JP2012235300A - Leaving or carrying-away detection system and method for generating leaving or carrying-away detection record - Google Patents

Abstract

PROBLEM TO BE SOLVED: To speedily provide reference information for specifying a person performing leaving or carrying-away in a leaving or carrying-away detection system.SOLUTION: The leaving or carrying-away detection system includes: a camera 20 imaging a monitoring area; a data storage medium 34 storing an imaged image; a leaving/carrying-away determining section 52; an occurrence time operating section 54 operating occurrence time of leaving or carrying-away; an image acquirement processing section 42 acquiring the imaged image of the monitored area imaged at occurrence time from the data storage medium 34; a peripheral person extracting section 43 extracting an image of a peripheral person within a prescribed distance range from a foreign matter being an object of leaving or carrying-away from the acquired imaged image; and a log data creating section 45 creating leaving log data or carrying-away log data, which include the image of the foreign matter being the object of leaving or carrying-away and the image of the peripheral person.

Description

  The present invention relates to a leaving or removal detection system and a method of generating a leaving or removal detection record, in particular, to detect a leaving or removal occurring in a monitoring area and to provide reference information for identifying the actor. The present invention relates to a leaving / taking-off detection system and a method for generating a leaving / taking-off detection record.

  Conventionally, an image (background image) of a monitoring area without a leftover or the like is acquired in advance by an imaging device, and the monitoring area is calculated from a difference image of luminance between the current monitoring area image (monitoring image) and the background image. A leaving or removal detecting system for detecting the leaving or removal of an object in the inside is known.

  As an example of the above-mentioned conventional leaving or taking away detection system, even when a part of an object left in the monitoring image becomes invisible by another object, the image of the object left from the partial image of the object left in the monitoring image is displayed. There is known a video monitoring device that detects the presence and highlights a left object on a display screen as a leftover object after a predetermined time has elapsed (see Patent Document 1).

  In addition, as another example of the above-mentioned conventional leaving or taking away detection system, each monitoring imaged until a predetermined time elapses after a foreign object appears in the monitoring area until the foreign object is detected as a leaving object. Based on the image, the images of all the persons who overlap the left-behind are displayed as candidate actors for the left-off, and the possibility that each candidate for the actor has the left-behind will be displayed. A leave monitoring apparatus that displays the degree of coincidence (partial coincidence) with an actor candidate image is also known (see Patent Document 2).

  However, in the video surveillance device described in Patent Document 1, the image of an object that has been left or taken away and the position thereof can be specified, but information that leads to the identification of the actor who has left or taken away is included. I can't get it. Moreover, in the leaving monitoring device described in Patent Document 2, although it is possible to obtain information about the leaving agent, was taken from the appearance of the foreign object until the foreign object is detected as a left object, Image processing must be performed for all monitoring images, and the amount of image processing is enormous. For this reason, it is a problem to reduce the amount of image processing and to provide information about a candidate for an actor etc. to a user more quickly.

JP 2008-104130 A JP 2011-49646 A

  The present invention has been made in view of the above-described conventional problems, and its object is to reduce the amount of image processing when performing leaving or removal detection, and to quickly provide reference information for identifying an actor. That is.

The present invention has a photographing means for photographing the monitoring area, a storage means for saving the photographed image, a leaving or taking away judging section, and a generation time calculating section for calculating the leaving or taking out occurrence time. A leaving or removal detection system that performs processing for identifying a person who is left or removed based on a captured image of the monitoring area, and is a captured image of the monitoring area that is captured from the storage unit at the time of occurrence And an image acquisition processing unit that acquires the image of a surrounding person that is within a predetermined distance range from a foreign object that is a target to be left behind or taken away from the captured image of the monitoring area that was captured at the time of occurrence Leaving log data or leaving log data each including a part, an image of the foreign object to be left or removed, and an image of the surrounding person A log data generation unit to be generated, a desertion or taking away sensing system having a.
In addition, the present invention provides a photographing means for photographing the monitoring area, a storage means for saving the photographed image, a leaving or taking away determination section, and a generation time calculating section for calculating the occurrence time of leaving or taking away. A method for generating a leaving or removal detection record in a removal or removal detection system that performs processing for specifying a person who is left or removed based on a captured image of a monitoring area, from the storage unit, A step of acquiring a photographed image of the monitoring area photographed at the time of occurrence, and from a photographed image obtained in the obtaining step, a surrounding person who is within a predetermined distance range from a foreign object to be left or removed An image extracting step, an image of the foreign object to be left or removed, and the surrounding person extracted in the extracting step Including image and respectively, and generating the left-behind log data or taking away the log data, a desertion or taking away detecting recording method of generating with a.

