the present embodiment. As illustrated in FIG. 8, the area information 322 is associated with an area ID, an area name, and a block position for each camera ID. In this embodiment, the place where the delivery act occurs is the square around the event venue, the entrance gate where the visitor passes the security check when entering the event venue, the visitor who has not received the security check The boundaries between the square and the entrance gate, which are partitioned by fences to restrict entry into the event venue, and the event venue are set in different areas. That is, the imaging device 2 to which the camera ID “100” is assigned images the square and the entrance gate as a monitoring area, and the imaging device 2 to which the camera ID “200” is assigned takes an image of only the square as the monitoring area. The imaging device 2 to which the camera ID “300” is assigned images only the inside of the event venue as a monitoring region, and the imaging device 2 to which the camera ID “400” is assigned monitors the inside of the event venue and the square. As can be seen from FIG. Note that the type of place set as an area is not limited to this embodiment, and may be determined as appropriate according to the purpose of monitoring.

  Further, as shown in FIG. 8, the importance of the area is set for each area in the area information 322. Specifically, in this embodiment, there are two types of areas, a normal area and an important area, according to the importance of guarding. For example, in this embodiment, a square, an entrance gate, and an entrance gate boundary are set as “normal areas”, and an event venue is set as an important area.

  The monitoring reference information 323 is information used for calculating the monitoring importance level of the delivery action. As described above, the monitoring importance is an index that represents the degree to which the monitor should pay attention as the monitoring target. In this embodiment, the monitoring importance of the delivery action is obtained based on the delivered article and the place where the delivery action occurs. For this reason, in this embodiment, the value of the monitoring importance according to the combination of the type of article and the area is set in advance and stored as the monitoring reference information 323.

  FIG. 9 is a diagram showing the monitoring reference information 323 in the present embodiment. As shown in FIG. 9, in the monitoring reference information 323, the item ID, the item name, the area ID, the area name, and the monitoring importance are associated with each monitoring reference. Here, the monitoring importance level indicates that the higher the value, the higher the degree of attention that the monitoring staff should pay attention to as the monitoring target. Note that the combination of the type of article and the type of location of occurrence is not limited to the present embodiment, and may be appropriately determined according to the purpose of monitoring. In this embodiment, the monitoring importance is represented by a numerical value, but may be represented by a category such as “high”, “medium”, and “low”.

  Next, returning to FIG. 2, the image processing unit 33 of the image monitoring apparatus 3 will be described in detail.

  The image processing unit 33 includes an arithmetic device such as a CPU (Central Processing Unit), a DSP (Digital Signal Processor), or an MCU (Micro Control Unit), and reads and executes various programs stored in the storage unit 32. The image processing unit 33 receives input images from the image input unit 31 and sequentially processes them. Here, the image processing unit 33 refers to the camera ID and the shooting time added to the input image, and processes the images in the order of shooting for each imaging device 2.

  The image processing unit 33 according to the present embodiment includes a person detection unit 331, a belonging detection unit 332, a delivery detection unit 333, a monitoring importance level calculation unit 334, and an output control unit 335. Hereinafter, each unit constituting the image processing unit 33 will be described.

  The person detection unit 331 detects a person existing in the monitoring area from the input image, and extracts an image of the person (hereinafter, a person image). In the present embodiment, the person detection unit 331 performs background difference processing on the input image, detects a person area, cuts out an image in the person area from the input image, and extracts a person image. For this reason, although not shown, a background image obtained by photographing the monitoring area in a state where no person is shown is prepared for each imaging device 2 and stored in the storage unit 32. When a person existing in the monitoring area possesses an article, the person and the possessed article appear as an integrated image on the input image. For this reason, the person image extracted by the person detection means 331 is extracted as an image including the image of the possessed article. In the present embodiment, the person image is extracted by the background difference process, but the person image extraction method is not limited to this, and various existing methods may be adopted. For example, a person image may be extracted from an input image by a learning discriminator that has learned a person image. In addition, the shape and brightness characteristics of a person may be stored in advance in the storage unit 32 as a person template, and a person image may be extracted from the input image by matching processing with the person template.

  Further, the person detecting means 331 in the present embodiment tracks a person existing in the monitoring area. Specifically, the person detection unit 331 associates a person image in the input image processed this time with a person image that can be determined to be the same person among the person images in the input image processed last time. That is, the person detection unit 331 realizes person tracking by associating person images of the same person with input images taken at different times. Here, the position, color, size, shape, texture, and the like of the person image may be used for the association of the person images. Note that the person detection unit 331 may associate person images not only between images captured by the same imaging device 2 but also between images captured by the imaging devices 2 having different camera IDs. In other words, the person detection unit 331 continues to track a person even when a person who is out of the monitoring area of a certain imaging apparatus 2 moves to a monitoring area of another imaging apparatus 2 adjacent to the monitoring area of the imaging apparatus 2. Like that. In order to realize such tracking, the adjacent relationship between the monitoring areas captured by each imaging device 2 is stored in advance in the storage unit 32 in association with the camera ID, and the person detection unit 331 uses this information. The person image may be associated with reference to FIG.

  In addition, when there is a person image that is not associated with the person image in the previously processed input image, the person detection unit 331 starts tracking the person image as a person image of a person who has newly appeared in the monitoring location. . In addition, the person detection unit 331 determines that the person of the person image that has been tracked up to the previous time has left the monitoring location when there is no person image associated with the current input image. The tracking for the person is finished.

  In addition, the person detection unit 331 obtains the position of the person in the monitoring area. In the present embodiment, the person detection means 331 is based on information such as the position (pixel position) of the person image on the input image, the size of the person image, the installation height of the image pickup apparatus 2 and the depression angle, and the like. Find the coordinates. In the present embodiment, the foot coordinates are set as the position of the person in the monitoring area. Here, the foot coordinates are obtained as XY coordinates in real space.

  The person detection means 331 assigns a person ID to each person being tracked (hereinafter referred to as a tracking person). The person detection unit 331 associates the person ID, the time-series person image of the person to be tracked, the shooting time of the person image, and the position of the person in time series with each other in the storage unit 32 for each person to be tracked. Remember. Hereinafter, this information is referred to as tracking person information.

  Note that the person tracking method and the person position calculation method by the person detection unit 331 are not limited to the present embodiment, and an existing tracking technique or position calculation technique may be employed. Further, in the present embodiment, the foot coordinates are obtained as XY coordinates in the real space as the position of the person. However, the present invention is not limited to this, and a person who can specify the position of the person in the monitoring place is appropriately adopted. Good.

  Next, the belonging detection unit 332 will be described. The belongings detection means 332 detects an article possessed by a person and identifies the type of the article. In the present embodiment, the belongings detection unit 332 determines whether the image of the possessed article is included in the person image extracted by the person detection unit 331. If the image of the possessed article is included, the type of the possessed article is identified. Specifically, the belongings detection unit 332 refers to the article information 321 stored in the storage unit 32 and collates the article image and the person image of the article information 321. Then, when a region similar in characteristics to the article image is detected in the person image, it is determined that the person of the person image possesses the article. Hereinafter, an area having characteristics similar to those of the article image is referred to as a belonging area. The belonging detection unit 332 refers to the article information 321 and reads the article ID and the article name of the article image having the most similar characteristics for each belonging area, thereby identifying the type of article possessed by the person.

  It should be noted that existing feature amounts such as luminance, color histogram, HOG (Histograms of Oriented Gradients) feature amount, LBP (Local Binary Patterns) feature amount, etc. may be adopted as the feature amount for detecting the belonging area.

  The belonging detection means 332 stores the article ID and the article name of the article possessed by the person in the tracking person information of the person who is determined to possess the article. When a person has a plurality of articles, a plurality of belonging areas are detected in the person image of the person. In this case, the belonging detection means 332 identifies the article ID and the article name for each belonging area and stores them in the tracking person information. At this time, it is conceivable that a person has a plurality of articles of the same type, and in this embodiment, a belonging ID is assigned to each article possessed by the tracking person. For example, when a person possesses two bags, two belonging areas are detected from the person's image of the person, and the same article ID “1” and article name “ “Bag” is identified. However, it is possible to distinguish the two bags by the belonging ID assigned to each possessed article. The belonging ID is given as “1” and “2” in the order of detection at the timing when the possessed area of the new possessed article is detected from the person. Thereafter, when the possessed items in the possessed area are the same, the assigned possessed ID is taken over.

  Also, the belonging detection unit 332 determines the appearance of the possessed article and the disappearance of the possessed article with respect to the person. Specifically, when a new possessed article is detected from the person image of the tracked person, the belonging detection unit 332 sets an appearance flag for the person image of the tracked person. Further, when the possessed article that has been detected up to the previous time is not detected from the current person image, a disappearance flag is set for the person image. Here, when the possessed article same as the possessed article in the person image for which the appearance flag was set last time is also detected from the current person image, the belonging detection means 332 sets the appearance flag for the person image. Absent. When determining whether the previous possessed article and the current possessed article are the same article, it is determined whether the article ID is the same, or the image features in each possessed area are similar. It may be determined whether or not to do so.

  That is, the belonging detection unit 332 adds information about the possessed article to the traced person information stored in the storage unit 32. Specifically, the belonging detection unit 332 performs the possession area for each possession ID, the presence / absence of the appearance flag for each possession ID, and the disappearance flag for the time-series human image stored in the tracking person information. Are stored in association with each other. In addition, the belonging detection unit 332 stores an article ID and an article name for each belonging ID in association with the person ID stored in the tracking person information.

  The possessed article detection method and the article type identification method by the possessed article detection means 332 are not limited to the present embodiment, and various existing methods may be adopted. For example, the belongings detection unit 332 may identify the type of article (article name or article ID) by a learning discriminator that has learned the article image. In the present embodiment, the belonging area is detected from the entire person image, but the present invention is not limited to this. For example, the belonging detection unit 332 may compare a past person image with a current person image, and detect a belonging area from areas where the color or texture of the person image has changed. Also, instead of detecting the belonging area from the person image as in the present embodiment, the belonging area is detected from the entire input image, and the position of the belonging area is obtained. You may make it match with a personal belonging area | region with the person image of the person who exists in the near position.

  Next, the delivery detection unit 333 will be described. The delivery detection means 333 detects a delivery action performed within the monitoring area. In the present embodiment, the delivery detection means 333 causes a new possessed article to appear from a certain tracked person within a predetermined time, and the possessed article disappears from another person existing in the vicinity of the person. When the article and the lost possessed article are the same article, it is determined that the possessed article has been delivered between these persons. Specifically, the delivery detection unit 333 refers to the tracking person information of the tracking person, and sets a pair of a person with an appearance flag and a person with a disappearance flag for the tracking persons with different person IDs. . In the present embodiment, a pair of a person whose appearance flag is set in the current process and a person whose disappearance flag is set within the past predetermined time is set. Here, a person whose disappearance flag has been set in the past predetermined time is targeted for pairing, not only when handing over directly by hand, but also when the handing person receives the goods once placed on the ground. This is so that, when the handed-over person picks up and delivers, and the handed-over person throws the possessed article to the handed-over person, it can be detected as a handing-over action. For this reason, it is preferable that the predetermined time is set to a time until the possessed article moves from the person on the delivery side to the person on the delivery side in consideration of these delivery methods. In the present embodiment, a person who has possessed an article and a person who has lost the possessed article within a predetermined time in the past are set as a pair. However, the present invention is not limited to this. For example, a pair of a person whose possessed article has disappeared and a person whose possessed article has appeared between the time when the person disappeared and the predetermined time has elapsed may be set.

  The delivery detection means 333 determines whether the possessed article that has appeared and the possessed article that has disappeared are the same article for each set pair. Here, when determining whether or not they are the same article, the article name and article ID associated with the belonging ID for which the appearance flag is set and the belonging ID for which the disappearance flag is set are associated. It is determined whether or not the article name and article ID are the same, and the image feature in the possessed area at the time when the appearance flag is set and the belonging ID with the disappearance flag set are given, What is necessary is just to determine whether the image characteristic in the possession area | region before the said disappearance is similar. In addition, in the set pair, when there are a plurality of possession IDs in which a disappearance flag is set for one tracking person, the identity of the articles may be determined for the possession areas of the respective possession IDs. Further, even when there are a plurality of belonging IDs in which the appearance flag is set for one tracking person in the set pair, it is only necessary to determine the identity of the articles for the belonging areas of the respective belonging IDs.

