JP6605104B2 - Suspicious person detection device and program - Google Patents

Description

  The present invention relates to a suspicious person detection apparatus and program, and more particularly to a suspicious person detection apparatus and program for detecting a suspicious person in a predetermined area.

  As a technique that can be applied to prevent the occurrence of a crime against a predetermined person in an area where an unspecified person enters and exits such as an open space, a park, etc., Patent Document 1 discloses that the object by photographing the target area. An imaging unit that acquires image information of a region in real time, a detection unit that detects a position of a person existing in the target region at a predetermined time interval based on the image information acquired by the imaging unit, and the detection unit The detected person is either a defender (a person who can prevent crimes), a defender (a person who wants to protect against crime), or an indefinite person (a person other than the above-mentioned defender and the defender). A first distance that is a distance between the defender and the defender based on a position of the person detected by the detection means and a determination result by the specification means; A distance deriving means for deriving a second distance, which is a distance between the indefinite person and the defender, and the defensive based on a ratio of the first distance and the second distance derived by the distance deriving means. A calculation means for calculating a crime risk value indicating the level of crime risk caused by the indeterminate person for a person, and a crime risk value calculated by the calculation means are used to execute a predetermined process for preventing crime And a crime prevention apparatus including a process execution means for performing the above-described process.

  Patent Document 2 discloses a photographing unit that captures image information of a target region in real time by photographing the target region, and a person existing in the target region based on the image information acquired by the photographing unit. Detecting means for detecting the position of the robot at predetermined time intervals, specifying means for specifying whether or not the person detected by the detecting means is a predetermined defender and an indefinite person excluding defenders, and the specifying Estimating means for estimating, based on the image information, characteristic information indicating personal characteristics including at least the age of a person identified as indeterminate by the means, and based on the characteristic information estimated by the estimating means A derivation means for deriving a crime execution rate, which is a probability that a person will execute a crime within a lifetime after the age included in the characteristic information, and a crime actuality derived by the derivation means Based on the rate, calculating means for calculating a crime risk value indicating the level of crime risk caused by the indefinite person to the defender, and using the crime risk value calculated by the calculating means to prevent crime A crime prevention device comprising: a process execution means for executing a predetermined process as:

Patent No. 5001808 JP 2010-250775 A

  By the way, in the technology disclosed in Patent Document 1 and Patent Document 2, an IC tag, a device capable of specifying its own position by GPS (Global Positioning Systems), and a wireless device are provided for the defender and the defender. A transmitter or the like is carried, and the positions of the defender and the defender are individually detected using these. And in these techniques, the person except the said defender and the defender is identified as an indefinite person from the persons detected based on the image information acquired by the imaging means. Note that the IC tag is sometimes referred to as an RFID tag, a wireless tag, an ID tag, an electromagnetic induction tag, or the like. In this specification, these are collectively referred to as an IC tag.

  As described above, in the technologies disclosed in Patent Document 1 and Patent Document 2, in order to detect defenders, defenders, and unidentified persons, the positions of the defenders and the defenders are identified. There was a problem that it was necessary to equip the device to do.

  In the techniques disclosed in Patent Document 1 and Patent Document 2 described above, persons other than the defender and the defender are referred to as “unidentified persons” as persons who cannot be identified, but in the present invention, The expression “suspicious person” is used for the purpose of detecting a person who is more likely to commit a crime than the unidentified person.

  The present invention has been made to solve the above-mentioned problems, and a suspicious person detecting device capable of detecting a suspicious person with high accuracy without requiring special equipment for a defender and a defender, and The purpose is to provide a program.

In order to achieve the above-described object, the suspicious person detection device of the present invention is configured to detect the behavior state of a person existing in the target area based on a plurality of captured images obtained by a photographing unit that continuously photographs the target area. A derivation means for deriving the physical quantity to be shown, an arithmetic means for computing an average value of the physical quantity derived by the derivation means within a predetermined period or with a predetermined number of persons, and a person existing in the target area based of the difference from the physical quantity of said average value calculated by the calculation means, a first specifying means for specifying the suspiciousness degree of the person present in the target area, the suspiciousness degree which is specified by the first specifying means based on a detection means for detecting a suspicious person from the person present in the target area, among the people existing in the captured image, in humans except suspicious individual detected by said detecting means, and the non And the defensive person to person close to a predetermined order from the nearest person who is provided with a second specifying means for specifying the others as defensive travelers,.

  According to the suspicious person detection device of the present invention, the physical quantity indicating the action state of the person existing in the target area based on the plurality of captured images obtained by the imaging unit that continuously images the target area by the deriving unit. Is calculated, and the arithmetic means calculates the average value of the physical quantity derived by the derivation means within a predetermined period or for a predetermined number of people. In addition, as said imaging | photography means, various cameras, such as a monochrome camera, a color camera, and an infrared camera which can image | photograph at least one of a moving image and a still image, are applicable.

And in this invention, based on the difference from the average value calculated by the said calculating means of the said physical quantity of the person who exists in the said target area by the 1st specific means, the suspicious degree of the person who exists in the said target area is The suspicious person is detected from the person existing in the target area based on the suspicious degree specified by the first specifying means and specified by the detecting means.

  That is, many of the conventional methods for detecting a suspicious person detect a person who is in an area where a person should not exist, but this method cannot be used in places where access is not restricted. In addition, there is a technology that detects specific movements set in advance as suspicious behavior, such as squatting and rampage. However, this technology cannot handle all suspicious behavior, and has not been considered in advance. When an action is a target, it is necessary to newly add a detection logic.

  On the other hand, a study that summarizes criminal viewpoints based on investigations into criminal planners (“Urban Crime Research Center, Survey Report on the Actual State of Burglary (4) Realities for the Invasion of Burglars in Houses, JUSRI Report 12, 1997 ”, Mamoru Amemiya et al., Examination of the effectiveness of crime prevention environment design from the viewpoint of criminals-Questionnaire survey report for detainees nationwide-, Urban Planning Report Collection No. 8, 2009 "), police officers' job question materials (" Theft Crime Investigation Study Group, Theft Crime Investigation Practice, Sakai Publishing, 1969 "," Police Practice Study Group, Job Question, Tachibana Shobo, 2007 ", etc. As is clear from the above, criminal planners are said to have unnatural behavior and behavior. In addition, as the behavior, a behavior related to the object color of the object, a behavior for avoiding surrounding eyes, and the like can be seen.