  ADVANTAGE OF THE INVENTION According to this invention, the amount of image processing at the time of performing leaving or a removal detection can be reduced, and the reference information for specifying the actor of the leaving or the removal can be provided quickly.

It is a block diagram which shows the structure of the leaving or taking away detection system which concerns on one Embodiment of this invention. It is a flowchart which shows the process sequence which produces | generates the leaving log data or the take away log data in the leaving or the take away detection system which concerns on this embodiment. It is a figure which shows an example of the monitoring image of the leaving or carrying out occurrence time in the leaving or taking away detection system which concerns on this embodiment. It is a figure which shows an example of the leaving log data or the leaving log data in the leaving or taking away detection system which concerns on this embodiment. It is a flowchart which shows the process sequence which displays the leaving log data or the leaving log data in the leaving or the taking away detection system which concerns on this embodiment. It is a figure which shows an example of the display screen of the leaving log data or the leaving log data in the leaving or taking away detection system which concerns on this embodiment.

An embodiment of a leaving or leaving detection system (hereinafter simply referred to as the present detection system) of the present invention will be described with reference to the drawings.
FIG. 1 is a block diagram showing the configuration of the present detection system according to an embodiment of the present invention.

  This detection system includes a control unit 100, a camera 20 that captures a monitoring area that is an external device connected to an input unit 16 that is an input side interface (I / F) of the control unit 100, and an instruction to the control unit 100. A display device (monitor) for displaying still images and moving images, which are external devices connected to the input device 22 such as a keyboard for inputting the image and the like, and the output unit 15 that is an output side interface (I / F) of the control unit 100 ) 30.

  Further, the present detection system stores the video of the monitoring area captured from the camera 20 via the input unit 16 with a shooting time for each video frame, for example, storage means such as a hard disk (HDD: Hard Disk Drive). The data storage medium 34 provided with. Here, the video of the monitoring area stored in the data storage medium 34 is a still image of the monitoring area in which each one of the video frames is taken at regular time intervals. Image processing is performed using the image of the video frame as a monitoring image. Hereinafter, in the present embodiment, the term “monitoring image” refers to each video frame that constitutes a video image of a monitoring area.

  The control unit 100 includes a CPU (Central Processing Unit) 10, a ROM (Read Only Memory) (not shown) in which programs and the like are written, RAMs 14a and 14b, and the input unit 16 described above. An output unit 15 and a data bus 26 for connecting the above-described units are provided.

  The RAM 14a has an extraction target range (peripheral area) when the surrounding person extraction unit 43 (described later) extracts an image of a left-behind or left-behind person on the monitoring image as various settings necessary for program execution. ) And the relationship degree calculation unit 44 (described later) uses a distance specific gravity, an overlap specific gravity, and the like to calculate a numerical value (degree of relation) indicating the degree of engagement between each neighboring person and a left-behind object or a left-behind object. And information such as the specific gravity of the person.

  The RAM 14b is a storage unit that stores images, and has three storage areas, an input area, a display area, and a holding area, depending on the type of image to be stored. This input area is an area for temporarily storing an image (monitoring image) of the monitoring area acquired from the camera 20 via the input unit 16. The holding area is an area for storing a background image or reading and storing a monitoring image from the data storage medium 34 during image processing. The display area is an image of a display image to be displayed on the monitor 30. An area for storing data.

  The control unit 100 controls the entire detection system, and as a function realizing unit realized by the CPU 10 executing a program, a left-behind / removal detection unit 50, an image acquisition processing unit 42, a surrounding person extraction unit 43, , A relationship degree calculating unit 44, a log data creating unit 45, and a log data collating unit 46. The computer program can be stored in any computer-readable storage medium.