  Moreover, the delivery detection means 333 calculates | requires the distance between tracking persons for each set pair. In the present embodiment, the distance between tracked persons refers to the time-series position of the person stored as tracked person information, and the position of the person when the appearance flag is set and the disappearance flag are set. Obtained based on the position of the person at the time. In addition, in the set pair, when there are a plurality of possession IDs for which the disappearance flag is set for one tracking person, the disappearance flag for the possession ID for which it is determined that the possessed items are the same is set. What is necessary is just to obtain | require the distance between persons using the position of a person. In addition, in the set pair, when there are a plurality of belonging IDs for which an appearance flag is set for one tracking person, an appearance flag for the belonging ID that is determined to be the same for the possessed article is set. What is necessary is just to obtain | require the distance between persons using the position of a person.

  The delivery detection unit 333 extracts pairs having the same possessed article and a distance between persons within a predetermined distance from the set pairs, and sets these pairs as “delivery pairs”. That is, the delivery detection unit 333 determines that a delivery action has occurred in the monitoring area when a delivery pair is extracted in the current process. Here, the predetermined distance takes into consideration not only the distance between persons when handing directly by hand, but also the distance between persons when handing goods to the person to whom the handing person is handed over It is preferably set to a value.

  Moreover, the delivery detection means 333 calculates | requires the kind of delivered goods, and the generation | occurrence | production position of delivery action. In the present embodiment, the delivery detection unit 333 includes, for the delivery pair, the article name of the possessed item ID associated with the possessed article area in which it is determined that the possessed article that has appeared and the possessed article that has disappeared are the same article. The article ID is determined as the type of article that has been delivered. In addition, the delivery detection unit 333 determines, with respect to the delivery pair, the “position of the person at the time when the appearance flag is set” and the “person at the time when the disappearance flag is set, which are determined that the distance between the persons is within a predetermined distance. The middle point is determined for “the position of”. And the delivery detection means 333 calculates | requires the XY coordinate in the real space of the said midpoint, and makes it the generation | occurrence | production position of a delivery act. In the present embodiment, the midpoint between the positions of the persons is set as the occurrence position of the delivery action, but the present invention is not limited to this, and the optimum delivery position is appropriately determined according to the monitoring purpose. Adopt it. For example, it is good also as a generation | occurrence | production position of the delivery action of either one person in a delivery pair. Further, the position of the belonging area may be obtained, and the position of the belonging area may be transferred and used as the generation position of the action. In this case, like the position of the person, the position of the possessed area may be obtained by XY coordinates in the real space so that the position of the possessed article at the monitoring place can be specified.

  Here, with reference to FIG. 10, the detection process of the delivery action in this Embodiment is demonstrated. FIG. 10 is a diagram schematically illustrating a detection example of a delivery action in the present embodiment. FIG. 10 is an enlarged view of a part of the input image for simplification of description, and the time series is shown in the order of (a), (b), and (c).

  First, FIG. 10A shows two persons, person 5 and person 6. The person 5 has a bag 7 and the person 6 does not have an article at the present time. That is, the person image of the person 5 is detected as an image in which the person 5 and the bag 7 are integrated. In addition, the black circle 51 and the black circle 61 shown in the figure have shown the foot position of the person 5 and the person 6, respectively. When the belonging detection unit 332 performs a belonging area detection process on the person image of the person 5 and the person image of the person 6, the belonging area is not detected from the person image of the person 6. The bag 7 is detected as a belonging area from the person image. At this time, the tracking person information of the person 5 includes the person ID of the person 5, the person image of the person 5, the shooting time of the person image, the position of the person indicated by the black circle 51, the belonging area of the bag 7, the bag 7 The possession ID assigned to the possession area, the article ID and article name associated with the possession ID, and the appearance flag associated with the person image of the person 5 are stored. Further, the tracking person information of the person 6 stores the person ID of the person 6, the person image of the person 6, the shooting time of the person image, and the person position indicated by the black circle 61.

  FIG. 10B shows the person 5 handing over the bag 7 to the person 6. Even at this time, the person image of the person 5 is detected as an image in which the person 5 and the bag 7 are integrated. In addition, the black circle 51 and the black circle 61 have shown the foot position of the person 5 and the person 6 at this time. When the belonging detection unit 332 performs a belonging area detection process on the person image of the person 5 and the person image of the person 6, the belonging area is not detected from the person image of the person 6. From the person image, the bag 7 is again detected as a belonging area. Here, the article ID of the belonging area and the image feature in the belonging area detected at this time by the belonging detection means 332 are the article ID and image of the belonging area of the person 5 at the time of FIG. It is compared with the characteristic and determined to be the same article. Accordingly, the tracking person information of the person 5 is associated with the person ID of the person 5, the person image of the person 5, the shooting time of the person image, and the person indicated by the black circle 51 at the time of FIG. And the belonging area of the bag 7 are stored in association with the same belonging ID as at the time of FIG. Further, in the tracking person information of the person 6, the person image of the person 6 at the time of FIG. 10B, the shooting time of the person image, and the person indicated by the black circle 61 are associated with the person ID of the person 6. The position is associated and stored.

  FIG. 10C shows a state in which the bag 7 is transferred from the person 5 to the person 6. That is, at this time, the person image of the person 6 is detected as an image in which the person 6 and the bag 7 are integrated. In addition, the black circle 51 and the black circle 61 have shown the foot position of the person 5 and the person 6 at this time. When the belonging detection unit 332 performs the belonging area detection process on the person image of the person 5 and the person image of the person 6, the belonging area is not detected from the person image of the person 5. The bag 7 is newly detected as a belonging area from the person image. Accordingly, the tracking person information of the person 5 is associated with the person ID of the person 5, the person image of the person 5, the shooting time of the person image, and the person indicated by the black circle 51 at the time of FIG. Is stored. Furthermore, since the possessed area of the possessed article that was detected last time is no longer detected from the person image of the person 5, the disappearance flag for the possessed article ID of the possessed article area is displayed in the person image at the time of FIG. Correspondingly stored. Further, the tracking person information of the person 6 is associated with the person ID of the person 6 and the person image of the person 6 at the time of FIG. Corresponding to the position of the bag 7, the belonging area of the bag 7, the belonging ID assigned to the belonging area of the bag 7, the article ID and the article name associated with the belonging ID, and the person image of the person 5 The added appearance flag is stored.

  Then, the delivery detection unit 333 sets the person 6 with the appearance flag raised and the person 5 with the disappearance flag raised within the past predetermined time as a pair at the time of FIG. In order to simplify the description, the description of the pair determination at the time of FIGS. 10A and 10B is omitted.

  Based on the tracking person information of the person 5 and the person 6, the delivery detection means 333 confirms that the possessed item is the same bag 7, the position of the person 5 (black circle 51) and the person 6 at the time of 10 (c). If it is determined that the distance to the position (black circle 61) is within a predetermined distance, the person 5 and the person 6 are detected as a “delivery pair”. Then, the article name of the possessed item ID and the article ID associated with the possessed article region for which the identity of the possessed article is determined is the type of the article that has been delivered. That is, the types of objects passed between the person 5 and the person 6 are “article ID: 1” and “bag”. Further, the midpoint (double circle 8) of the black circle 51 and the black circle 61 at the time of FIG. 10C used for the calculation of the distance is determined as the delivery occurrence position.

  The delivery detection unit 333 stores the type of delivered article and the delivery occurrence position in the storage unit 32 in association with each delivery pair. Then, the delivery detection unit 333 outputs these pieces of information to the monitoring importance calculation unit 334 described later.

  In the present embodiment, the delivery detection unit 333 detects the delivery action based on the appearance and disappearance of the possessed article by the tracking process, but the present invention is not limited to this. For example, the image feature of the person who is delivering the possessed article from the input image may be detected by image analysis, and the delivery action may be detected. Specifically, without using the tracking process, the person detection unit 331 detects a person in the input image, and the belongings detection unit 332 detects an article in the input image. Then, the delivery detection means 333 may determine that the delivery action is performed when two persons are positioned so as to sandwich the article as shown in FIGS. 10B and 10C. Further, as shown in FIG. 10 (b), a pattern having a shape composed of a possessed article, a person who delivers the possessed article, and a person who delivers the possessed article is stored in advance, and delivery detection is performed. The means 333 may determine that the action is a delivery action when a region matching the pattern is detected from the input image. In the above example, the delivery detection unit 333 detects the delivery action from the positional relationship between the plurality of persons and the article, but can also detect the delivery action by another method. In other words, the delivery detection unit 333 may detect the delivery action using a learning discriminator that has learned in advance an image showing a delivery action between persons. In addition, the posture of a person in the input image is estimated using an existing posture estimation technique, and the transfer detection unit 333 transfers the transfer when a posture of passing an object or a posture of receiving an object is detected. You may make it determine with an action. In the case of these detection methods, the position of the article and the type of the article may be determined by the belonging detection means 322 from the input image after detecting the delivery action. In addition, an appropriate position may be adopted as the delivery position. For example, the position where the delivery act is generated may be obtained by the above-described method, or the position of the center of gravity of the region where the pattern such as the shape is matched, the detection position by the learning discriminator, or the like may be used.

  Next, the monitoring importance calculation means 334 will be described. The monitoring importance calculation means 334 calculates the monitoring importance of the delivery action. In the present embodiment, the monitoring importance level calculation means 334, for each “delivery pair” detected by the delivery detection means 333, based on the type of the article delivered in the delivery action and the delivery occurrence position, Calculate the monitoring importance. Specifically, the monitoring importance calculation means 334 first determines which block in the monitoring area corresponds to the position where the delivery action occurs. Here, the block may be determined based on the correspondence relationship between the XY coordinates and the block position in the real space of the monitoring area for each imaging device 2 stored in the storage unit 32. Then, the monitoring importance level calculation means 334 refers to the area information 322 and determines to which area the block at the position where the delivery action occurs belongs. Thereafter, the monitoring importance calculation means 334 refers to the monitoring reference information 323 stored in the storage unit 32, and determines the combination of the article type and area (occurrence location type) determined by the delivery action detection means 333. Read the corresponding monitoring importance. Then, the read monitoring importance is calculated as the monitoring importance of the delivery action. In the present embodiment, the monitoring importance is set to 0 when there is no combination of the article type and the generation location type corresponding to the monitoring reference information 323.

  The monitoring importance level calculation means 334 stores the type of the delivered article, the location where the delivery action occurs, the type of the location where the delivery occurs, and the monitoring importance level in association with each delivery pair in the storage unit 32.

  With reference to FIG. 11, the monitoring importance calculation processing in the present embodiment will be described. FIG. 11 shows the state of the monitoring area captured by the imaging device 2 with the camera ID “100” in the present embodiment. The imaging device 2 captures the area around the event venue, and the monitoring area is an area indicated by a dotted line in the figure. The monitoring area includes a square around the event venue and an entrance gate attached to the square, and there are many visitors who have visited the event venue. Visitors who visit the event venue pass through the square and head to the entrance gate partitioned by the fence 9. The fence 9 is provided with two entrance gate entrances 10. At the entrance 10, an event management staff 11 performs security checks on visitors. During security checks, visitors are checked for entrance tickets and checked baggage. Visitors who pass the security check pass through the entrance gate and head into the event venue.

  The monitoring area captured by the imaging device 2 with the camera ID “100” corresponds to the block or area shown in FIG. 4. In FIG. 11, the area set as the monitoring area is indicated by a dotted line. ing. As shown in FIG. 11, the area set in the monitoring area includes an area with an area ID “001” that is a square around the event venue, an area with an area ID “002” that is in the entrance gate, and a fence. 9 is composed of an area having an area ID “003”, which is a boundary portion between a square set to enclose 9 and an entrance gate.

  In the monitoring area shown in FIG. 11, there are a large number of visitors and event management staff 11, and among these persons, a large number of persons performing a delivery act are seen. When the transfer detection process by the transfer detection unit 334 is performed on the input image obtained by photographing the monitoring area, as shown in FIG. 11, 10 transfer actions A to J are detected.