  In consideration of these, in the present invention, suspicious behavior is quantitatively detected by statistically detecting unnatural behavior that does not fit in the field in the target region.

  As described above, according to the suspicious person detection device of the present invention, the physical quantity indicating the action state of the person existing in the target region based on the plurality of captured images obtained by the imaging unit that continuously images the target region. And calculating an average value of the derived physical quantity within a predetermined period or with a predetermined number of persons, and calculating a difference from the calculated average value of the physical quantity of a person existing in the target area. Based on the suspicious degree of the person existing in the target area, and the suspicious person is detected from the person existing in the target area based on the specified suspicious degree. Therefore, it is possible to detect a suspicious person with high accuracy without requiring special equipment.

  In the present invention, the calculating means calculates the average value using a physical quantity excluding the physical quantity derived by the deriving means based on the captured image obtained at the timing when the crime occurred in the target area. May be. Thereby, a suspicious person can be detected with higher accuracy.

  Further, according to the present invention, the calculating means further calculates a standard deviation of the physical quantity used for calculating the average value in addition to the average value, and the detecting means determines the suspicious degree based on the standard deviation. A person who exceeds the threshold may be detected as a suspicious person. Thereby, a suspicious person can be detected with higher accuracy.

  In the present invention, the physical quantity is at least one of movement-related information related to a movement state of a person existing in the target area and stay-related information related to a stay state of a person existing in the target area in the target area. It may be a thing. Thereby, the suspicious degree according to the actual condition can be easily specified.

  Further, according to the present invention, the deriving means derives at least one of a movement speed and a movement direction of a person existing in the target area as the movement related information, and the person existing in the target area as the stay related information. At least one of a stay time in the target area, a stay time in a specific range in the target area, and a stay position in the target area may be derived. Thereby, the suspicious degree according to the actual condition can be easily specified.

On the other hand, in order to achieve the above object, the program according to the present invention allows a computer to perform an action of a person existing in a target area based on a plurality of captured images obtained by imaging means for continuously capturing the target area. Deriving means for deriving a physical quantity indicating a state, computing means for computing an average value of the physical quantity derived by the deriving means within a predetermined period or with a predetermined number of persons, and existing in the target area Based on the difference from the average value calculated by the calculation means of the physical quantity of the person to be identified, the first identification means for identifying the suspicious degree of the person existing in the target area, and the first identification means based on the suspiciousness degree, divided detecting means for detecting a suspicious person from the person present in the target area, among the people existing in the captured image, the suspicious individual detected by said detecting means In humans, and human closer to a predetermined order from the nearest person to the suspicious person and the defensive person is intended to function with others as a second specifying means for specifying as a defensive person.

  Therefore, according to the program of the present invention, the computer can be caused to act similarly to the suspicious person detection device of the present invention. It is possible to detect a suspicious person with high accuracy without obliging to wear a special equipment.

  According to the present invention, a physical quantity indicating a behavior state of a person existing in the target area is derived based on a plurality of captured images obtained by an imaging unit that continuously captures the target area, and the physical quantity of the derived physical quantity is derived. Calculate an average value within a predetermined period or with a predetermined number of people, and exist in the target area based on a difference from the calculated average value of the physical quantity of the person existing in the target area Since the suspicious person is identified from the person existing in the target area based on the specified suspicious degree, the suspicious degree of the person is identified. The effect that a suspicious person can be detected with high accuracy is obtained.

It is a mimetic diagram showing the whole crime prevention system composition concerning an embodiment. It is a perspective view which shows the arrangement | positioning relationship etc. of each part which comprises the crime prevention system which concerns on embodiment. It is a block diagram which shows the principal part structure of the electric system of the crime prevention apparatus which concerns on embodiment. It is a schematic diagram which shows the main memory content of the secondary memory | storage part with which the crime prevention apparatus which concerns on embodiment was equipped. It is a schematic diagram which shows the structure of the positional infomation database which concerns on embodiment. It is a schematic diagram which shows the structure of the obstruction information database which concerns on embodiment. It is a schematic diagram which shows the structure of the crime risk value database which concerns on embodiment. It is a graph which shows an example of the probability distribution D (x) of the walking speed which concerns on embodiment. It is a graph which shows an example of evaluation value E (x) regarding walking speed concerning an embodiment. It is a schematic diagram which shows an example of a structure of the fault tree with which it uses for description of the detection method of a suspicious person which concerns on embodiment. It is a schematic diagram with which it uses for description of the computing equation which calculates the crime risk value which concerns on embodiment. It is a graph with which it uses for description of the validity of the computing equation which calculates the crime risk value which concerns on embodiment. It is a flowchart which shows the flow of a process of the action distribution measurement program which concerns on embodiment. It is a flowchart which shows the flow of a process of the crime prevention program which concerns on embodiment.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. Here, the case where the present invention is applied to a crime prevention system for monitoring a park in order to prevent the occurrence of a crime in the park will be described.

  First, an overall configuration of a crime prevention system 10 to which the present invention is applied will be described with reference to FIG.

  As shown in the figure, the crime prevention system 10 according to the present embodiment includes a crime prevention device 20 that plays a central role in the system 10 and an area to be monitored in the system 10 (here, a park, which is described below). And a plurality of cameras 40 that take images of the inside of the surveillance area, and each camera 40 is electrically connected to the crime prevention apparatus 20.

  On the other hand, in the crime prevention system 10 according to the present embodiment, as shown in FIG. 2, the cameras 40 are located at a plurality of positions (four locations in the present embodiment) around the monitoring target area 60. The target region 60 is provided at a position and a state where the target region 60 can be photographed. In addition, since the camera 40 is for monitoring the monitoring target area 60, it is preferable that the camera 40 is provided in a state where no blind spots are generated in the entire monitoring target area 60. The shape of the monitoring target area, The number of installations and installation positions are determined according to the state of obstacles, usage, etc. Further, in the crime prevention system 10 according to the present embodiment, a digital camera capable of shooting a moving image in color is applied as the camera 40, and the image of the monitoring target region 60 is captured by the camera 40. The image information shown can be acquired.

  Next, with reference to FIG. 3, the structure of the main part of the electrical system of the crime prevention apparatus 20 having a particularly important role in the crime prevention system 10 will be described.