The abandoned / taken away detection unit 50 includes a leftover / takeaway determination unit 52 and an occurrence time calculation unit 54.
The abandonment / removal determination unit 52 generates a luminance difference image between the current monitoring image and the background image input from the camera 20 and detects the appearance or disappearance of a foreign object in the monitoring area based on the generated difference image. Sometimes, time measurement is started by a timer (not shown), and it is determined whether or not an object has been left or removed depending on whether or not the detection state has continued for a predetermined time.

  The occurrence time calculating unit 54 subtracts the predetermined time from the time when the leaving or taking away determination unit 52 determines that the leaving or taking away has occurred, and the leaving or The time of occurrence of take-away (the time of appearance or disappearance of a foreign object) is calculated.

  The image acquisition processing unit 42 is a monitoring image at the occurrence time of leaving or taking away calculated by the occurrence time calculation unit 54, or within a certain time between the occurrence times when the occurrence time includes a calculation error (for example, A predetermined number of monitoring images taken continuously (within the calculation error) are read from the data storage medium 34 and stored in the holding area of the RAM 14b. Here, the predetermined number can be about several sheets according to the calculation error of the generation time of the leaving or taking away calculated by the generation time calculation unit 54.

For the predetermined number of monitoring images, for example, a video recording server that continuously records video for a predetermined time is used as the data storage medium 34, and video after a specified time is specified from the recorded video. It can also be obtained by using a so-called “retro recording function” that extracts only the time.
Also, instead of reading the monitoring image stored in the data storage medium 34 to the holding area of the RAM 14b as in the present embodiment, the leaving / determining determination unit 52 detects a foreign object and starts measuring the time of the timer. In this case, the storage destination of the video imaged by the camera 20 can be switched from the data storage medium 34 to the RAM 14b and directly stored in the holding area of the RAM 14b.

  The surrounding person extraction unit 43 uses a monitoring object stored at the holding area of the RAM 14b or a monitoring image at the time of occurrence of removal or a predetermined number of monitoring images taken within a certain time with the generation time interposed, Extract images of people around your belongings.

Here, in the case where the image of the surrounding person is extracted from the monitoring image at the time of occurrence, when the image of the left object or the left object is not reflected in the monitoring image, the time closest to the shooting time of the monitoring image is displayed. It is assumed that a monitoring image which is a monitoring image and shows the left behind or left behind is read from the data storage medium 34 and images of the surrounding persons are extracted.
In addition, when extracting images of surrounding people from the predetermined number of monitoring images, from the frame of the past, the oldest frame in which a leftover is shown, or the video of the newest frame in which a leftover is shown. In addition, an image of a person around the left or left item is extracted.

By the way, it is desirable that the range (peripheral area) in which the peripheral person extraction unit 43 extracts the peripheral person is a certain range centered on the left behind or the left behind.
However, on the monitoring image, an object far from the camera 20 is small and a close object is large, so if the surrounding person extraction range on the monitoring image, for example, a pixel range (pixel area) is constant, the object is far from the camera 20. The peripheral area is large for a position image (for example, an image near the upper end of the monitoring image), and the peripheral area is small for an image near the camera 20 (for example, an image near the lower end of the monitoring image).

For this reason, in the present embodiment, peripheral area information in which the size of the pixel area from which the peripheral person is extracted is determined according to the distance from the camera 20 to the object is stored in the RAM 14a in advance.
For example, the peripheral area information is centered on the center pixel so that the pixel area gradually decreases as it goes to the top of the screen according to the number of pixels from the lower end of the screen to the center pixel of the left behind image or the left behind image. The radius of the circular area to be used can be information represented by the number of pixels.

  The degree-of-relation calculation unit 44 is an information generation unit that generates relevance information indicating the degree of association between a nearby person and a left-behind object or a left-behind object. A numerical value (relationship degree) as relevance information is determined by a predetermined method described later. The degree of this relationship is based on, for example, the distance L from the monitoring image at the time of occurrence to the surrounding person and the left or left thing, the overlapping area S of the surrounding person and the left or left thing, or a predetermined direction. The direction θ of the surrounding person to be calculated is calculated and determined from the distance L, area S, or direction θ by a predetermined method described later.