  When the monitoring importance level calculation means 334 refers to the monitoring reference information 323 shown in FIG. 9 and the monitoring importance level is calculated for the delivery actions A to J, the result is as follows. First, since the delivery action A is delivering a bag in the open space, the monitoring importance is 7. Moreover, the delivery act B delivers the leaflet in the open space. Since this is a combination that is not registered in the monitoring reference information 323, the monitoring importance is zero. In the delivery action C, the camera is delivered in the open space. Since this is a combination that is not registered in the monitoring reference information 323, the monitoring importance is zero. Moreover, the delivery action D delivers the plastic bottle in the open space. Since this is a combination that is not registered in the monitoring reference information 323, the monitoring importance is zero. In addition, the delivery action E delivers a flyer at the boundary of the entrance gate. Since this is a combination that is not registered in the monitoring reference information 323, the monitoring importance is zero. In addition, the delivery action F delivers a ticket in the entrance gate. Since this is a combination that is not registered in the monitoring reference information 323, the monitoring importance is zero. In addition, since the delivery action G delivers a ticket at the boundary of the entrance gate, the monitoring importance level is 9. Moreover, since the delivery action H delivers the camera in the entrance gate, the monitoring importance level is 5. Moreover, the delivery action I delivers the bag in the entrance gate. Since this is a combination that is not registered in the monitoring reference information 323, the monitoring importance is zero. Moreover, since the delivery action J delivers the plastic bottle at the boundary of the entrance gate, the monitoring importance is 10. That is, in the monitoring area shown in FIG. 11, the monitoring importance increases in the order of the delivery action J, the delivery action G, the delivery action A, and the delivery action H.

  In the delivery action J, a person outside the entrance gate delivers a plastic bottle through the fence 9 to a person inside the entrance gate. That is, this is an act of bringing a plastic bottle into the entrance gate without passing the security check. Therefore, in the delivery action J, there is a possibility that the liquid in the PET bottle may be a dangerous substance, and therefore, the degree to which the monitoring personnel should pay attention as a monitoring target is high. On the other hand, the delivery action D is also a delivery of a plastic bottle. However, in the plaza, it is normal for visitors to hand over plastic bottles purchased at stores etc. to their companions. Even if it is dangerous, it is necessary to bring it into the event venue. You need to pass the security check at the entrance gate entrance. For this reason, the remarkable degree of the delivery action D is low at the present time.

  In the delivery action G, a person inside the entrance gate passes the entrance ticket to a person outside the entrance gate through the fence 9. This is an act in which a person who has entered the entrance gate after the ticket is confirmed by the security check passes his / her entrance ticket to a person who does not have an entrance ticket outside the entrance gate. That is, there is a possibility that a person who did not have an admission ticket will pass through the security check in the name of re-entry and enter the event venue. Therefore, such a suspicious behavior has a high degree of attention that should be noted by the monitor as a monitoring target. On the other hand, the delivery act F is also a delivery of the admission ticket. However, in the entrance gate, tickets are normally delivered between the visitors and the event management staff 11 for security checks. For this reason, the noticeable degree of the delivery action F becomes low.

  The delivery act A is an act of delivering a bag in the open space. Of course, visitors may hand over their bags for temporary storage by their companions. However, as described above, it is necessary to be wary of the possibility of passing dangerous materials such as explosives placed in bags between suspicious individuals. If the contents of the bag are dangerous goods, it is necessary to pass the security check at the entrance gate to bring them into the event venue, but if the contents of the bag are explosives, crimes are being done in the plaza etc. Will cause great damage. For this reason, since it is necessary to carefully confirm whether or not the delivery action A is suspicious, the degree to which the observer should pay attention as a monitoring target is increased. On the other hand, the delivery action I is also an action of handing over the bag. However, since the bag is handed over in the entrance gate after passing the security check, it is unlikely that the bag contains dangerous goods. For this reason, the remarkable degree of the delivery action I becomes low.

  The delivery act H is an act of delivering the camera in the entrance gate. There is a high possibility that the act of the visitor taking out the camera and delivering it to a third party is a preliminary operation for taking a picture. If the event venue is prohibited to shoot, the security check may confirm the camera you have, but the camera itself is not a dangerous item and you can shoot even if you bring the camera into the event venue. If there is no problem. For this reason, the camera staff is rarely confiscated, although the management staff may urge attention not to shoot. However, the delivery action H takes a preliminary operation for taking a picture after passing through the entrance gate. Therefore, after that, there is a high possibility that actual shooting will be performed. For this reason, about the delivery action H, the degree which should be watched as a monitoring object by the supervisor is increased. On the other hand, the delivery C also delivers the camera, but since the shooting in the open space is not prohibited, the degree of attention of the delivery act C is low.

  The delivery action B and the delivery action E are delivering flyers. The act of handing over leaflets and pamphlets is usually performed around the event venue. For this reason, the remarkable degree of the delivery action B becomes low. In addition, the delivery action E is a delivery action from a person inside the entrance gate to a person outside the entrance gate, but since the delivered article is a non-hazardous flyer, The degree of attention is low. Although not shown in FIG. 11, even if there is no danger in the handing over of the leaflet itself, if the handing over of the leaflet is performed in a place where the leaflet is prohibited from being distributed, You may make it raise the degree which should be noted as a monitoring object.

  As described above, the degree to which the observer should pay attention as the monitoring target varies depending on the combination of the delivered item and the place where the delivery act occurs, and is determined only by the delivered item. It is not determined only by the place where the delivery act occurs. In the image monitoring system 1 according to the present embodiment, the monitoring person should pay attention as the monitoring target by obtaining the monitoring importance according to the combination of the delivered article and the place where the delivery act is generated. The degree can be determined with high accuracy and in accordance with the purpose of monitoring.

  Next, the output control means 335 will be described. The output control means 335 controls the output method of the delivery information to the outside according to the monitoring importance calculated by the monitoring importance calculation means 334. In the present embodiment, the output control means 335 outputs the delivery information of the delivery action detected by the delivery detection means 333 to the monitoring center device 4 and displays the delivery information on the display unit 41 of the monitoring center 4. And stored in the recording unit 42. The output control means 335 of this embodiment controls to change the display method of the delivery information to the display unit 41 according to the monitoring importance. Specifically, the output control means 335 allows the monitoring person who is viewing the video displayed on the display unit 41 to preferentially confirm the delivery action with high monitoring importance over the delivery action with low monitoring importance. Thus, the display mode is changed. In addition, as described above, the delivery information of the present embodiment includes an image showing a delivery state, a delivery action occurrence time, a delivery action occurrence position, a type of a delivery action occurrence place, a delivery article The type, the image of the delivered item, the monitoring importance level, etc. In the present embodiment, an image at the current time in which the persons who have performed the transfer action are also handled as the transfer information. Note that the delivery information refers to the tracking person information stored in the storage unit 32 for the delivery pair, the type of the delivered article, the delivery occurrence position, the delivery occurrence location type, and the like. It is generated as appropriate using image information such as a time-series input image stored in 32.

  With reference to FIG. 12, a display mode on display unit 41 in the present embodiment will be described. FIG. 12 is a diagram showing a display example of the display unit 41 in the present embodiment. In addition, in FIG. 12, the delivery information of the delivery action detected in the monitoring area | region shown in FIG. 11 is displayed. As shown in FIG. 12, the display unit 41 displays various delivery information related to the delivery action J. First, in the area 411, an image showing the state of delivery is displayed. The area 412 displays the time of occurrence of the delivery action, the type of place where the delivery action occurs, the type of the delivered article, and the monitoring importance. In the area 413, an image of the delivered article is displayed. Also, in the area 414 and the area 415, images at the current time of the persons who performed the delivery action are displayed.

  In the area 416 of FIG. 12, thumbnail images are listed for each delivery action detected by the delivery detection unit 333. When the monitoring person selects a thumbnail image of the delivery action whose details are to be confirmed from the list of thumbnail images, the display unit 41 displays delivery information regarding the selected delivery action in an area 411, an area 412, an area 413, an area. 414, displayed in area 415. In FIG. 12, the thumbnail image of the delivery action J is selected.

  The output control means 335 of the present embodiment displays the thumbnail images on the display unit 41 so that the display order of the thumbnail images is in descending order of monitoring importance. For example, in the area 416 in FIG. 12, thumbnail images are displayed in the order of the delivery action J, the delivery action G, the delivery action A, and the delivery action H from the top. Further, below the thumbnail image of the delivery action H, a thumbnail image of the delivery action B having a monitoring importance of 0 is displayed. Although not shown in FIG. 12, the thumbnail image of the delivery action B having the monitoring importance level of 0 as in the delivery action B is followed by the thumbnail image of the delivery action B. The thumbnail images of these delivery actions are displayed when the monitor operates the arrow button 417 to scroll the thumbnail images up and down.

  As shown in FIG. 12, the list of thumbnail images of the delivery action is displayed on the display unit 41 in descending order of the monitoring importance, so that the monitor can monitor the importance of monitoring from among the detected many delivery actions. It is possible to easily grasp a delivery action with a high level and a delivery action with a low level of monitoring importance. In addition, since those with higher monitoring importance are always displayed on the list, the thumbnail image of the delivery action with higher monitoring importance becomes easier for the monitor to select. Thereby, even if it is a case where many delivery acts are detected from the monitoring area | region, the monitoring person can monitor a delivery act efficiently.

  In addition, the display mode according to the monitoring importance is not limited to this. Hereinafter, another display mode on the display unit 41 will be described with reference to FIGS. 13 and 14.

  FIG. 13 is a diagram showing another display example of the display unit 41 in the present embodiment. Note that FIG. 13 also displays the delivery information of the delivery action detected in the monitoring area shown in FIG. 11 as in FIG. In FIG. 13, an image showing the entire monitoring area shown in FIG. 11 is displayed, and the detected delivery action is displayed surrounded by a rectangular frame. At this time, the output control means 335 causes the display unit 41 to display so that the thickness of the rectangular frame changes according to the monitoring importance. Specifically, as illustrated in FIG. 13, the output control unit 335 displays the rectangular frame so that the thickness increases in the order of the delivery action J, the delivery action G, the delivery action A, and the delivery action H. Further, the output control means 335 may change the type of the rectangular frame according to the monitoring importance. For example, as shown in FIG. 13, the output control means 335 indicates the delivery action B, the delivery action C, the delivery action D, the delivery action E, the delivery action F, and the delivery action I, whose monitoring importance is 0, in the solid line. A rectangular frame is displayed with a dotted line. Further, the output control means 335 may display the value of the monitoring importance as a rectangular frame of the delivery action J. Further, the output control means 335 may display the color of the rectangular frame or the shape of the rectangular frame.

  As shown in FIG. 13, on the image showing the entire surveillance area, the position of the delivery action is indicated by a rectangular frame so that the supervisor can easily confirm the delivery action performed in the surveillance area. Can do. In addition, by displaying a delivery action with a high monitoring importance so that it is more prominent than a delivery action with a low monitoring importance, the supervisor can select a delivery action with a high monitoring importance from a number of detected delivery actions. And it is possible to easily grasp the delivery action with low monitoring importance. Thereby, even if it is a case where many delivery acts are detected from the monitoring area | region, the monitoring person can monitor a delivery act efficiently.

  FIG. 14 is a diagram showing another display example of the display unit 41 in the present embodiment. 14 also displays the delivery information of the delivery action detected in the monitoring area shown in FIG. 11, as in FIGS. In FIG. 14, an image in which the delivery action J is shown in the area 418 is displayed. In addition, an image in which the state of the delivery action G is shown in the area 419 is displayed. In addition, an image showing the delivery action A is displayed in the area 420. In addition, an image showing the delivery action H in the area 421 is displayed. As shown in FIG. 12, the output control means 335 according to the present embodiment causes the delivery information to be displayed in the forefront as the delivery action has a higher monitoring importance. In FIG. 12, an image showing the state of delivery is displayed for the delivery action detected in the monitoring area, but delivery action J, delivery action G, delivery action A, delivery action H overlap in this order. Is displayed. That is, the monitoring person can preferentially confirm a delivery action with a high monitoring importance level over a delivery action with a low monitoring importance level from among the detected many delivery actions. Thereby, even if it is a case where many delivery acts are detected from the monitoring area | region, the monitoring person can monitor a delivery act efficiently.

  As described above, various display modes on the display unit 41 have been described. However, the display mode is not limited thereto, and the display mode is changed so that the monitoring staff can efficiently monitor the delivery action according to the monitoring importance. Anything is acceptable. For example, although not shown in the figure, the detected occurrence position of the delivery action may be mapped on the map image of the monitoring place, and the higher the importance of monitoring may be displayed more conspicuously. At this time, when the monitoring person selects the mapped occurrence position, the delivery information may be displayed as shown in FIG. Moreover, you may make it display the movement locus | trajectory to the present time of the persons who performed delivery on a map image. At this time, the display mode of the movement trajectory may be changed according to the monitoring importance.