  As shown in the figure, the crime prevention apparatus 20 according to the present embodiment includes a CPU (Central Processing Unit) 20A that controls the entire operation of the crime prevention apparatus 20, and a work area when the CPU 20A executes various processing programs. A RAM (Random Access Memory) 20B used, a ROM (Read Only Memory) 20C in which various control programs, various parameters, and the like are stored in advance, and a secondary storage unit (storage means used for storing various information) Here, the hard disk device 20D, a keyboard 20E used for inputting various information, a display 20F used for displaying various information, and an external I / F (interface) 20G, and these parts are electrically connected by a system bus BUS. They are mutually connected.

  Accordingly, the CPU 20A accesses the RAM 20B, ROM 20C, and secondary storage unit 20D, acquires various input information via the keyboard 20E, displays various information on the display 20F, and external devices via the external I / F 20G. Various signals can be exchanged between them. Each camera 40 described above is electrically connected to the crime prevention apparatus 20 via the external I / F 20G.

  On the other hand, FIG. 4 schematically shows the main storage contents of the secondary storage unit 20D provided in the crime prevention apparatus 20.

  As shown in the figure, the secondary storage unit 20D stores a database area DB for storing various databases, a control program for controlling the crime prevention apparatus 20, a program for performing various processes, and the like. A program area PG is provided.

  The database area DB includes a position information database DB1, an obstacle information database DB2, and a crime risk value database DB3. Hereinafter, the configuration of each database will be described in detail.

  As shown in FIG. 5 as an example, the position information database DB1 is configured so that information on camera ID (Identification), pixel position, and position coordinates is stored for each camera 40.

  Note that the camera ID is given in advance as different information (information combining alphabets and numbers in the present embodiment) for each camera in order to identify the camera 40.

  By the way, in the crime prevention system 10 according to the present embodiment, each camera 40 is fixed in the shooting range (view angle) within a predetermined range by fixing the shooting direction and the enlargement ratio, and the state In the XY coordinate system, the pixel position by the solid-state image sensor incorporated in each camera 40 and the corresponding base point position when the monitoring target region 60 is viewed in plan is corresponding to the pixel position. Are stored in association with each other, and based on this information, the position of a person existing in the image of the monitoring target area 60 indicated by the image information obtained from the camera 40 is specified.

  The pixel position and the position coordinate in the position information database DB1 correspond to the pixel position by the solid-state image sensor and the position coordinate in the XY coordinate system. The position of a person existing in the displayed image can be specified.

  In the position information database DB1 according to the present embodiment, as shown in FIG. 5, a pair of diagonal coordinates of a rectangular pixel area (imaging area) in the solid-state imaging device is used as the pixel position. The position of the person in the XY coordinate system is specified by applying the value and referring to the position coordinates corresponding to the pixel area in which the person's area existing most in the captured image is included. As described above, in the present embodiment, information having a predetermined range is applied as the pixel position. However, the present invention is not limited to this, and for example, one point of the pixel region in the solid-state imaging device is used as the pixel position. And the position coordinates corresponding to the pixel position closest to the predetermined position (for example, the center of gravity of the whole body, the center position of the head, etc.) of the person in the captured image is set as the XY of the person. Needless to say, it is possible to adopt a form specified as a position in the coordinate system.

  On the other hand, as shown in FIG. 6 as an example, the obstacle information database DB2 includes information on the obstacle ID (Identification), the area, and the type, which are present in the monitoring target area 60 and prevent the person from walking. It is configured to be stored for each of the obstacles.

  Note that the obstacle ID is different information for each obstacle (this embodiment) in order to identify the obstacle existing in the monitoring target area 60 (an obstacle 62 implanted as shown in FIG. 2 as an example). Then, the information is given in advance as information combining alphabets and numerals.

  The area is information indicating an area (position) where a corresponding obstacle exists. As shown in the figure, in the obstacle information database DB2 according to the present embodiment, the area is represented by a pair of diagonal coordinate values of a circumscribed rectangular frame of the corresponding obstacle by the XY coordinate system. However, the present invention is not limited to this, and it is needless to say that other forms such as a form represented by coordinate values of the contour position of the corresponding obstacle in the XY coordinate system can be used.

  Moreover, the said classification is information which shows classifications, such as implantation, a fountain, and a heel, of a corresponding obstacle.

On the other hand, as shown in FIG. 7 as an example, the crime risk value database DB3 stores crime risk values indicating the level of crime risk caused by a suspicious person (to be described later) with respect to a defender (to be described later) in chronological order. It is configured as follows. In the figure,..., T _x , t _{x + 1} , t _{x + 2} ,... Represent crime risk value derivation times, S ₁ , S ₂ ,. ₁ , V ₂ ,... Represent different defenders.

  By the way, normally, there is a certain range in the behavior state of a person in normal times. For example, as shown in FIG. 8 as an example, the moving speed of the pedestrian is in the range of about 0.5 (m / s) to about 2.3 (m / s). The frequency of pedestrian movement speed is close to the normal distribution (see, for example, “Yohei Inukai et al., Development of a multi-agent simulator for evacuation behavior based on measurement, Japan Society of Civil Engineers, Applied Mechanics, Vol. 8, 2005”). .)

  On the other hand, the action state when the crime planner executes the crime is likely to deviate from the range of the normal action state.

  Therefore, in the crime prevention system 10 according to the present embodiment, a physical quantity (hereinafter, also referred to as “normal physical quantity”) indicating a normal behavior state of a person existing in the monitoring target area 60 is measured in advance. A person taking an action deviating from the normal physical quantity is detected as a suspicious person.

  More specifically, the crime prevention system 10 according to the present embodiment derives a physical quantity of the same type as the normal physical quantity of the person existing in the monitoring target area 60, and the deviation of the physical quantity from the normal physical quantity. The suspicious degree of the person is specified according to the amount (difference). Then, the crime prevention system 10 detects that a person whose specified suspicious degree exceeds a predetermined threshold is a suspicious person. In the crime prevention system 10 according to the present embodiment, since all persons existing in the monitoring target area 60 are subject to determination as to whether or not they are suspicious persons, all persons existing in the monitoring target area 60 will be described below. Is also referred to as “determination subject person”.