  In the present embodiment, the relevance information is calculated as a numerical value (relationship degree). However, the present invention is not limited to this. For example, the relevance information is determined according to the distance L, the area S, or the direction θ. The rank (A rank, B rank, etc.) can also be used as the relevance information.

  The log data creation unit 45 is a leftover log data in which an image of a leftover object, an image of a person in the vicinity of the leftover object, and a degree of relationship for each of the surrounding persons are recorded for each leftover object, and an image of the leftover object. Taking away log data in which the images of the surrounding persons and the degree of relationship between the surrounding persons are recorded for each thing to be taken is generated and stored in the data storage medium 34.

  The log data collating unit 46 collates the left log data stored in the data storage medium 34 and the image of the left object included in the removed log data with the image of the left object, and the same object is left behind and removed. The left log data and the left log data are specified. In addition, the neighboring persons recorded in the specified leaving log data and the leaving log data are compared with each other, and the same person recorded as the surrounding person in both the leaving log data and the leaving log data is specified. .

This detection system has the above-described configuration, and stores the video of the monitoring area captured by the camera 20 in the data storage medium 34 via the input unit 16 after the user turns on the power.
At this time, the abandonment / removal determination unit 52 generates a luminance difference image with respect to the background image for each monitoring image, and detects the appearance or disappearance of a foreign object based on the difference image. Then, after a predetermined time has elapsed from the detection, it is determined that the foreign object has been left or taken away.

  Next, the occurrence time calculation unit 54 calculates the occurrence time of the leaving or taking away (refers to the time when the appearance or disappearance of a foreign object is detected), and the image acquisition processing unit 42 Several monitoring images taken continuously within a certain time (for example, within the generation time calculation error) sandwiching the occurrence time are read from the data storage medium 34 and stored in the holding area of the RAM 14b.

  Subsequently, the peripheral person extraction unit 43 extracts an image of a peripheral person of a left-off object or a left-off object (referred to as a foreign object to be left or removed) from the monitoring image stored in the holding area of the RAM 14b. To do. In addition, the relationship degree calculation unit 44 determines a numerical value (degree of relationship) indicating the degree of the relationship between the left object or the left object for each neighboring person by a method described later.

  Then, the left log data or the left log data including the image of the left object or the left object, the extracted image of the surrounding person, and the degree of relationship for each surrounding person are generated, and the data storage medium 34 is generated. To remember.

  In other words, in the present detection system, the leaving log data or the leaving log data can be generated only from the monitoring image at the time of occurrence of the leaving or taking away or a few monitoring images before and after the occurrence time. Unlike a monitoring device, an enormous amount of image processing is not required. As a result, this detection system can promptly provide the user with the image of the object left behind or the object around the object and the degree of relationship as reference information for identifying the person who has left or removed. .

  The generated leave log data and take away log data are read from the data storage medium 34 and displayed on the monitor 30 in accordance with an instruction from the user via the input device 22 or the like. At that time, the log data collating unit 46 performs collation between the left-out log data and the left-behind thing recorded in the left-out log data, and collation between neighboring persons. Then, the leaving log data and the leaving log data for the same object are associated with each other and displayed on the monitor 30.

  Next, a processing procedure for generating left log data or take away log data in the detection system will be described with reference to the flowchart of FIG. Here, it is assumed that the background image of the monitoring area is previously captured by the camera 20 and stored in the holding area of the RAM 14b.

  First, the CPU 10 of the control unit 100 inputs a video of a monitoring area (that is, a monitoring image) captured by the camera 20 via the input unit 16 (S101). Here, the CPU 10 stores the input video in the data storage medium 34 with a shooting time for each video frame.

  In addition, the abandonment / removal determination unit 52 generates a difference image of luminance with respect to the background image stored in the holding area of the RAM 14b for each input monitoring image, and appearance of a foreign object based on the generated difference image. Alternatively, disappearance is detected (S102). At this time, the abandoned / removal determination unit 52 starts a timer (not shown) and measures the duration of the detection state of the foreign object.