  The output control means 335 is not limited to changing the display mode on the display unit 41. For example, among the detected delivery actions, only those whose monitoring importance is greater than or equal to a predetermined value (for example, 5 or more) may be displayed on the display unit 41. For example, FIG. 14 shows only four pieces of delivery information of delivery action J, delivery action G, delivery action A, and delivery action H among the delivery actions detected in the monitoring area shown in FIG. In order to realize such a display, only the delivery information of the delivery action having a monitoring importance level equal to or higher than a predetermined value may be output to the monitoring center 4, or delivery of all detected delivery actions may be performed. After outputting the information to the monitoring center 4, only the delivery information of the delivery action with the monitoring importance level exceeding a predetermined value may be outputted to the display unit 41. That is, the output control means 335 in the present embodiment may control whether or not the delivery information is output to the outside according to the monitoring importance. As a result, the monitoring person can preferentially confirm the delivery action having a high monitoring importance level over the delivery action having a low monitoring importance level from among the detected many delivery actions. For this reason, even if a large number of delivery actions are detected from the monitoring area, the delivery actions can be efficiently monitored.

  The output control means 335 may output a notification display on the display unit 41 as delivery information. For example, when a delivery action with a monitoring importance level equal to or higher than a predetermined value (for example, 5 or higher) occurs, a notification display indicating that a delivery action with a high monitoring importance level has occurred is output to the display unit 41 of the monitoring center 4. You may do it. In this case, after confirming the notification display displayed on the display unit 41, the monitor checks the details by referring to the image information as described above for the delivery action having a high degree of monitoring importance. The notification display is not limited to the image display on the display unit 41, but may be a sound, an alarm sound, lighting of an alarm lamp, or the like. That is, the output control unit 335 may control the presence / absence of notification to the outside according to the monitoring importance. Further, the notification display mode may be changed according to the monitoring importance. As a result, the monitor can easily know that a delivery action having a high monitoring importance level has occurred at the monitoring location by the notification display, and can confirm it with priority over a delivery action having a low monitoring importance level. . For this reason, even if a large number of delivery actions are detected from the monitoring area, the delivery actions can be efficiently monitored.

  In the present embodiment, the output control unit 335 stores the delivery information in the recording unit 42 of the monitoring center device 4. At this time, the output control means 335 may control the presence / absence of output to the recording unit 42, that is, the presence / absence of recording to the recording unit 42, according to the monitoring importance. For example, the output control means 335 may record in the recording unit 42 only the delivery information of a delivery action having a monitoring importance level of a predetermined level or higher (for example, 3 or higher). Further, the output control means 335 records all the delivery action delivery information detected from the monitoring area in the recording unit 42, and the delivery action delivery information having a monitoring importance of a predetermined level or higher (for example, 5 or more). May be displayed on the display unit 41, or the threshold of monitoring importance for determining the presence / absence of recording and the threshold of monitoring importance for determining the presence / absence of display may be set to different values. . For example, delivery information with a monitoring importance level of 3 or higher may be recorded in the recording unit 42, and further, delivery information with a monitoring importance level of 5 or higher may be displayed on the display unit 41. As a result, the supervisor can confirm the delivery action having a high monitoring importance with a higher priority than the delivery action having a low monitoring importance, and can post the recorded delivery information for the delivery action having a low monitoring importance. It is possible to re-verify whether it is suspicious by confirming from For this reason, even if a large number of delivery actions are detected from the monitoring area, the delivery actions can be efficiently monitored.

  As described above, the output control unit 335 according to the present embodiment outputs the delivery information corresponding to the monitoring importance level for the delivery action detected in the monitoring area. As a result, even if there are a large number of unspecified persons such as event venues, and monitoring a place where a large number of delivery actions occur, the monitoring center's observers can perform a highly suspicious delivery action, It is possible to efficiently monitor the delivery action while easily distinguishing the delivery action with a low suspicious degree.

  In the present embodiment, the output control means 335 has been described as a function of the image processing unit 33 in the image monitoring apparatus 3, but the present invention is not limited to this. For example, the image monitoring system 1 may be realized by using the output control unit 335 as a function of the monitoring center device 4. In this case, the output control unit 335 may input information necessary for generating transfer information from the image monitoring device 33. Further, various image information of the delivery information displayed on the display unit 41 may be generated by the output control unit 335 and then output to the outside, or externally based on an instruction from the output control unit 335. You may make it produce | generate.

  Next, returning to FIG. 1, the monitoring center apparatus 4 constituting the image monitoring system 1 will be described.

  The monitoring center device 4 is a device installed in a monitoring center where a monitoring person is stationed. In the present embodiment, the monitoring center device 4 includes a display unit 41 and a recording unit 42. In the present embodiment, the monitoring center is provided at an event venue that is a monitoring location, and the monitoring staff can perform a delivery action that occurs in the monitoring location based on the delivery information input from the image monitoring device 3. Confirm in detail whether it is suspicious or not, and take appropriate measures such as issuing instructions to security personnel at the site. For this purpose, the monitoring center device 4 is connected to the image monitoring device 3 by wired communication or wireless communication, and inputs and outputs various types of information.

  The display unit 41 is an information display device such as a liquid crystal display, and displays delivery information and the like based on the signal from the output control means 335 described above. The display unit 41 may include a speaker or a lighting device such as an alarm lamp.

  The recording unit 42 includes a semiconductor device such as a ROM (Read Only Memory) and a RAM (Random Access Memory), and a memory device such as a hard disk, and records the delivery information based on the signal from the output control unit 335 described above. . In the present embodiment, the recording unit 42 is provided in the monitoring center device 4, but may be provided in the image monitoring device 3. Further, the recording unit 42 may be provided at a location away from the display unit 41.

  Note that the monitoring center may not be provided at the monitoring location, but may be provided at a remote location away from the monitoring location. In this case, the monitoring center device 4 may transmit necessary information to a security guard at the site via a communication network or the like, and output a handling instruction for a delivery action. Also, a plurality of monitoring centers may be provided such as a monitoring center provided in a remote place and a monitoring center provided in a monitoring place. That is, the image monitoring system 1 may be realized using a plurality of display units 41 and recording units 42.

  Next, processing of the image monitoring apparatus 3 in the present embodiment will be described with reference to FIG. FIG. 15 is a flowchart of image monitoring processing in the present embodiment.

  When each device constituting the image monitoring system 1 is started and then monitoring is started, the image monitoring device 3 repeats the processing from step S1 to step S9 shown in the flowchart of FIG. Information such as the background image is generated before monitoring is started and stored in the storage unit 32.

  First, in step S <b> 1, the image input unit 31 inputs an input image to the storage unit 32 and the image processing unit 33.

  In step S2, the image processing unit 33 performs a delivery detection process. The delivery detection process in this embodiment will be described later.

  In step S3, the process branches depending on whether or not a delivery action is detected by the delivery detection means 333. In the present embodiment, the process branches depending on whether or not the delivery pair is detected by the delivery detection means 333. If a delivery action is detected by the delivery detection means 333, the process proceeds to step S4. If no delivery action is detected by the delivery detection unit 333, the process returns to step S1 to acquire the next input image.

  From step S4 to step S8, the processing from step S5 to step S7 is performed for each delivery action detected by the delivery detection means 333, and the process is repeated until all delivery actions are processed. That is, in the present embodiment, the processing from step S5 to step S7 is performed for each delivery pair, and the processing is repeated until the processing is performed for all delivery pairs.

  In step S5, processing for determining the type of the delivered article is performed for the delivery action detected by the delivery detection means 333. In the present embodiment, the delivery detection unit 333 refers to the tracking person information of the delivery pair that is the processing target. Then, the delivery detection means 333 determines the article name and the article ID of the same possessed article that has disappeared from one person and appeared from the other person as the kind of article delivered.

  In step S <b> 6, processing for determining the occurrence position of the delivery action is performed for the delivery action detected by the delivery detection means 333. In the present embodiment, the delivery detection unit 333 refers to the tracking person information of the delivery pair that is the processing target. Then, in the delivery detection means 333, among the positions of the persons stored in the tracked person information, the position of one person at the time when the possessed article disappears and the other person at the time when the possessed article appears. The midpoint of the position is determined as the occurrence position of the delivery action.

  In step S7, the monitoring importance calculation means 334 refers to the monitoring reference information 323 in the storage unit 32, and calculates the monitoring importance based on the type of the delivered article and the position where the delivery action occurs. In the present embodiment, first, the monitoring importance level calculation means 334 refers to the area information 322 in the storage unit 32, and determines the occurrence area of the delivery action based on the occurrence position of the delivery action. Then, the monitoring importance level calculation means 334 refers to the monitoring standard information 323 and obtains the monitoring importance level corresponding to the type of the delivered article and the area where the delivery action occurs.

  When step S5 to step S7 are performed for all delivery actions, the process proceeds to step S9. In step S9, the output control unit 335 performs output control of the delivery information according to the monitoring importance level obtained by the monitoring importance level calculation unit 334. When the output control by the output control means 335 is completed, the process returns to step S1 to acquire the next input image.

  Next, with reference to FIG. 16, the delivery detection process (step S2 in FIG. 15) of the image processing unit 33 in the present embodiment will be described in detail. FIG. 16 is a flowchart of the delivery detection process in the present embodiment.

  The delivery detection process (step S2) in the present embodiment starts when the input image is input to the image processing unit 33 by the image input unit 31.

  First, in step S201, the person detection unit 331 performs a person detection process and a person tracking process. In the present embodiment, the person detection unit 331 performs a person image extraction process from the input image, and performs a tracking process on the extracted person image. Here, a new person ID is assigned to a person newly detected by the person detection means 331, and the same person ID is inherited for a person who has already started tracking by the previous time. Thereafter, the position of the tracking person is obtained by the person detecting means 331. The person detection unit 331 stores the person image extracted this time, the shooting time of the person image, and the position of the tracked person in the storage unit 32 as tracked person information in association with the person ID. Here, for a person who has already started tracking up to the previous time, the person image extracted this time, the shooting time of the person image, and the position of the person to be tracked are associated with the person ID assigned last time. Is stored in the storage unit 32.

  In step S202, the process branches depending on whether or not there is a person (tracked person) tracked by the person detection unit 331. If there is a tracking person, the process proceeds to step S203. If no tracked person exists, that is, if no person image is extracted, or if all tracked persons tracked up to the previous time have been tracked, the process proceeds to step S214. .

  Steps S203 to S206 are repeated until the processing from step S204 to step S205 is performed for each tracked person, and the process is performed for all the tracked persons.

  In step S <b> 204, the possessed article detection unit 332 performs possessed article disappearance determination processing on the tracking person. In the present embodiment, when the possessed article detection unit 332 does not detect the possessed article detected up to the previous time from the current person image, the possessed article ID of the possessed article with respect to the person image. Disappearance flag is set. Note that the possessed article detection unit 332 detects the possessed article and determines whether or not the possessed article is the same as the previous article with reference to the article information 321 and the tracking person information in the storage unit 32.

  In step S <b> 205, the possessed article detection unit 332 performs possessed article appearance determination processing on the tracked person. In the present embodiment, when a new belonging is detected in the person image of the tracked person by the belonging detection means 332, a belonging ID is given to the possessed area of the possessed article, The belonging area is stored in the tracking person information in the storage unit 32 in association with the person image. Thereafter, the article information 321 in the storage unit 32 is referred to, and the article ID and the article name are stored in the tracking person information in association with the belonging ID. Then, the belonging detection unit 332 sets an appearance flag of the belonging ID for the person image. In addition, when the possessed article detected from the current person image is the same possessed article as the possessed article in the previous person image, the possessed article detection unit 332 applies to the possessed area detected from the present person image. Then, the personal belonging ID assigned last time is associated and stored in the tracking person information. At this time, the belonging detection unit 332 does not set the appearance flag of the belonging ID for the person image. Note that the possessed article detection means 332 detects the possessed article and determines whether or not it is the same possessed article as the previous one, with reference to the article information 321 and the tracking person information in the storage unit 32.

  When step S204 to step S205 are performed for all tracked persons, the process proceeds to step S207. In step S207, a pair of a person whose possessed article appears in the current process and a person whose possessed article has disappeared within the past predetermined time is set. Specifically, referring to the tracking person information, a combination of the tracking person with the appearance flag set in the current process and the tracking person with the disappearance flag set in the past predetermined time is created.

  From step S208 to step S215, the processing from step S209 to step S213 is performed for each pair set in step S207, and the processing is repeated until all pairs are processed.