  As shown in Table 1, in the crime prevention system 10 according to the present embodiment, as an action state item used for detection of a suspicious person, a “movement related item” that is an item related to a person's movement state and a person's staying state Two types of “retention related items” are used. The movement-related items include physical quantities such as “person movement speed”, “person movement direction”, “person passage position”, etc., and the stay-related items include “retention time when stopped”. ”,“ Time for staying in the monitoring target area ”,“ position for staying ”, and the like.

  The “person movement speed” is simply a physical quantity indicating the speed at which a person moves. The “person movement direction” is a physical quantity indicating the direction in which the person moves, and the “person passage position” is a physical quantity indicating the position through which the person passes. That is, on a commuting road, a school road, etc., a person normally walks in substantially the same position in substantially the same direction. For this reason, as the physical quantities corresponding to these, the physical quantities of “person movement direction” and “person passage position” can be employed as physical quantities that deviate from the normal behavior state by the suspicious person.

  On the other hand, the “residence time when stopped” is a physical quantity indicating the time during which the user stays at a certain position. Further, the “time staying in the monitoring target area” is a physical quantity indicating the time of staying in the monitoring target area 60, and the “position to stay” is a physical quantity indicating the position where the staying area may stay normally. It is. In addition, as a position which may stay in the said normal time, in the case where a vending machine is provided in the monitoring target area 60, a position near the front of the vending machine or a case where a vending machine is also provided. The inside of the store can be exemplified.

  In other words, the time that a person stays at a certain position, the time that a person stays in an area to be monitored, and the position where a person may stay in a normal period depends on the use and timing of the area to be monitored. Accordingly, it is fixed to some extent. For this reason, as physical quantities corresponding to these, the physical state of the above-mentioned “staying time when stopped”, “time staying in the monitoring target area”, and “position to stay” deviates from the normal state by the suspicious person. It can be adopted as a physical quantity.

  In addition, in the crime prevention system 10 according to the present embodiment, the reason why the two items of “movement related item” and “stay related item” are adopted is as follows. In other words, if the person to be judged is moving, the suspicious person can be detected only by the physical quantity belonging to the “movement-related item”. However, if the person to be judged is not moving, the physical quantity is suspicious. Cannot be detected. In order to compensate for this, the physical quantity belonging to the “retention related item” in addition to the “movement related item” is adopted. Hereinafter, information indicating a physical quantity belonging to “movement related item” is also referred to as “movement related information”, and information indicating a physical quantity belonging to “stay related item” is also referred to as “stay related information”.

  In the crime prevention system 10 according to the present embodiment, a park is adopted as the monitoring target area 60 as described above. For this reason, in the crime prevention system 10 according to the present embodiment, as the “movement-related information”, the “personal movement direction” and the “person passage position” that are considered to be less likely to contribute to detection of suspicious persons in the park. Except for “”, only the physical quantity of “person movement speed” is adopted. On the other hand, as the “retention related information”, “the stay time when stopped”, “the time to stay in the monitoring target area”, and “the position to stay”, which are considered to be highly contributed, The following three types of physical quantities are adopted.

  In the crime prevention system 10 according to the present embodiment, as a normal physical quantity, a monitoring target region is obtained during a normal physical quantity measurement period (hereinafter referred to as “behavioral state measurement period”) obtained by the following equation (1). The average value xa of the physical quantities of all persons detected from 60 is applied. In the formula (1), x represents the value of the physical quantity for each person, and n represents the number of samples (the number of x).

  As described above, in the crime prevention system 10 according to the present embodiment, a simple average value is applied as the normal physical quantity, but the present invention is not limited to this. For example, since the physical quantity obtained when the crime occurs in the monitoring target area 60 is extremely low in reliability, an average value excluding this physical quantity may be applied as the normal physical quantity. In addition, an average value excluding at least one of the maximum value and the minimum value of the physical quantity in the behavior state measurement period may be applied as the normal physical quantity.

  Further, in the crime prevention system 10 according to the present embodiment, the physical quantity x ′ and the average value xa indicating the above-described behavior state of the determination target person himself / herself, obtained by the following equation (2), as the suspicious degree h of the determination target person: The difference is applied.

  As described above, in the crime prevention system 10 according to the present embodiment, the difference itself between the physical quantity x ′ and the average value xa is applied as the suspicious degree h. However, the present invention is not limited to this. For example, as shown in FIG. 8 as an example, the probability distribution D (x) of the physical quantity x is derived, and as shown in FIG. 9 as an example, the evaluation value E (x) obtained by the following equation (3) ), The value corresponding to the physical quantity x ′ of the person to be judged may be applied as the suspicious degree h. In the example shown in FIGS. 8 and 9, the walking speed probability distribution D (x) and the evaluation value when the average walking speed is 1.3 (m / s) and the standard deviation is σ = 0.3. It is an example of E (x).

  And in crime prevention system 10 concerning this embodiment, as a threshold applied when detecting a suspicious person, physical quantity x which shows the above-mentioned action state in the above-mentioned action state measurement period obtained by the following formula (4) is obtained. A value based on the standard deviation σ is applied.

  In the crime prevention system 10 according to the present embodiment, 3σ, which is three times the standard deviation σ, is applied as the threshold value. However, the present invention is not limited to this. For example, a fixed value may be applied according to the security level required for the monitoring target area, the usage of the monitoring target area, or the like, or the user may input via the keyboard 20E or the like. .

  In crime prevention system 10 according to the present embodiment, as described above, “person movement speed”, “staying time when stopped”, “time staying in monitored area”, and “position to stay” The following four types of physical quantities are used. For this reason, in the crime prevention system 10 according to the present embodiment, each of these four types of physical quantities is compared with the corresponding threshold value to determine whether or not it is abnormal. Then, the crime prevention system 10 performs an analysis using the fault tree shown in FIG. 10 as an example in order to comprehensively determine the determination results based on the four types of physical quantities.

  In the example shown in the figure, if the “person's moving speed” is abnormal, that is, if the corresponding physical quantity exceeds the corresponding threshold value, the person to be judged is unconditionally detected as a suspicious person. . Similarly, if the “stay time when stopped” is abnormal, or the “time staying in the monitoring target area” is abnormal and the “position to stay” is abnormal, Detect as a suspicious person.

  And in crime prevention system 10 concerning this embodiment, based on the suspicious person detected by the above processing, it is possible to prevent the occurrence of a crime against a defender who is a person who wants to protect from a crime and a defender. Identify the possible defenders.