Next, the abandonment / removal determination unit 52 determines whether the abandonment / removal has occurred or not, depending on whether a predetermined time has elapsed for the detection state (S103).
When it is determined that the leaving or taking away has occurred (S103, Yes), the occurrence time calculating unit 54 subtracts the predetermined time from the time at which it has been determined that leaving or taking away has occurred. The occurrence time of leaving or taking away (that is, the time when the appearance or disappearance of a foreign object is detected) is calculated (S104).
On the other hand, when leaving or taking away has not occurred (No in S103), the process returns to the image input in step S101.

  Next, the image acquisition processing unit 42 reads out from the data storage medium 34 a monitoring image at the time of occurrence or a predetermined number of monitoring images taken continuously within a certain time with the time of occurrence (S105). The surrounding person extraction unit 43 extracts an image of a person who is left behind or left behind (a surrounding person) from the read monitoring image (S106). Note that when extracting the surrounding person, the surrounding person extracting unit 43 determines a range in which the surrounding person is extracted from the monitoring image based on the surrounding area information stored in the RAM 14a.

  Subsequently, the relationship degree calculation unit 44 calculates the degree of relationship between each of the neighboring persons with the left-behind object or the left-behind object (S107). Extract the image of the object, and leave log data including the image of the object left, the image of the person around the object, and the degree of relationship between each person, or the image of the object and the area around the object After taking away log data including the image of the person and the degree of relationship between the surrounding persons is generated and stored in the data storage medium 34 (S108), the process returns to the image input in step S101.

  Here, the degree of relationship r calculated in step S107 is the distance L between each of the neighboring persons and the left object or left object, and the overlapping area S of the left object or left object (for example, the person and the left object or left object) The number of pixels that overlap with the leftovers and the orientation θ of the person (for example, since a monitoring image of a certain time sandwiching the occurrence time is stored, it is estimated from several stored monitoring images. The angle of the direction of the left-behind or left-behind object seen from the person with respect to the direction of movement of the person, or the angle of the direction of the left-behind or left-behind object seen from the person with respect to the direction of the person's face or line of sight) From this, it can be determined as follows.

That is, left behind thereof or carried away was a peripheral person existing in the surrounding area and n (1 to N) of the relationship of h _n of each peripheral persons n, the distance L _n to the left-behind object or taking away _thereof, desertion product Alternatively, it is calculated by the following equation from the overlapping area S _n with the carry-away object and the angle θ _n representing the direction of the person.

Here, α _L , α _S , and α _θ are weighting factors for the distance L, the overlapping area S, and the human orientation θ, respectively. Then, after calculating the degree of relationship h _n with respect to all peripheral person, the relationship degree r _n of each peripheral persons n, is calculated by the following equation.

The above weighting coefficients are determined in advance by, for example, experiments and stored in the RAM 14a as “distance specific gravity α _L ”, “overlap specific gravity α _S ”, and “human orientation specific gravity α _θ ”. be able to. For example, it may be necessary to reduce the distance specific gravity in a quiet area because the number of persons is small, and conversely, in a congested area, the number of persons may be increased, so the distance specific gravity may be increased. .

  In addition, the behavior pattern of the surrounding person, for example, the movement of the hand and the body direction is analyzed, and the angle φ between the moving direction of the hand and the body and the direction of the left behind or removed thing viewed from the surrounding person. May be reflected in the calculation of the relationship degree r. Thereby, the calculation accuracy of the relationship degree r can be improved.

Next, examples of leave log data and take away log data created by the detection system will be described.
FIG. 3 is a diagram illustrating an example of a monitoring image of the time of leaving or taking away in the detection system.
In FIG. 3, there is a bag 60 that is left behind in the upper right part of the center, and a peripheral area 62 that is a range for extracting a peripheral person is indicated by a dotted circle centered on the bag. Three persons 64a, 64b, and 64c are shown in the peripheral area 62, and these persons 64a, 64b, and 64c are specified as the peripheral persons. Then, the images of the persons 64a, 64b, and 64c are extracted as the images of the surrounding persons from the monitoring image.

  FIG. 4 is a diagram illustrating an example of the leaving log data and the leaving log data in the detection system. 4A shows an example of left log data, and FIG. 4B shows an example of left log data.