  In step S209, the delivery detection means 333 determines the identity of the possessed article. More specifically, each tracking person information is referred to by the delivery detection means 333, and the article name and article ID of the belonging ID for which the appearance flag is set and the article name of the belonging ID for which the disappearance flag is set. Whether or not the article IDs are the same is determined. At this time, it may be determined whether or not the image features in the belonging area are similar. In the set pair, when there are a plurality of possession IDs for which one tracking person has a disappearance flag, the identity of the articles is determined for the possession areas of the respective possession IDs. Also, in the set pair, even when there are a plurality of belonging IDs for which an appearance flag is set for one tracking person, the identity of the articles is determined for the belonging areas of the respective belonging IDs.

  In step S210, the delivery detection unit 333 determines whether the distance between persons is within a predetermined distance. Specifically, the passing detection means 333 refers to each tracked person information, and calculates the distance between the position of the person when the appearance flag is set and the position of the person when the disappearance flag is set. Is done. Then, the delivery detection means 333 determines whether or not the calculated distance is within a predetermined distance. At this time, in the set pair, if there are a plurality of possession IDs for which one tracked person has a disappearance flag, the disappearance flag for the possession ID determined to be the same in step S209 is the possession ID. Using the position of the person at the time of standing, the distance between the persons is obtained. In addition, in the set pair, even when there are a plurality of belonging IDs for which an appearance flag is set for one tracking person, an appearance flag for the belonging ID determined to be the same in step S209 is set. The distance between the persons is obtained using the positions of the persons at that time.

  In step S211, when it is determined in step S209 that the possessed items are the same, and in step S210, it is determined that the distance is within a predetermined distance, the pair to be processed is delivered in step S212. Is determined to have been performed (delivery pair). In other cases, it is determined in step S213 that the pair to be processed is not performing a delivery act.

  When step S209 to step S213 are performed for all pairs, the process proceeds to step S3 shown in FIG. In step S202, if there is no tracking person, it is determined in step S214 that there is no delivery action, and the process proceeds to step S3 shown in FIG.

  According to the present embodiment, a delivery act occurring at a monitoring location is automatically detected, and for each detected delivery act, the supervisor monitors based on the delivered goods and the location where the delivery act occurs. The monitoring importance level indicating the degree of attention as a target is obtained. And the detection result of a delivery action is output according to the calculated | required monitoring importance. As a result, in a place where an unspecified number of people gather, such as in an event venue, the monitor can efficiently monitor the delivery action. For this reason, it becomes possible to deal with various abnormal situations occurring in the monitoring place at an early stage with the delivery act as an opportunity.

  Hereinafter, another embodiment of the image monitoring system 1 of the present invention will be described. Note that a description of processing common to the above-described embodiment will be omitted, and processing unique to the present embodiment will be described.

  In the image monitoring system 1 of the above-described embodiment, the monitoring importance is calculated based on the type of the delivered article and the type of the place where the delivery act is generated. However, the image monitoring system 1 in the present embodiment. Uses other information to calculate the monitoring importance of the delivery action. Specifically, in the present embodiment, the time when the delivery act occurs, the frequency of occurrence of the delivery act at the location where the delivery act occurs, the degree of congestion around the location where the delivery act occurs, the location from the location where the delivery act occurs to the important area The monitoring importance of the delivery action is calculated using the distance, the distance from the occurrence position of the delivery action to the predetermined monitoring person, and these information.

  First, an embodiment in which the monitoring importance is calculated based on the time when the delivery action occurs will be described. In the present embodiment, as the monitoring reference information 323, the article ID, the article name, the area ID, the area name, the time zone information, and the monitoring importance are associated and stored in the storage unit 32. In the present embodiment, the delivery detection unit 333 obtains the occurrence time of the delivery action in addition to the type of the delivered article and the occurrence position of the delivery action. Specifically, the delivery detection unit 333 refers to the tracking person information, and obtains the shooting time of the person image where the possessed article appears among the person images for which the occurrence position of the delivery action has been obtained, as the occurrence time of the delivery action. Then, the monitoring importance level calculation means 334 refers to the monitoring standard information 323 in the storage unit 32 and calculates the monitoring importance level corresponding to the type of the delivered item, the occurrence position of the delivery action, and the occurrence time of the delivery action. To do.

  Here, the time zone information stored in the monitoring reference information 323 includes, for example, a time zone from the opening time of the event venue to the start time of the event, a time zone from the start time of the event to the end time of the event, an important person Is the time to visit the event venue. The time zone from the opening time of the event venue to the start time of the event is a time zone in which the vicinity of the entrance gate is congested, and there is a possibility that suspicious delivery acts may be performed at the entrance gate boundary due to congestion. For this reason, at the entrance gate boundary, the monitoring importance of the act of delivering a plastic bottle, the act of delivering a ticket, the act of delivering a bag, etc. is set higher than that outside this time period. In addition, it is necessary to increase alertness in the event venue, particularly during the time period from the start time of the event to the end time of the event. For this reason, the monitoring importance of the delivery action performed in the event venue is set higher than the time zone before the start of the event or after the end of the event. For example, the monitoring importance of the camera delivery action in the event hall is set high, or the monitoring importance of the bag delivery action in the event hall is set high. Further, in the time zone when the important person visits the event venue, it is necessary to increase the alertness in the area related to the movement path of the important person. For this reason, an area related to the movement route of the important person is set in the area information 322 in advance, and the monitoring importance level of the delivery action performed in the area in the time zone when the important person visits the event venue is expressed as follows. Set higher than other time zones.

  According to the present embodiment, it is possible to obtain the monitoring importance level in consideration of not only the type of the delivered item and the type of the place where the delivery action occurs, but also the status of the monitoring area that changes depending on the time zone.

  The time zone information is not limited to the time but may be set by date, day of the week, or the like. According to this, the monitoring importance level of the delivery action performed on the date of the event can be set higher than the monitoring importance level of the delivery action performed on the date other than the event date. In this case, as the information indicating the shooting time, not only the shooting time but also the shooting date and the shooting day are added to the input image. What is necessary is just to memorize the date when it was taken and the day of the week when it was taken.

  In the present embodiment, the time zone information is stored in the monitoring reference information 323 in association with the article name or area, but the present invention is not limited to this. For example, a predetermined addition value may be added to the monitoring importance level according to the occurrence time of the delivery action. Specifically, a predetermined addition value is associated with each time zone information and stored in the storage unit 32 in advance. Then, the monitoring importance calculation means 334 calculates the monitoring importance from the type of the delivered item and the occurrence position of the delivery action, and then reads the added value corresponding to the occurrence time of the delivery action from the storage unit 32, The added value may be added to the monitoring importance.

  Note that the time zone information is not limited to that exemplified in the present embodiment, and may be appropriately determined according to the monitoring purpose.

  Next, an embodiment in which the monitoring importance is calculated based on the frequency of occurrence of the delivery action will be described. Baggage deposit boxes and shops located at event venues usually carry baggage such as bags, and deliver goods and money. Therefore, in such a place, the frequency of occurrence of the delivery action increases. In addition, the frequency of occurrence of delivery acts also increases around the person distributing the flyer. That is, when a delivery action is detected in a place where the frequency of occurrence of the delivery action is high, there is a high possibility that it is a delivery action with a low suspicious degree that is normally performed. On the other hand, when a delivery action is detected in a place where the frequency of occurrence of the delivery action is low, it means that the delivery action is performed in a place where delivery is not normally performed. Therefore, it is necessary for the observer to be wary of the delivery action. In the present embodiment, this is used to calculate the monitoring importance based on the frequency of occurrence of the delivery action.

  In the present embodiment, the image processing unit 33 of the image monitoring apparatus 3 further includes a frequency calculating unit 336 (not shown). The frequency calculation means 336 calculates the frequency of the delivery action in the monitoring area for each predetermined area and stores it in the storage unit 32 as frequency information. Then, the monitoring importance level calculation unit 334 refers to the monitoring reference information 323 stored in the storage unit 32 and calculates the monitoring importance level based on the type of the delivered article and the occurrence position of the delivery action. Thereafter, the monitoring importance level calculation means 334 refers to the frequency information stored in the storage unit 32 and obtains the frequency of the delivery action at the position where the delivery action occurs. When the frequency is equal to or lower than the first predetermined value, the monitoring importance level calculation unit 334 sets a predetermined value (for example, 5) to the monitoring importance level calculated based on the type of the delivered article and the position where the delivery act is generated. Is added. Further, when the frequency is equal to or higher than the second predetermined value, the monitoring importance level calculation unit 334 sets a predetermined value (for example, the monitoring importance level calculated based on the type of the delivered article and the occurrence position of the delivery action) 5) is subtracted.

  Here, the frequency calculation means 336 will be described. As described above, the frequency calculation unit 336 obtains the frequency of occurrence of the delivery action in the monitoring area for each predetermined area. Specifically, the monitoring area is divided into predetermined frequency calculation areas, and the position of the frequency calculation area and the block position of the monitoring area are stored in association with each other. Then, the frequency calculation unit 336 counts the number of delivery actions that have occurred within the frequency calculation area based on the delivery action occurrence position detected by the delivery detection means 333. In the present embodiment, for each frequency generation area, the number of delivery actions counted within the past predetermined time (for example, one hour) is set as the occurrence frequency of the delivery action, and stored in the storage unit 32 as frequency information. In addition, if the place where the frequency of occurrence of delivery acts is high, such as a baggage storage or a store, the predetermined time for counting the frequency may be set to a time longer than one hour. However, since there are things where the frequency of occurrence of delivery acts changes in time, such as around the person who distributes the flyer, in consideration of such a change in the frequency of occurrence in time, in this embodiment, The predetermined time is set to 1 hour.

  The frequency calculation area will be described with reference to FIG. FIG. 17 is a diagram illustrating a setting example of a monitoring area block and a frequency calculation area according to the present embodiment. FIG. 17 corresponds to the monitoring area shown in FIG. As shown in FIG. 17, the monitoring area is divided into blocks of a predetermined size in advance, and all blocks are set to area ID “001” as in FIG. In FIG. 17, one frequency calculation area in the present embodiment is indicated by a bold rectangular frame K. The size of the frequency calculation area is an actual distance of 10 m × 10 m, and is 4 blocks in the X-axis direction and 4 blocks in the Y-axis direction. In the present embodiment, the frequency calculation areas of this size are arranged at intervals of 5 m with respect to the monitoring area.

  The frequency information calculation process will be described with reference to the flowchart of FIG. Although not shown, the frequency calculation means 336 performs frequency information calculation processing between step S8 and step S9. First, the frequency calculation means 336 refers to the occurrence position of the delivery action determined in step S6 for the delivery action detected in step S2, and determines the number of delivery actions that occurred in the area for each frequency calculation area. Count. By repeating this counting process for each input image, the frequency information in the storage unit 32 is updated. Here, the frequency calculation means 336 sets the total number of delivery actions that have occurred within the past predetermined time from the present time as the occurrence frequency for each frequency calculation area. That is, for a predetermined time, it is subtracted from the total number. Note that the frequency information at the time of starting monitoring is set to “0” as an initial value for each frequency calculation area.

  Next, with reference to FIG. 18, the monitoring importance degree calculation process based on the frequency of occurrence of the delivery action in the present embodiment will be described. FIG. 18 is a diagram schematically showing an occurrence frequency area and a delivery occurrence position in the present embodiment. Here, an example will be described in which two delivery actions occur on the monitoring area shown in FIG. FIG. 18 shows the generation positions 81 and 82 of the delivery action on the monitoring area divided into blocks, and the frequency calculation areas corresponding to the respective generation positions. In the two delivery actions, the goods delivered together are “bags” and the place of occurrence is “open space”. First, the monitoring importance level calculation means 334 refers to the monitoring standard information 323 in the storage unit 32 and calculates the monitoring importance level for the detected delivery action based on the type of the delivered item and the type of the occurrence location. To do. According to the monitoring reference information 323 shown in FIG. 9, the monitoring importance level of each delivery action is calculated as the same value “7”. Thereafter, the monitoring importance degree calculation means 334 obtains a frequency calculation area corresponding to the position where each delivery action occurs. As shown in FIG. 18, the occurrence position 81 corresponds to the frequency calculation area K, the frequency calculation area L, and the frequency calculation area M, and the occurrence position 82 corresponds to only the frequency calculation area N. Here, when a plurality of frequency calculation areas correspond, such as the generation position 81, the monitoring importance calculation unit 334 selects the one whose generation position is the center of the frequency calculation area. That is, for the generation position 81, the frequency calculation area M is selected. The monitoring importance level calculation unit 334 reads the frequency in the frequency calculation area corresponding to the occurrence position with reference to the frequency information in the storage unit 32. Here, it is assumed that the frequency of the frequency calculation area M is “8” and the frequency of the frequency calculation area N is “0”. The monitoring importance level calculation unit 334 determines whether each frequency is equal to or higher than a first predetermined value (for example, 5), and subtracts 5 from the monitoring importance level when the frequency is higher than the predetermined value. Since the frequency of the frequency calculation area M is “8”, the monitoring importance level of the delivery action at the occurrence position 81 is “2”. Further, since the frequency of the frequency calculation area N is “0”, the monitoring importance of the delivery action at the occurrence position 82 is not subtracted. Moreover, the monitoring importance calculation means 334 determines whether each frequency is below a second predetermined value (for example, 1), and adds 5 to the monitoring importance when it is below a predetermined value. Since the frequency of the frequency calculation area M is “8”, the monitoring importance level of the delivery action at the occurrence position 81 remains “2”. Further, since the frequency of the frequency calculation area N is “0”, the monitoring importance of the delivery action at the occurrence position 82 is “12”.