  That is, in the case of a crime targeting humans, the crime is established by physically approaching the victim. Therefore, in the crime prevention system 10 according to the present embodiment, the person closest to the position of the suspicious person S in the monitoring target area 60 is defined as the defender V, and the persons other than the suspicious person S and the defender V are defined as the defender. Specify as G.

Here, in the present embodiment, as shown in FIG. 11 as an example, the position of the defender V on the XY coordinates is represented as (X _V , Y _V ), and the XY of the defender G is described above. The position on the coordinates is represented as (X _G , Y _G ), and the position on the XY coordinates of the suspicious person S is represented as (X _S , Y _S ). In addition, in the crime prevention system 10 according to the present embodiment, the distance D _VG between the defender V and the defender G and the distance D _VS between the defender V and the suspicious person S are _expressed by the following formula (5 ) And equation (6).

Then, in the crime prevention system 10 according to this embodiment, assessing the crime risk based on the ratio _D VG _{/ D VS} of the distance _{D VG} and the distance _{D VS.}

  Here, in the crime prevention system 10 according to the present embodiment, the crime risk value R is calculated by the following equation (7), where C is a constant for normalizing the crime risk value.

Here, in the crime prevention system 10 according to the present embodiment, the crime risk value R is set to 0.5 when the suspicious person S and the defender G are equidistant from the defender V, that is, When the distance D _VS is equal to the distance D _VG and the ratio D _VG / D _VS is 1, the calculation result of the expression (7) is 0.5, C = 1.

Table 2 shows a part of the crime risk value R when the constant C = 1 and the distance D _VS and the distance D _VG are changed from 1 to 10, and FIG. 12 shows the entire distribution state. Criminal risk value R decreases as defender G is closer to defender V (distance D _VG is smaller), and crime risk value R increases as suspicious person S is closer to defender V (distance D _VS is smaller). Become.

Here, when the ratio D _VG / D _VS is calculated, if the distance D _VS becomes 0 (zero), it diverges. However, in this case, if the distance between the defensive person V and the suspicious person S is 0 (zero), it means that the defensive person V and the suspicious person S are in contact with each other. It is not handled in the calculation of the value R (0 <D _VS <∞).

  By calculating the crime risk value R using the equation (7), the crime risk in the monitoring target area 60 can be quantitatively evaluated.

  Next, with reference to FIG. 13 and FIG. 14, the effect | action of the crime prevention system 10 which concerns on this Embodiment is demonstrated. Here, a case where the position information database DB1 is constructed in advance in order to avoid complications will be described.

  First, with reference to FIG. 13, the operation of the crime prevention system 10 when measuring the above-described normal physical quantity will be described. FIG. 13 shows the flow of processing of the action distribution measurement program executed by the CPU 20A when an instruction to start measurement of a physical quantity at normal times is accepted via the keyboard 20E of the crime prevention apparatus 20 or the like. It is a flowchart. The program is stored in advance in the program area PG of the secondary storage unit 20D.

  In step 100 in the figure, all information is read from the position information database DB1, and in the next step 102, image information indicating captured images is acquired from all the cameras 40 and stored in a predetermined area of the secondary storage unit 20D. To do.

  In the next step 104, a person existing in the image is detected from the image indicated by the image information acquired by the processing in step 102. In step 104, the position of the person detected by the process of step 104 is detected by the above-described method using the information read from the position information database DB1 by the process of step 100, and the predetermined position in the secondary storage unit 20D is detected. Store in the area. In addition, the detection of the person who exists in the said image from the said image is a technique etc. which are used by FieldAnalyst (registered trademark) (NEC Soft Corp.), IMS (Intelligent Monitoring System, Exit Corp.) etc. as an example. This can be done by a conventionally known technique.

  In the next step 106, it is determined whether or not the execution of the processing in steps 102 to 104 is the first execution after the execution of the action distribution measurement program is started. If the determination is affirmative, the process proceeds to step 118 described later, whereas if the determination is negative, the process proceeds to step 108.

  In step 108, the same person as the person detected by the process of the previous step 104 is started from the person existing in the image indicated by the image information obtained so far after the execution of this action distribution measurement program is started. To detect. In the next step 110, it is determined whether or not the same person has been detected in the processing of step 108. If a negative determination is made, the process proceeds to step 118 described later, whereas if an affirmative determination is made, step 112 is performed. Migrate to The identification of the same person in the process of step 108 is performed by the image information (hereinafter referred to as “past image information”) obtained up to the previous time by the process of step 102 and the image information (hereinafter referred to as “past image information”) obtained this time. It can be performed by an image recognition technique such as a conventionally known pattern matching technique using “current image information”).

  In step 112, the physical quantities of the movement related information and the stay related information related to the same person detected by the processing in step 108 are derived as shown below.

  Regarding the physical quantity of “person movement speed” in the movement-related information, first, among the images indicated by the past image information, the position in the newest (most recent) image where the same person exists and the image obtained this time The distance between the same persons and the position at (hereinafter referred to as “distance between the same persons”). Then, the physical quantity of “human movement speed” is derived by dividing this distance by the elapsed time indicated by the difference in time when the image information corresponding to each of these images is acquired. In the following, the distance between the same persons between the position of the same person in the past image and the position in the image obtained this time is referred to as “distance between the same persons”.

  In addition, regarding the physical quantity of “stay time when stopped” in the stay related information, first, the distance between the same persons is continuously in advance over a predetermined period (5 seconds in the present embodiment). It is determined whether or not there is an image combination that is equal to or less than a predetermined first threshold value. As a result of this determination, when it is determined that this combination exists, the time corresponding to the interval at which image information is acquired by the processing of step 102 (hereinafter referred to as “acquisition”) is added to the number of consecutive images of the combination. The physical quantity of “stay time when stopped” is derived by multiplying by “interval time”). As the first threshold, a value stored in advance as a value that can be considered that the corresponding person has not moved from the spot is applied if the distance between the same persons is equal to or less than the first threshold. be able to. However, the present invention is not limited to this form, and a form in which the user inputs the value corresponding to the use of the crime prevention system 10 or the security level required for the crime prevention system 10 as the first threshold value via the keyboard 20E or the like. Etc.