The leaving log data (FIG. 4A) is configured in, for example, a table format, and information about each leaving thing is recorded for each line. In the leftmost column of each row, an image of the left behind is recorded, and on the right side, the image of the surrounding person and the degree of relationship are recorded vertically.
4A is the left log data created based on the monitoring image shown in FIG. 3, and the three persons extracted from the peripheral area in FIG. Images 64a, 64b, and 64c are recorded as images of surrounding persons.

  The take away log data (FIG. 4B) has the same configuration as the leave log data shown in FIG. 4A. In the leftmost column, an image of the take away object is recorded instead of the image of the leave article. ing.

Next, a processing procedure for displaying the left log data or the left log data on the monitor 30 will be described with reference to the flowchart shown in FIG.
When the process is started, the CPU 10 of the control unit 100 displays on the monitor 30 the leftover log data stored in the data storage medium 34 and the leftover image and the leftover image recorded in the left away log data (S201). .

  With respect to the left behind image and the left behind image displayed on the monitor 30, the user can use the input device 22 to select a left behind item or a left behind item whose detailed information is to be displayed.

  When the user selects a leftover item or a leftover item, the log data collating unit 46 displays an image of the leftover item or leftover item (user selected item) selected by the user and a leftover image recorded in the leftover log data. Then, the left-side log data and / or the left-out log data for the user-selected item are specified by collating with the left-behind object image recorded in the removed log data (S202).

  Next, whether there is left log data and left log data that make the user selection a left item and left item, respectively (whether there is left log data and left log data related to the user selection) (S203), and when there is such a leave log data and take away log data (S203, Yes), the peripheral person recorded in the leave log data and the periphery recorded in the leave log data The person is collated (S204), the contents of the leaving log data and the carried away log data and the collation result of the surrounding person are displayed on the monitor 30 (S205), and the process is terminated.

  As a result, for example, a person who forgot his / her baggage came to pick up his / her baggage from the displayed log data and the image of the surrounding person included in the log data. It may be used as reference information for detecting crimes by determining whether the actor is the same) or a person who has nothing to do with the baggage (the left actor is different from the removed actor) it can. In addition, since the degree of relationship between each peripheral person displayed on the monitor 30 can be used for this determination, it is possible to accurately determine the coincidence / disagreement between the leaving actor and the leaving actor.

  On the other hand, in step S203, when there is no left-out log data and left-out log data for the user-selected items as left-and-left items (S203, No), the user-selected items specified in step S202 are identified. The contents of the leaving log data or the leaving log data are displayed on the monitor 30 (S206), and the process is terminated.

  Thus, for example, when the person who has lost the item comes to pick up the lost item, the user displays the leaving log data about the lost item, and the image of the displayed peripheral person (candidate for the leaving agent) and the relationship thereof. The degree can be used as reference information when returning a lost item.

FIG. 6 is a diagram illustrating an example of a display screen of left log data or left log data in the detection system.
FIG. 6A is an example of a leftover image and a leftover image displayed on the monitor 30, and FIG. 6B is a display example of the leftover log data and the left away log data for the selected leftover or leftover item. It is.

  When the above-described display processing is started, a leftover image and a leftover object included in the log data are obtained based on the leftover log data and the left away log data shown in FIGS. 4A and 4B, for example, in step S201 of FIG. The images are displayed side by side on the monitor 30 as shown in FIG. 6A.

  For example, when the user selects a left object or a left object surrounded by a dotted circle in the figure on the display screen, the left log data or the left log data for the user selection is displayed on the monitor 30. (FIG. 6B). Here, with respect to the user selection surrounded by a dotted circle in FIG. 6A, leaving log data in which the user selection is left behind is recorded in the bottom line of FIG. 4A, and the user selection is taken away. Taken away log data is recorded in the uppermost line (excluding the title line) in FIG. 4B.

  Therefore, the log data collating unit 46 collates the image of the user-selected item with the image of the left-behind item and the left-behind item in step S203 of FIG. As the data and the take-away log data, the left-behind log data in the bottom row in FIG. 4A and the take-out log data in the top row in FIG. 4B are specified. As a result, these log data are associated and displayed side by side on the monitor 30 (FIG. 6B).