  The delivery action at the generation position 81 and the delivery action at the generation position 82 are both actions of delivering a bag in a plaza. However, if the monitoring importance is calculated based on the frequency of occurrence of the delivery at the occurrence position, the respective monitoring importances are “2” and “12”, and the degree to which the monitor should pay attention changes. Since the delivery action at the generation position 81 is performed at a position where the frequency of the delivery action is high, there is a high possibility that it is a bag delivery action performed at a baggage custody provided in the open space. On the other hand, since the delivery action at the generation position 82 is a bag delivery performed in a place where the delivery is hardly performed, there is a possibility that dangerous goods have been delivered between suspicious persons. According to the present embodiment, the monitoring staff confirms the delivery act that may have been delivered of the dangerous goods between the suspicious persons in preference to the delivery act that may have been performed at the baggage storage. can do.

  Note that the frequency calculation method is not limited to this embodiment. For example, the occurrence position of the delivery action detected in the past predetermined time is stored in the storage unit 32, and the monitoring importance calculation means 334 reads the stored occurrence position when calculating the monitoring importance. Alternatively, the number of occurrences within a predetermined range centered on the occurrence position of the delivery action that is the processing target may be counted and obtained as the occurrence frequency of the delivery action. In addition, the frequency information may be stored in the storage unit 32 as frequency information in advance by counting the frequency of occurrence of the monitoring area passing action before the monitoring by the image monitoring system 1 starts.

  In the present embodiment, the monitoring importance is subtracted when the occurrence frequency is high, and is added when the occurrence frequency is low. However, the present invention is not limited to this. For example, in a place where the frequency of occurrence is high, it is determined that there is a possibility of passing suspiciously in a normal delivery action, and the monitoring importance is added to the delivery action detected in a place where the occurrence frequency is high. Also good. Further, the monitoring importance level may be changed depending on the occurrence frequency by either one of the above-described addition and subtraction.

  Further, in the present embodiment, when a plurality of frequency calculation areas are associated with one occurrence position, the one where the occurrence position is closest to the center of the frequency calculation area is selected. Absent. For example, the one with the highest frequency or the one with the lowest frequency may be selected. Further, in the present embodiment, for one selected frequency calculation area, the added value corresponding to the frequency of the area is added to the monitoring importance level, but the present invention is not limited to this. For example, instead of selecting one frequency calculation region, an addition value corresponding to each frequency may be added to the monitoring importance level for all frequency calculation regions associated with the occurrence position.

  Further, in the present embodiment, the frequency of occurrence of the delivery action is calculated regardless of the delivered article. However, the present invention is not limited to this, and the frequency of occurrence may be calculated for each delivered article. In this case, the frequency calculation unit 336 may count the number of delivery actions for each delivered article for each frequency calculation area and store it as frequency information. Then, the monitoring importance calculation means 334 may calculate the monitoring importance based on the frequency of occurrence of the delivery action of the same article type as the delivery action that is the processing target.

  In the present embodiment, the predetermined value is added to or subtracted from the monitoring importance level based on the type of the delivered article and the type of the place where the delivery occurs, but the present invention is not limited to this. For example, the monitoring standard information 323 is stored in association with the type of the delivered item, the location of the delivery, and the frequency of the delivery, and the monitoring importance calculation means 334 stores the monitoring criteria information 323. With reference to, the importance of monitoring determined by the type of the delivered article, the location where the delivery occurs, and the frequency of the delivery may be calculated. For example, as the monitoring reference information 323, a bag, a plaza, an occurrence frequency “low”, and a monitoring importance level “7” may be associated with each other.

  Further, in the present embodiment, the monitoring importance is calculated based on the type of the delivered article and the type of the place where the delivery occurs, but the present invention is not limited to this. For example, the monitoring importance level of the detected delivery action may be calculated using only the occurrence frequency information. In this case, as the monitoring reference information 323, the monitoring importance is set in association with the occurrence frequency, and the monitoring importance calculation unit 334 may calculate the monitoring importance by referring to the monitoring reference information 323. . Further, the monitoring importance may be calculated based on the information on the occurrence frequency and the type of the delivered article or the location where the delivery occurs.

  According to the present embodiment, it is possible to obtain the monitoring importance in consideration of the suspicious degree based on the frequency of occurrence of the delivery action as well as the type of the delivered article and the location of the delivery action.

  Next, an embodiment in which the monitoring importance is calculated based on the degree of congestion around the location where the delivery action occurs will be described. When a suspicious person conducts dangerous goods such as a dangerous goods delivery action, there is a high possibility that a place with less human eyes will be selected as a delivery place. On the other hand, there is a possibility of avoiding human eyes by performing a delivery act in a crowded manner. That is, when a delivery action is detected in a place where the degree of congestion is low, there is a high possibility that the delivery action is highly suspicious. In addition, when a delivery action is detected in a place where the degree of congestion is high, there is a high possibility that the delivery action is highly suspicious. On the other hand, in places where people are reasonably present, people's eyes can reach well, so if a delivery act is detected in such a place, it is compared to a place with a high degree of congestion or a place with a low degree of congestion. Therefore, the possibility of suspicious delivery is reduced. In this embodiment, this is used to calculate the monitoring importance based on the degree of congestion around the location where the delivery action occurs.

  In the present embodiment, the monitoring importance calculation means 334 first refers to the monitoring reference information 323 stored in the storage unit 32, and monitors the importance of monitoring based on the type of the delivered article and the occurrence position of the delivery action. Calculate the degree. Thereafter, the monitoring importance calculation means 334 refers to the position of the tracking person stored as the tracking person information, and obtains the degree of congestion around the position where the delivery action occurs. Then, when the congestion degree is equal to or less than the first predetermined value, the monitoring importance degree calculation unit 334 sets a predetermined value (for example, the monitoring importance degree calculated based on the type of the delivered article and the occurrence position of the delivery action) 5) is added. In addition, when the congestion degree is equal to or greater than the second predetermined value, the monitoring importance degree calculation unit 334 sets a predetermined value (for example, the monitoring importance degree calculated based on the type of the delivered article and the position where the delivery action is generated) 5) is added.

  Here, the calculation process of the degree of congestion will be described. The monitoring importance level calculation means 334 counts the number of persons existing within a predetermined range centered on the occurrence position of the delivery action detected by the delivery detection means 333, and the degree of congestion around the delivery action occurrence position. And Specifically, the monitoring importance level calculation means 334 reads the current position of the tracked person with reference to the tracked person information of the tracked person currently present in the monitoring area. Then, based on the current position of the tracking person, the monitoring importance level calculation means 334 detects the tracking person existing within a predetermined range (for example, within a radius of 5 m) centered on the position where the delivery action that is the processing target occurs. Count the number. The method for calculating the degree of congestion is not limited to this. For example, similar to the method of calculating the frequency of occurrence of the delivery action, the monitoring area is divided into areas of a predetermined size, the number of tracking persons is counted for each area, the congestion degree is calculated, and the processing target The degree of congestion in the area corresponding to the position where the delivery action occurs may be set as the degree of congestion around the position where the delivery action occurs. Further, instead of calculating the degree of congestion using the number of tracking persons, the degree of congestion may be obtained by image processing. For example, an area of a predetermined size centered on the occurrence position of the delivery action is set on the input image, and the area of all human image areas existing in the area obtained by background difference processing or the like from the area of the area The value obtained by subtracting is obtained. The monitoring importance level calculation means 334 may take the subtracted value as the congestion level around the position where the delivery action occurs. In this case, the subtracted value becomes an index indicating that the lower the value, the higher the degree of congestion.

  Next, with reference to FIG. 19, the calculation process of the monitoring importance based on the degree of congestion in the present embodiment will be described. FIG. 19 is a diagram schematically showing the position of the tracking person and the position where the delivery occurs in the present embodiment. FIG. 19 corresponds to the monitoring area shown in FIG. As shown in FIG. 19, the monitoring area is divided into blocks of a predetermined size in advance, and all blocks are set to area ID “001” as in FIG. Here, an example in which three delivery actions have occurred on the monitoring area shown in FIG. 5 will be described. In FIG. 19, delivery positions 83, 84, and 85 are shown on the monitoring area divided into blocks, and the positions of the tracking persons existing in the monitoring area are shown by black circles. In all three delivery actions, the delivered goods are “bags” and the place of occurrence is “open space”. First, the monitoring importance level calculation means 334 refers to the monitoring standard information 323 in the storage unit 32 and calculates the monitoring importance level for the detected delivery action based on the type of the delivered item and the type of the occurrence location. To do. According to the monitoring reference information 323 shown in FIG. 9, the monitoring importance level of each delivery action is calculated as the same value “7”. Thereafter, the monitoring importance calculation means 334 obtains the congestion level corresponding to the position where each delivery action occurs. Here, the predetermined range centering on the generation position is indicated by a bold circle. As shown in FIG. 19, there are nine persons around the generation position 83 including the person who is delivering. Further, only the person who is delivering is present around the generation position 84. In addition, around the generation position 85, there are four persons including the person who is delivering. That is, the degree of congestion around the generation position 83 is “9”, the degree of congestion around the generation position 84 is “2”, and the degree of congestion around the generation position 85 is “4”. The monitoring importance level calculation means 334 determines whether each congestion level is equal to or less than a first predetermined value (for example, 2), and adds 5 to the monitoring importance level when it is equal to or less than the predetermined value. Since the degree of congestion around the generation position 83 is “9”, the monitoring importance level of the delivery action at the generation position 83 remains “7” without being added. Further, since the degree of congestion around the generation position 84 is “2”, the monitoring importance level of the delivery action at the generation position 84 is added to “12”. In addition, since the degree of congestion around the generation position 85 is “4”, the monitoring importance level of the delivery action at the generation position 85 remains “7” without being added. Further, the monitoring importance level calculation means 334 determines whether each congestion level is equal to or higher than a second predetermined value (for example, 9), and adds 5 to the monitoring importance level when it is equal to or higher than the predetermined value. Since the degree of congestion around the generation position 83 is “9”, the monitoring importance level of the delivery action at the generation position 83 is added to “12”. Further, since the degree of congestion around the generation position 84 is “2”, the monitoring importance of the delivery action at the generation position 84 remains “12” without being added. In addition, since the degree of congestion around the generation position 85 is “4”, the monitoring importance level of the delivery action at the generation position 85 remains “7” without being added.

  The delivery action at the generation position 83, the delivery action at the generation position 84, and the delivery action at the generation position 85 are all actions of delivering a bag in a plaza. However, if the monitoring importance is calculated based on the degree of congestion in the vicinity of the occurrence position, the respective monitoring importances are “12”, “12”, and “7”, and the degree to which the monitor should pay attention changes. Since the delivery action at the occurrence position 83 is performed in a place where the degree of congestion is low, there is a high possibility that the suspicious person is carrying out the delivery action while avoiding eyes. Moreover, since the delivery action of the generation | occurrence | production position 84 was performed in the place where the degree of congestion is high, there is a high possibility that a suspicious person avoids a person's gaze by performing the delivery action in a crowded state. Moreover, the delivery action of the generation | occurrence | production position 85 is performed in the place where the degree of congestion is neither high nor low, and is the delivery action performed in the place where a person exists moderately. In other words, since the delivery action at the occurrence position 85 is performed in a situation where human eyes can reach well, there is a possibility that a suspicious delivery action is performed compared to a place where the degree of congestion is high or a place where the degree of congestion is low. Lower. According to the present embodiment, the monitoring staff can preferentially confirm the delivery action in which the dangerous goods may have been delivered between the suspicious persons.