  Furthermore, the physical quantity of “time to stay in the monitoring target area” in the stay related information is obtained by multiplying the number of images in which the same person exists among the images indicated by the past image information by the acquisition interval time. Derived by

  On the other hand, the physical quantity of “position to stay” in the stay-related information is a physical quantity indicating a position where it can stay at normal times, and should not be obtained as a normal physical quantity in advance. Exclude from In the crime prevention system 10 according to the present embodiment, information indicating the position where the person may stay in the normal state (hereinafter referred to as “stayable position information”) is stored in advance in a predetermined area of the secondary storage unit 20D. Has been.

  It should be noted that the above derivation method of each physical quantity is an example, and it goes without saying that the present invention is not limited to this.

  In the next step 114, it is determined whether or not a crime has occurred at this time. If a negative determination is made, the process proceeds to step 118 described later, whereas if an affirmative determination is made, the process proceeds to step 116. . In this behavior distribution measurement program, the determination processing in step 114 is performed by storing in advance information indicating the normal range of each physical quantity to be measured, and the physical quantity derived within the corresponding range is stored. This is done by determining whether it is in or not. However, the present invention is not limited to this form. For example, a form in which it is determined that no crime has occurred when the difference from the average value of the corresponding physical quantities obtained so far is within a predetermined threshold. Needless to say, other forms may be adopted.

  In step 116, each physical quantity of the movement related information and the stay related information derived by the processing in step 112 is stored in a predetermined area of the secondary storage unit 20D, and then the process proceeds to step 118.

  In step 118, it is determined whether or not a predetermined end condition is satisfied. If the determination is negative, the process proceeds to step 120 and waits for a predetermined time (1 second in the present embodiment). Then, the process returns to step 102. If the determination at step 118 is affirmative, the routine proceeds to step 122.

  In the behavior distribution measurement program according to the present embodiment, the number of persons detected by the process of step 104 after the execution of the program is started is determined as the end condition (this embodiment In the form, the condition of exceeding 1000) is applied. However, the present invention is not limited to this. For example, a condition that a predetermined period (for example, one week) has elapsed since the execution of the behavior distribution measurement program is started may be applied as the end condition. In addition, other conditions such as a condition that an instruction input for instructing the end of the action distribution measurement program has been performed via the keyboard 20E, a condition that the remaining storage capacity of the secondary storage unit 20D is equal to or less than a predetermined amount, and the like. May be applied as the termination condition.

  In the behavior distribution measurement program according to the present embodiment, when repeatedly executing the processing from step 102 to step 120, in step 116, the person's residence related information overlapping with the information stored so far is stored. There is. In this case, in this action distribution measurement program, only the physical quantity exceeding the physical quantity indicated by the information stored so far is overwritten and stored in the residence related information regarding the person.

  In step 122, all the information stored by the process of step 116 is read from the secondary storage unit 20D. Then, in the next step 124, based on the read information, the average value xa and the standard value are calculated for each of the physical quantities of the movement related information and the stay related information by using the above formulas (1) and (4). The deviation σ is calculated. In step 124, the calculated average value xa and standard deviation σ are stored in a predetermined area of the secondary storage unit 20D, and then the action distribution measurement program is terminated.

  Next, the operation of the crime prevention system 10 when executing the crime prevention process for preventing the occurrence of crime in the monitoring target area 60 will be described with reference to FIG. FIG. 14 is a flowchart showing the flow of the crime prevention program executed by the CPU 20A when an instruction to execute crime prevention processing is received via the keyboard 20E or the like. This program is also stored in advance in the program area PG of the secondary storage unit 20D.

  In step 200 in the figure, all the information in the position information database DB1, and the average value xa and the standard deviation σ relating to each physical quantity of the movement related information and the stay related information stored by the processing of the behavior distribution measurement program are secondarily stored. Read from the unit 20D. In the next step 202, image information indicating captured images is acquired from all the cameras 40 and stored in a predetermined area of the secondary storage unit 20D.

  In the next step 204, a person (determination target person) existing in the image is detected from the image indicated by the image information acquired by the process of step 202 by the same process as the process of step 104 of the behavior distribution measurement program. To do. In step 204, the detected position of each person to be determined is detected by the above-described method using the information read from the position information database DB1 by the process of step 200, and is stored in a predetermined area of the secondary storage unit 20D. Remember.

  In the next step 206, it is determined whether or not the execution of the processing in steps 202 to 204 is the first execution after the execution of the crime prevention program is started. If the determination is affirmative, the routine proceeds to step 228 described later, whereas if the determination is negative, the routine proceeds to step 208.

  In step 208, in the same manner as the processing of step 108 of the behavior distribution measurement program, the same person as the determination target detected by the processing of the previous step 204 is executed until the previous time from the start of execution of the crime prevention program. It detects from the image shown by the obtained image information. In the next step 210, it is determined whether or not the same person has been detected in the processing of step 208. If a negative determination is made, the process proceeds to step 228, which will be described later. Migrate to

  In step 212, movement-related information and stay-related information related to the same person detected by the processing in step 208 are derived in the same manner as in step 112 of the behavior distribution measurement program.

  In the next step 214, the suspicious degree h is derived for each physical quantity of the movement related information and the stay related information for all the determination target persons in which the same person is detected using the above-described equation (2).

  In the next step 216, based on the suspicious degree h obtained as described above and the corresponding physical quantity threshold value (in this embodiment, a value that is three times the corresponding standard deviation σ), the aforementioned fault tree (see also FIG. 10). .) Is used to determine whether or not there is a suspicious person S as a determination target person. If the determination is negative, the process proceeds to step 228 described later, and if the determination is affirmative, the process proceeds to step 218. At this time, the physical quantity of “position to stay” in the stay related information is abnormal when the person to be judged exists in an area excluding a predetermined area including the position indicated by the stayable position information described above. It is determined that

In step 218, the person closest to the position of the suspicious person S determined to be present in the process of step 216 among the persons to be detected detected by the process of step 204 is defined as the defender V, A person other than the defender V is identified as a defender G. In the next step 220, the distance D _VG between the defender G and the defender V obtained by the above processing is calculated by the above-described equation (5), and then in the next step 222, the distance D _VG is obtained by the above processing. The distance D _VS between the suspicious person S and the defender V is calculated by the above-described equation (6).

In the next step 224, the crime risk value R is calculated for each suspicious person S using the above-described equation (7) using the distance D _VG and the distance D _VS obtained by the above processing, and then the crime risk value. Store (register) in the database DB3. In the next step 226, the crime risk value R obtained by the process of step 224 is used to execute a process predetermined as preventing crime, and then the process proceeds to step 228.