  In addition, the collation result of the peripheral person in step S204 in FIG. 5 is obtained by, for example, displaying a line (() on the display screen shown in FIG. 6B. It can be displayed by tying them (not shown). In this case, in the example of FIG. 6B, a person with a degree of relationship of 80% in the leaving log data and a person with a degree of relation of 100% in the leaving log data are connected by a line and displayed.

  As described above, in the present embodiment, only the monitoring image at the time when the appearance or disappearance of the foreign object is detected, or only a few monitoring images taken before and after the monitoring image, the left-handed object or the person around the left-handed object is left behind. An image is extracted and stored, and a numerical value (degree of relationship) indicating the degree of the relationship between each neighboring person and the left-behind or left-out thing is calculated and stored. For this reason, this detection system does not perform an enormous amount of image processing as in the prior art, and provides reference information (i.e., a peripheral person image and a degree of relationship) for identifying a leaving person or a leaving person. It can be quickly provided to the user.

10 ... CPU, 14a, 14b ... RAM, 15 ... output unit, 16 ... input unit, 20 ... camera, 22 ... input device, 26 ... data bus, 30. ... Monitor 34... Data storage medium 42... Image acquisition processing unit 43 .. peripheral person extraction unit 44 .. relationship degree calculation unit 45. ... Log data collating unit, 50... Left and removed detection unit, 52... Left and removed judgment unit, 54.

Claims (5)

An image capturing means for capturing the monitoring area; a storage means for storing the captured image; a leaving or leaving determination section; and an occurrence time calculating section for calculating an occurrence time of leaving or taking away. A leave or removal detection system that performs processing for identifying a person who is left or removed based on a captured image,
An image acquisition processing unit for acquiring a captured image of the monitoring area captured at the generation time from the storage unit;
A surrounding person extraction unit that extracts an image of a surrounding person within a predetermined distance range from a foreign object that is a target to be left behind or taken away from a captured image of the monitoring area photographed at the time of occurrence;
A log data generation unit for generating left log data or left log data, each including an image of the foreign object and an image of the surrounding person that is the target of the left or left;
A leave or take away detection system. In the leaving or taking away detection system according to claim 1,
In the captured image of the monitoring area photographed at the time of occurrence, calculate a distance from the surrounding person to the foreign object, an overlapping area between the surrounding person and the foreign object, or an angle representing the direction of the surrounding person, From the distance, area, or angle, comprising information generation means for generating relevance information indicating the degree of association between the surrounding person and the foreign object,
The leaving log data or the leaving log data further includes a leaving or leaving detection system including the relevance information of the surrounding person. In the leaving or taking away detection system according to claim 1 or 2,
A log data collation unit that collates the foreign object image included in the leaving log data with the foreign object image included in the removed log data;
A left-off or left-behind detection system that displays the left log data and the left log data in association with each other on the basis of the result of the collation and displays the left log data and the left log data as images of foreign objects related to left and left. In the leaving or taking away detection system according to claim 3,
The log data collating unit includes, for the leaving log data and the leaving log data including images of the same object as the images of the leaving and taking away foreign objects, and the surrounding person images included in the leaving log data and the Check the image of the person in the surrounding area included in the log data,
A leave or take-away detection system that displays a display unit in association with neighboring persons who are the same person based on the result of the collation. An image capturing means for capturing the monitoring area; a storage means for storing the captured image; a leaving or leaving determination section; and an occurrence time calculating section for calculating an occurrence time of leaving or taking away. A method for generating a leaving or removal detection record in a leaving or removal detection system that performs processing for identifying a person who is leaving or taking away based on a captured image,
Acquiring from the storage means a captured image of the monitoring area captured at the occurrence time;
A step of extracting, from the captured image acquired in the step of acquiring, an image of a surrounding person who is within a predetermined distance range from a foreign object to be left or removed; and
Generating a leaving log data or a leaving log data, each including an image of the foreign object to be left or taken away and an image of the surrounding person extracted in the extracting step;
A method for generating a leaving or taking away detection record.

Images