  In this embodiment, the monitoring importance is added when the degree of congestion is high or when the degree of congestion is low. However, the present invention is not limited to this. For example, the monitoring importance may be subtracted when the degree of congestion is neither high nor low. Further, the change of the monitoring importance level based on the degree of congestion may be performed by either adding when the degree of congestion is high or adding when the degree of congestion is low.

  In the present embodiment, the predetermined value is added to the monitoring importance based on the type of the delivered article and the type of the place where the delivery occurs, but the present invention is not limited to this. For example, the type of delivered item, the location where the delivery occurred, and the congestion level information are stored in association with the monitoring standard information 323, and the monitoring importance calculation unit 334 refers to the monitoring standard information 323. Thus, the monitoring importance level determined by the type of the delivered item, the location where the delivery occurs, and the degree of congestion may be calculated. For example, as the monitoring reference information 323, a bag, a plaza, a congestion level “low”, and a monitoring importance level “7” may be associated with each other.

  Further, in the present embodiment, the monitoring importance is calculated based on the type of the delivered article and the type of the place where the delivery occurs, but the present invention is not limited to this. For example, the monitoring importance level of the detected delivery action may be calculated using only the congestion level information. In this case, the monitoring importance level associated with the congestion level is set as the monitoring standard information 323, and the monitoring importance level calculation unit 334 calculates the monitoring importance level with reference to the monitoring standard information 323. . Further, the monitoring importance may be calculated based on the information on the degree of congestion and either the type of the delivered item or the type of the place where the delivery occurs.

  According to the present embodiment, it is possible to obtain the monitoring importance based not only on the type of the delivered item and the type of the place where the delivery act occurs but also on the suspicious degree that changes depending on the degree of congestion around the place where the delivery act occurs. it can.

  Next, an embodiment in which the monitoring importance is calculated based on the distance from the position where the delivery action occurs to the important area will be described. As described above, in the area information 322 stored in the storage unit 32, the area set as the monitoring area is divided into a normal area and an important area according to the importance of security for the area. According to the area information 322 of this embodiment shown in FIG. 8, the inside of the event venue is set as an important area, and the other areas are set as normal areas. When a suspicious person delivers dangerous goods, there is a high possibility that dangerous goods will be used in an event venue, which is an important area. At this time, the dangerous goods are not necessarily delivered in the important area, but may be brought into the event venue and used after being delivered in a normal area such as a plaza. In other words, if a suspicious delivery act is confirmed in the surveillance area, the observer gives instructions on how to deal with the delivery act to the on-site guard, but the suspicious delivery act is close to the event venue. The more it is done, the more urgent it is to deal with it. Therefore, it is necessary for the observer to confirm with priority the delivery action performed near the event venue. In the present embodiment, this is used to calculate the monitoring importance based on the distance from the position where the delivery action occurs to the important area.

  In the present embodiment, the monitoring importance level calculation means 334 refers to the monitoring standard information 323 stored in the storage unit 32, and determines the monitoring importance level based on the type of the delivered item and the occurrence position of the delivery action. calculate. Thereafter, the monitoring importance calculation means 334 refers to the area information 322 stored in the storage unit 32 and obtains the distance from the position where the delivery action occurs to the important area. When the distance to the important area is equal to or less than the predetermined distance, the monitoring importance level calculation unit 334 sets a predetermined value (for example, 5) to the monitoring importance level calculated based on the type of the delivered item and the occurrence position of the delivery action ) Is added.

  Here, the process of calculating the distance to the important area will be described. In the present embodiment, the monitoring importance calculation means 334 obtains the shortest distance from the occurrence position of the delivery action detected by the delivery detection means 333 to the important area. Specifically, the monitoring importance calculation unit 334 reads the block position set in the important area with reference to the area information 322. Moreover, the monitoring importance calculation means 334 refers to the area information 322, and obtains a block corresponding to the position where the delivery action that is the processing target is generated. Here, when determining the block corresponding to the occurrence position of the delivery action, based on the correspondence relationship between the XY coordinates and the block position in the real space of the monitoring area for each imaging device 2 stored in the storage unit 32. Judgment. The monitoring importance calculation means 334 calculates the distance from the block position of the read important area to the block at the position where the delivery action is processed, and the shortest distance among the calculated distances is the distance to the important area. To do. Here, the important area block to be read may be limited to the important area block set in the same monitoring area as the delivery position. At this time, if the important area block is not set in the same monitoring area as the delivery action occurrence position, calculate the distance between the important block set in the other monitoring area and the delivery action occurrence position, What has become the shortest distance may be the distance to the important area. Further, when the position where the delivery action occurs is within the important area, the distance to the important area may be set to “0”. Note that the method for calculating the distance to the important area is not limited to this. For example, the distance to the important area may be calculated and stored in advance for each block, or the distance to the important area may be calculated and stored in advance for each imaging device 2.

  Next, the monitoring importance degree calculation process based on the distance to the important area in the present embodiment will be described with reference to FIG. FIG. 20 is a diagram schematically showing the delivery position and the distance to the important area in the present embodiment. FIG. 20 corresponds to the monitoring area shown in FIG. As shown in FIG. 20, the monitoring area is divided into blocks of a predetermined size in advance, and the blocks are set to area ID “001” and area ID “004” as in FIG. Here, an example will be described in which two delivery actions occur on the monitoring area shown in FIG. FIG. 20 shows the generation positions 86 and 87 of the delivery action on the monitoring area divided into blocks. In the two delivery actions, the goods delivered together are “bags” and the place of occurrence is “open space”. First, the monitoring importance level calculation means 334 refers to the monitoring standard information 323 in the storage unit 32 and calculates the monitoring importance level for the detected delivery action based on the type of the delivered item and the type of the occurrence location. To do. According to the monitoring reference information 323 shown in FIG. 9, the monitoring importance level of each delivery action is calculated as the same value “7”. Thereafter, the monitoring importance level calculation unit 334 reads the block position of the important area with reference to the area information 322. In the monitoring area shown in FIG. 7, an area with area ID “004”, that is, an important area is set. Here, in order to simplify the description, an example in which only the block position of the important area of the monitoring area shown in FIG. 7 is read will be described. The monitoring importance calculation means 334 first determines the corresponding block for each delivery action occurrence position. As for the generation position 86, 5 blocks (block 5-8) correspond to the X-axis direction and 8 in the Y-axis direction. The generation position 87 corresponds to 10 blocks in the X-axis direction and 3 blocks (block 10-3) in the Y-axis direction. The monitoring importance calculation means 334 obtains the distances between the blocks at the respective occurrence positions and the blocks in the important area, and selects the shortest distance. As for the generation position 86, 5 blocks in the X-axis direction and 11 blocks (block 5-11) in the Y-axis direction become the shortest important area, and the distance is “5 m” for two blocks. Regarding the generation position 87, 10 blocks in the X-axis direction and 11 blocks (block 10-11) in the Y-axis direction are important areas of the shortest distance, and the distance is “17.5 m” for 7 blocks. The monitoring importance calculation means 334 determines whether each distance is a predetermined distance (for example, 10 m) or less, and adds 5 to the monitoring importance when it is equal to or less than the predetermined distance. As for the delivery action of the occurrence position 86, the distance to the important area is “5 m”, so the monitoring importance is “12”. In addition, since the distance to the important area is “17.5 m”, the monitoring importance is not added to the delivery action of the generation position 87 and remains “7”.

  The delivery action at the generation position 86 and the delivery action at the generation position 87 are both actions of delivering a bag in a plaza. However, when the monitoring importance is calculated based on the distance from the occurrence position to the important area, the respective monitoring importance becomes “12” and “7”, and the degree of attention of the monitor changes. Since the delivery action at the generation position 86 is performed at a position closer to the important area than the delivery action at the generation position 87, the urgency of the countermeasure is high. According to the present embodiment, the monitoring staff can preferentially confirm from the delivery act of the generation position 86 performed near the event venue.

  In the present embodiment, the monitoring importance is added when the distance to the important area is equal to or less than the predetermined distance, but the present invention is not limited to this. For example, the monitoring importance may be subtracted when the distance to the important area is a predetermined distance or more.

  In the present embodiment, the predetermined value is added to the monitoring importance based on the type of the delivered article and the type of the place where the delivery occurs, but the present invention is not limited to this. For example, the monitoring reference information 323 is stored in association with the type of the delivered item, the location where the delivery has occurred, and the information on the distance to the important area, and the monitoring importance calculation means 334 stores the monitoring reference information 323. With reference to the above, the importance of monitoring determined by the type of the delivered article, the location where the delivery has occurred, and the distance to the important area may be calculated. For example, as the monitoring reference information 323, a distance “10 m or less” to a bag, a plaza, or an important area may be associated with a monitoring importance “7”.

  In the present embodiment, the monitoring importance level is determined based on whether the distance to the important area is equal to or less than a predetermined distance, but the present invention is not limited to this. For example, the added value may be changed so as to increase as the distance to the important area becomes shorter. Further, a plurality of predetermined distances such as “10 m or less”, “50 m or less”, and “500 m or less” may be prepared, and different addition values may be set for each. In the present embodiment, an example in which the area set as the monitoring area is divided into two types of “normal area” and “important area” is shown, but the present invention is not limited to this. For example, it may be divided into a plurality of types of important areas according to the importance of security. Specifically, “important area 1”, “important area 2”, and “important area 3” are set according to the importance of security. The value of the predetermined distance or the value added to the monitoring importance may be changed in each important area. In this case, the distance from the occurrence position of the delivery action is calculated for all the important areas, and the blocks of the important area having the shortest distance are “important area 1”, “important area 2”, and “important area 3”. Determine which is applicable. Then, the monitoring importance level changing process may be performed according to the determined type of the important area.

  Further, in the present embodiment, the monitoring importance is calculated based on the type of the delivered article and the type of the place where the delivery occurs, but the present invention is not limited to this. For example, the monitoring importance of the detected delivery action may be calculated using only the distance to the important area. In this case, the monitoring importance degree associated with the distance to the important area is set as the monitoring reference information 323, and the monitoring importance degree calculation unit 334 calculates the monitoring importance degree with reference to the monitoring reference information 323. do it. At this time, the monitoring importance may be continuously changed so as to become a higher value as the distance to the important area becomes shorter. Further, the monitoring importance may be calculated based on the distance to the important area and either the type of the delivered item or the type of the place where the delivery occurs.

  According to the present embodiment, monitoring is important in consideration of not only the type of goods delivered and the type of place where the delivery act occurs, but also the urgency of the action based on the distance from the location where the delivery act occurs to the critical area. Degree can be obtained.

  Next, an embodiment in which the monitoring importance is calculated based on the distance from the occurrence position of the delivery action to a predetermined monitoring person will be described. When guarding an event venue, monitoring is performed not only by image monitoring but also by visually observing the security guards by arranging a plurality of security guards at the site. A guard arranged at the site receives an instruction from a supervisor at the monitoring center, copes with an abnormal situation, or looks at the monitoring place to find a person or a suspicious object taking a suspicious action. In addition, by arranging security guards at the monitoring location, there is also an effect of restraining criminal and deterring criminal acts. For this reason, when a suspicious person performs a dangerous delivery action such as a dangerous goods delivery action, there is a high possibility that a place where the guards cannot reach will be selected. On the other hand, since it is difficult to think of a suspicious person performing a suspicious delivery act where the guardian's eyes reach, it can be said that the suspicious degree is low when the act is performed where the guardian's eyes reach. In the present embodiment, this is used to calculate the monitoring importance based on the distance from the position where the delivery action occurs to the monitoring person such as a guard.

  In the present embodiment, the image processing unit 33 of the image monitoring apparatus 3 further includes a monitoring person position calculating unit 337 (not shown). The monitoring person position calculation means 337 determines whether or not the person existing in the monitoring area is a monitoring person that satisfies a predetermined standard. Then, the tracking person information of the person determined as the monitoring person is referred to, and the position of the monitoring person at the current time is stored in the storage unit 32 as the monitoring person position information. Thereafter, the monitoring importance calculation means 334 refers to the monitoring reference information 323 stored in the storage unit 32, and calculates the monitoring importance based on the type of the delivered item and the occurrence position of the delivery action. Then, the monitoring importance calculation unit 334 refers to the monitoring person position information stored in the storage unit 32 and obtains the distance from the position where the delivery action occurs to the position of the monitoring person. When the distance is a predetermined distance (for example, 10 m) or more, the monitoring importance level calculation unit 334 sets a predetermined value (for example, the monitoring importance level) calculated based on the type of the delivered article and the occurrence position of the delivery action. 5) is added.