In the crime prevention program according to the present embodiment, the following three types of processing are applied as the predetermined processing in step 226.
Process 1. A process of displaying the crime risk value R on the display 20F in association with the corresponding person's video.
Process 2. Processing that issues a warning when there is a crime risk value R that is greater than or equal to a predetermined threshold.
Process 3. When the crime risk value R is greater than or equal to a predetermined threshold value, a process of notifying that the crime risk is high to the defender closest to the target defender.

  In addition, when the said process 2 is applied, the speaker can be provided in the monitoring object area | region 60, and the form which issues a warning by the said speaker can be applied. In this case, various forms such as a form of issuing a warning to a monitor or a guard in the monitoring target area 60 via a terminal device such as a PDA (Personal Digital Assistant, portable information terminal) can be applied. .

  In addition, when applying the process 3 described above, a speaker is provided in the monitoring target area 60, and the speaker notifies the defender closest to the target defender that the crime risk is high. The form to be applied can be applied. In this case, various forms such as a form for notifying the defender that the crime risk is high via a terminal device such as a PDA can be applied.

  Here, the predetermined threshold value in the case of applying each of the processes 2 to 3 is automatically set according to the required security level, usage, etc., as well as by the user of the crime prevention system 10 It can also be set as the form input via keyboard 20E etc.

  As described above, in the crime prevention program according to the present embodiment, the three types of processes 1 to 3 are applied as the predetermined process. Any one or a combination of the two may be applied.

  In step 228, it is determined whether or not a predetermined end condition is satisfied. If the determination is negative, the process proceeds to step 230 and waits for a predetermined time (1 second in this embodiment). Then, the process returns to step 202. If the determination in step 228 is affirmative, the crime prevention program is terminated.

  In the crime prevention program according to the present embodiment, the condition that the time set in advance as the time to end the monitoring of the monitoring target area 60 has been applied as the end condition to be applied in the processing of step 228 above. However, the present invention is not limited to this. For example, a condition that an instruction input for instructing the end of the crime prevention program is made via the keyboard 20E or the like, and the remaining storage capacity of the secondary storage unit 20D is a predetermined amount. It goes without saying that other conditions such as the following conditions can also be applied.

The process of step 112 of the behavior distribution measurement program corresponds to the derivation means of the present invention, and the process of step 124 corresponds to the calculation means of the present invention. Further, the process of step 212 of the crime prevention program is also performed by the derivation unit of the present invention, the process of step 214 is performed by the first specifying unit of the present invention, the process of step 216 is performed by the detection unit of the present invention, and the process of step 218 is performed. Each corresponds to the second specifying means of the present invention .

  As described above in detail, in the present embodiment, based on a plurality of photographed images obtained by the photographing means (here, the camera 40) for continuously photographing the target region (here, the monitoring target region 60). Calculating a physical quantity indicating a behavior state of a person existing in the target area, calculating an average value of the derived physical quantity for a predetermined number of persons, and calculating the physical quantity of the person existing in the target area. The suspicious degree (here, suspicious degree h) of the person existing in the target area is specified based on the difference from the average value, and the suspicious person from the person existing in the target area is determined based on the specified suspicious degree Therefore, it is possible to detect a suspicious person with high accuracy without requiring special equipment for the defender and the defender.

  Further, in the present embodiment, since the average value is calculated using a physical quantity excluding a physical quantity derived based on the captured image obtained at the timing when a crime occurs in the target area, higher accuracy can be obtained. Can detect a suspicious person.

  Further, in the present embodiment, in addition to the average value, a standard deviation of the physical quantity used for the calculation of the average value is further calculated, and a person whose suspicious degree exceeds a threshold based on the standard deviation is regarded as a suspicious person. Therefore, the suspicious person can be detected with higher accuracy.

  Further, in the present embodiment, the physical quantity is both the movement related information related to the movement state of the person existing in the target area and the stay related information related to the stay state of the person existing in the target area in the target area. Since it is assumed that there is something, it is possible to easily specify a suspicious degree that is more realistic.

  Furthermore, in the present embodiment, as the movement related information, a movement speed of a person existing in the target area is derived, and as the stay related information, a staying time of the person existing in the target area in the target area, In addition, since the residence time within a specific range in the target area is derived, it is possible to easily specify the degree of suspiciousness that further matches the actual situation.

  As mentioned above, although this invention was demonstrated using embodiment, the technical scope of this invention is not limited to the range as described in the said embodiment. Various modifications or improvements can be added to the above-described embodiment without departing from the gist of the invention, and embodiments to which such modifications or improvements are added are also included in the technical scope of the present invention.

  The above embodiments do not limit the invention according to the claims (claims), and all the combinations of features described in the embodiments are essential for the solution means of the invention. Is not limited. The embodiments described above include inventions at various stages, and various inventions can be extracted by appropriately combining a plurality of disclosed constituent elements. Even if some constituent requirements are deleted from all the constituent requirements shown in the embodiment, as long as an effect is obtained, a configuration from which these some constituent requirements are deleted can be extracted as an invention.

  For example, in the above embodiment, the case where the crime risk value R is derived by the equation (7) has been described. However, the present invention is not limited to this, for example, the crime risk value by the following equation (8): R may be derived.

  Here, P in Formula (8) is the probability that the suspicious person S will commit a crime. Note that the probability P is, for example, Japanese Patent Laid-Open No. 2010-250775 by the inventors of the present invention, “Masayoshi Kondo et al., Crime probability evaluation method based on human attributes for crime risk evaluation—crime risk evaluation method (6), Summary of Academic Lectures of Architectural Institute of Japan, 2010, pp. 1017-1018 ”and the like. In this case, the crime risk value R can be derived with higher accuracy than in the above embodiment.

  Further, in the above embodiment, the case where physical quantities of both movement-related information and stay-related information are used for detection of a suspicious person has been described. However, the present invention is not limited to this, and any of these pieces of information is used. It may be configured to detect a suspicious person using only one of these physical quantities. In this case, compared with the above-described embodiment, the calculation load can be reduced although the detection accuracy of the suspicious person is lowered.