  Here, the monitoring person position calculation means 337 will be described. The monitoring person position calculation means 337 obtains the position of the monitoring person existing in the monitoring area at the current time. In this embodiment, the monitoring person is described as a guard. Specifically, the image characteristics of the guards are stored in advance in the storage unit 32 as person identification information, and the monitoring person position calculation unit 337 detects the tracking person when the person detection unit 331 detects the tracking person. The person image of the person is checked against the person identification information. When the image features match, the traced person is determined to be a guard. Then, the monitoring person position calculation means 337 refers to the tracking person information of the tracking person determined to be a guard, and stores the position at the current time in the storage unit 32 as the monitoring person position information. At this time, the position information is stored for each tracking person determined to be a guard. Note that the method for calculating the position of the monitoring person is not limited to the present embodiment. For example, instead of determining whether or not the person is a monitoring person based on the image feature, the position information of the terminal previously held by the monitoring person is acquired, and the position information acquired at the current time is stored as the monitoring person position information. You may do it. That is, the possession of the terminal is a reference for the monitoring person. In the present embodiment, the monitoring person is described as a predetermined person such as a guard, but the present invention is not limited to this. For example, the reference of the monitoring person may be “a person other than the person who is performing the delivery action that is present at the closest position to the delivery position”. That is, a person who exists near the position where the delivery action occurs is likely to have witnessed the delivery action, and is therefore treated as a monitoring person like a guard. In this case, the monitoring person position calculation means 337 calculates the distance from the position where the delivery action occurs to the position of the tracking person who is not in the delivery action delivery pair existing in the monitoring area, and the tracking that has become the shortest distance is performed. What is necessary is just to determine a person as a monitoring person.

  Next, the process for calculating the distance to the monitoring person will be described. In the present embodiment, the monitoring importance calculation unit 334 obtains the distance from the generation position of the delivery action detected by the delivery detection unit 333 to the monitoring person. Specifically, the monitoring importance calculation unit 334 reads the position of each monitoring person existing in the monitoring area with reference to the monitoring person position information in the storage unit 32. Then, the distance from the position of the read monitoring person to the position where the delivery action to be processed is generated is calculated. The distance calculation method may be performed by an existing method.

  Next, with reference to FIG. 21, the monitoring importance level calculation process based on the distance to the monitoring person in the present embodiment will be described. FIG. 21 is a diagram schematically illustrating a delivery occurrence position and a monitoring person position in the present embodiment. 21 corresponds to the monitoring area shown in FIG. As shown in FIG. 21, the monitoring area is divided into blocks of a predetermined size in advance, and all the blocks are set to area ID “001” as in FIG. Here, an example will be described in which two delivery actions occur on the monitoring area shown in FIG. It is assumed that there are two guards on the monitoring area. FIG. 21 shows the generation positions 88 and 89 of the delivery action on the monitoring area divided into blocks. Also, the positions of the guards present in the monitoring area are indicated by white circles, which are a monitoring person position 90 and a monitoring person position 91, respectively. In addition, the positions of the tracking persons other than the guards are indicated by black circles. In the two delivery actions, the goods delivered together are “bags” and the place of occurrence is “open space”. First, the monitoring importance level calculation means 334 refers to the monitoring standard information 323 in the storage unit 32 and calculates the monitoring importance level for the detected delivery action based on the type of the delivered item and the type of the occurrence location. To do. According to the monitoring reference information 323 shown in FIG. 9, the monitoring importance level of each delivery action is calculated as the same value “7”. Thereafter, the monitoring importance calculation means 334 reads the position of the monitoring person with reference to the monitoring person position information. The monitoring importance calculation means 334 obtains the distance from each monitoring person position with respect to each occurrence position of the delivery action. Here, when there are a plurality of monitoring persons in the monitoring area as shown in FIG. 21, the distance to the monitoring person with the shortest distance is selected. As for the generation position 88, the distance from the monitoring person position 91 is the shortest, which is “22.5 m”. The generation position 89 is “7.5 m” because the distance between the monitoring person position 90 and the monitoring person 91 is the same. The monitoring importance calculation unit 334 determines whether each distance is a predetermined distance (for example, 10 m) or more, and adds 5 to the monitoring importance when the distance is equal to or more than the predetermined distance. In the delivery action of the generation position 88, the distance to the monitoring person is “7.5 m”, so the monitoring importance level remains “7”. Further, since the distance to the monitoring person is “22.5 m”, the monitoring importance is “12” for the delivery action at the generation position 89.

  The delivery action at the generation position 88 and the delivery action at the generation position 89 are both actions of delivering a bag in a plaza. However, when the monitoring importance is calculated based on the distance from the generation position to the monitoring person, the respective monitoring importances are “7” and “12”, and the degree to which the monitoring person should pay attention changes. Since the delivery action at the occurrence position 88 is performed in a place where the guards cannot reach, it is highly likely that the delivery action is a suspicious delivery action. On the other hand, since the delivery action at the generation position 89 is performed at a place where the guards can reach, the possibility of being a suspicious delivery action is low. According to the present embodiment, the monitoring person can preferentially confirm from the delivery action of the generation position 88 performed without the eyes of the security guard at the site.

  In the present embodiment, the monitoring importance is added when the distance to the monitoring person is equal to or greater than the predetermined distance, but the present invention is not limited to this. For example, the monitoring importance may be subtracted when the distance to the monitoring person is equal to or less than a predetermined distance.

  In the present embodiment, the predetermined value is added to the monitoring importance based on the type of the delivered article and the type of the place where the delivery occurs, but the present invention is not limited to this. For example, the monitoring standard information 323 stores the type of delivered article, the location where the delivery occurred, and the information about the distance to the monitoring person in association with each other, and the monitoring importance calculation unit 334 stores the monitoring standard information 323. With reference to the above, the importance of monitoring determined by the type of the delivered article, the place where the delivery has occurred, and the distance to the monitoring person may be calculated. For example, as the monitoring reference information 323, the distance “10 m or more” to the bag, the plaza, the monitoring person, and the monitoring importance “7” may be associated with each other.

  Further, in the present embodiment, the monitoring importance level is obtained based on whether or not the distance to the monitoring person is equal to or greater than a predetermined distance, but the present invention is not limited to this. For example, the added value may be changed so as to increase as the distance to the monitoring person increases. Further, a plurality of predetermined distances such as “5 m or more”, “10 m or more”, and “100 m or more” may be prepared, and different addition values may be set for each.

  Further, in the present embodiment, the monitoring importance is calculated based on the type of the delivered article and the type of the place where the delivery occurs, but the present invention is not limited to this. For example, the monitoring importance of the detected delivery action may be calculated using only the distance to the monitoring person. In this case, the monitoring importance level associated with the distance to the monitoring person is set as the monitoring reference information 323, and the monitoring importance level calculation unit 334 calculates the monitoring importance level with reference to the monitoring reference information 323. do it. At this time, the monitoring importance may be continuously changed so as to increase as the distance to the monitoring person increases. Further, the monitoring importance may be calculated based on the distance to either the type of the delivered item or the type of the place where the delivery occurs and the monitoring person.

  According to the present embodiment, not only the type of the delivered item and the type of the place where the delivery act occurs, but also the monitoring importance in consideration of the suspicious degree based on the distance to a predetermined monitoring person such as a guard Can be requested.

  As described above, in addition to or instead of the type of goods delivered and the type of place where delivery acts occur, using the time of occurrence, the frequency of occurrence, the degree of congestion, the distance from the important area, the distance from the monitoring person Each of the embodiments for calculating the monitoring importance of the delivery action has been described. It goes without saying that the image monitoring system 1 of the present invention may be realized by combining these embodiments. For example, the monitoring importance may be calculated using both the frequency of occurrence of the delivery action and the degree of congestion around the location where the delivery action occurs.

  As described above, according to the image monitoring system 1 of the present invention, in consideration of the delivery act occurring at the surveillance place, the delivered article, the place where the delivery act was performed, the surrounding situation, and the like are considered. It is possible to determine the degree to which the monitor should pay attention as the monitoring target. Moreover, the image monitoring system 1 of the present invention outputs the detection result of the delivery action according to the degree to which the monitor should pay attention as the monitoring target. As a result, even if there are a large number of unspecified persons such as event venues, and monitoring a place where a large number of delivery actions occur, the monitoring center's observers can perform a highly suspicious delivery action, It is possible to efficiently monitor the delivery action while easily distinguishing the delivery action with a low suspicious degree.

DESCRIPTION OF SYMBOLS 1 ... Image monitoring system, 2 ... Imaging device, 3 ... Image monitoring device, 4 ... Monitoring center apparatus, 31 ... Image input part, 32 ... Memory | storage part, 33 ... Image processing unit, 41 ... display unit, 42 ... recording unit, 321 ... article information, 322 ... area information, 323 ... monitoring reference information, 331 ... person detection means, 332 ..Carrying object detection means, 333 ... delivery detection means, 334 ... monitoring importance calculation means, 335 ... output control means

Claims (9)

An image monitoring system that uses a captured image of a monitoring area to monitor the behavior of a person existing in the monitoring area,
Person detecting means for detecting a person existing in the monitoring area from the image;
Possession detection means for identifying the type of article possessed by the person from the image;
Detecting the delivery action of the article by a plurality of the persons, and determining the occurrence position of the delivery action in the monitoring area for the delivery action, and referring to the identification result of the belonging detection means, Delivery detection means for determining the type of delivered article;
A storage unit that stores monitoring reference information that associates a monitoring importance level that represents a degree of attention as the monitoring target for each combination of a position in the monitoring area and a type of article;
Monitoring importance calculation means for obtaining the monitoring importance corresponding to the combination of the occurrence position and the type of the delivered article for each delivery action with reference to the monitoring standard information;
An image monitoring system comprising: output control means for outputting a detection result of the delivery action according to the monitoring importance of the delivery action.   The said monitoring importance calculation means calculates the congestion degree of the said person around the generation | occurrence | production position of the said delivery action, and changes the said monitoring importance according to the said congestion degree. Image surveillance system. The storage unit further stores, for each delivery action, the occurrence position of the delivery action detected in the past predetermined time in the monitoring area,
The monitoring importance calculation unit changes the monitoring importance according to a total number of delivery actions detected within the past predetermined time around a position where the delivery action occurs. Item 3. The image monitoring system according to Item 2. The storage unit further stores the position of the important area set in the monitoring area,
The monitoring importance calculation means calculates a distance from a position where the delivery action occurs to a position of the important area, and changes the monitoring importance according to the distance. The image monitoring system according to any one of the above. The image monitoring system further includes monitoring person position calculation means for obtaining a position in the monitoring area of a monitoring person that satisfies a predetermined criterion,
The monitoring importance calculation unit calculates a distance from a position where the delivery action occurs to the position of the monitoring person, and changes the monitoring importance according to the distance. 5. The image monitoring system according to any one of 4 above. The delivery detection means further obtains the occurrence time of the delivery action,
The image monitoring system according to any one of claims 1 to 5, wherein the monitoring importance calculation unit further calculates the monitoring importance using a time when the delivery action occurs. The image monitoring system includes a display unit that displays a detection result of the delivery action,
7. The image monitoring system according to claim 1, wherein the output control unit causes the display unit to display so that a delivery action having a higher degree of monitoring importance is more likely to be noticed. 8. .   The output control means outputs the detection result of the delivery action for the delivery action with the monitoring importance level equal to or higher than a predetermined value, and the detection result of the delivery action for the delivery action with the monitoring importance level less than the predetermined value. The image monitoring system according to claim 1, wherein the image monitoring system is not output. An image monitoring system that uses a captured image of a monitoring area to monitor the behavior of a person existing in the monitoring area,
Possession detection means for identifying the type of article existing in the monitoring area from the image;
A pattern composed of a plurality of persons existing in the monitoring area and articles transferred between the persons is detected from the image to detect a passing action between the persons, and the passing action is performed in the monitoring area. A delivery detecting means for obtaining a position where the delivery action occurs, and for obtaining a type of the article delivered in the delivery action with reference to an identification result of the belonging detection means;
A storage unit that stores monitoring reference information that associates a monitoring importance level that represents a degree of attention as the monitoring target for each combination of a position in the monitoring area and a type of article;
Monitoring importance calculation means for obtaining the monitoring importance corresponding to the combination of the occurrence position and the type of the delivered article for each delivery action with reference to the monitoring standard information;
An image monitoring system comprising: output control means for outputting a detection result of the delivery action according to the monitoring importance of the delivery action.

Images