  In the above embodiment, the case where two types of physical quantities of movement-related information and stay-related information are applied to derivation of the suspicious degree has been described. However, the present invention is not limited to this. When hiding, it is good also as a form which raises a suspicious degree compared with other cases.

  In other words, criminal planners may be hiding their faces with helmets, sunglasses, masks, etc. in order to avoid seeing their faces by surrounding people and identifying the human phase by being photographed by a surveillance camera. It is considered high. Further, although the area of the part that hides the face is different between the full-face helmet and the mask, the larger the area of the hidden part, the higher the possibility of being a suspicious person.

On the other hand, depending on the season, the number of people who usually wear masks for hay fever and influenza countermeasures may increase. For this reason, in this embodiment, the ratio f _h of people who are concealed and cannot estimate at least one of age and sex among the people in the monitoring target area is calculated by the following equation (9) as an example. Incidentally, n _h in equation (9) is the number of people who can not at least one of the estimated age and sex, n is the number of people present in the monitored area.

  Here, the estimation of age and gender can be performed by a conventionally known technique such as the technique used in the IMS or the like as an example.

In this embodiment, when at least one of the age and sex of the determination target person cannot be estimated, a larger value is applied as the ratio f _h is smaller as the suspicious degree of the determination target person. In addition, it cannot be overemphasized that the suspicious degree in this form may be used independently, or may be used together with the suspicious degree as described in the said embodiment. In these cases, the same effects as in the above embodiment can be obtained.

  Moreover, although the said embodiment demonstrated the case where it was determined comprehensively whether a determination object person is a suspicious person using a fault tree, this invention is not limited to this. For example, it is determined whether or not each of the physical quantities adopted for deriving the suspicious degree is a suspicious person. It is good also as a form determined to be.

  In the above embodiment, the case where only the physical quantity of “human movement speed” is used as the movement-related information has been described, but the present invention is not limited to this. For example, in addition to the physical quantity of “person movement speed”, one or a combination of physical quantities of “person movement direction” and “person passage position” may be used. In this case, a suspicious person can be detected with higher accuracy than in the above embodiment.

  Moreover, although the said embodiment demonstrated the case where the person nearest to the suspicious person S was specified as the defender V, this invention is not limited to this, For example, the person nearest to the suspicious person S It is good also as a form which specifies the person close | similar to a predetermined order | rank as the defender V. The order in this form may be fixedly set according to the use of the monitoring target area 60, the required security level, or the like, or may be entered by the user via the keyboard 20E or the like. Good. In this case as well, the same effects as in the above embodiment can be obtained.

  Moreover, although the said embodiment demonstrated the case where the suspicious degree h was calculated by the said Formula (2) or Formula (3), this invention is not limited to this. For example, the suspicious degree h ′ may be derived by the following equation (10). In Equation (10), A represents the average value xa of the physical quantities to be employed, B represents the suspicious degree h according to the above-described embodiment corresponding to the physical quantities, α represents the weighted value of the average value A, and β Represents the weight value of the suspicious degree B. Here, the total value of the weighting value α and the weighting value β is 1.

  In this embodiment, the weighting value β is increased as the security level is increased. In this case as well, the same effects as in the above embodiment can be obtained.

  Moreover, although the said embodiment demonstrated the case where a crime prevention process was performed after measuring a normal physical quantity, this invention is not limited to this. For example, a normal physical quantity may be measured while executing crime prevention processing, that is, a normal physical quantity may be updated sequentially. In this case, the suspicious degree can be derived with higher accuracy than in the above embodiment.

  Moreover, although the said embodiment demonstrated the case where the normal physical quantity was measured for the monitoring object area | region 60 made into the execution object of a crime prevention process, this invention is not limited to this. For example, the normal physical quantity may be measured for an area different from the monitoring target area 60. However, it is preferable that the area to be applied in this case is an area that is as close as possible to conditions such as the use, position, and time zone of the monitoring target area 60.

  In the above-described embodiment, the case where the crime risk value is derived without considering any obstacles present in the monitoring target area 60 has been described. However, the present invention is not limited to this, It is also possible to derive a crime risk value in consideration of obstacles.

As an example of this case, when deriving the distance D _VG between the defender V and the defender G and the distance D _VS between the defender V and the suspicious person S, the distance between the defender V and the defender G In the meantime, if there is an obstacle area indicated by the area information registered in the obstacle information database DB2, the distance D _VG is derived to bypass the obstacle area, and the defensive person V and In the case where the obstacle area exists with the suspicious person S, a form in which the distance D _VS is derived as an alternative to the obstacle area can be exemplified. Thereby, crime can be prevented with higher accuracy.

  Further, in the above embodiment, the timing for measuring the physical quantity in normal times is not particularly mentioned. For example, the season, the month, the beginning, the middle, and the end of the month, every predetermined time in the day (for example, 1 It is good also as a form measured beforehand for every period, such as every time. In this form, when executing crime prevention processing, the normal physical quantity acquired at the time corresponding to the execution time is applied. In this case as well, crime can be prevented with higher accuracy.

  Moreover, although the case where the physical quantity obtained at the timing when the crime occurs is excluded from the normal physical quantity derivation target has been described in the above embodiment, the physical quantity may not be excluded. With this configuration, although the accuracy of detecting a suspicious person is slightly reduced, the calculation load for excluding the physical quantity can be reduced.

  In the above embodiment, the case where a suspicious person is detected by comprehensively determining a plurality of physical quantities has been described. However, a suspicious person may be detected based on a single physical quantity.

  Further, in the above embodiment, as a process for preventing a crime, a case where a process for notifying a crime risk value or a process for issuing a warning or a warning when the crime risk value is equal to or greater than a predetermined threshold has been described. However, the present invention is not limited to this. For example, processing to prevent crime, such as processing to raise (mark) the attention level of a suspicious person with the highest crime risk value, or to issue an alarm when the crime risk value suddenly changes It is good also as a form applied as.

  Moreover, although the case where four cameras 40 were provided was demonstrated in the said embodiment, this invention is not limited to this, Only one camera 40 may be provided and several units other than four are provided. Needless to say. In these cases, the same effects as in the above embodiment can be obtained.

In the above embodiment has described the case of applying the Euclidean distance as the distance D _VG and the distance D _VS, the present invention is not limited thereto, for example, apply other distances, such as a Manhattan distance It can also be set as the form to do. In this case as well, the same effects as in the above embodiment can be obtained.

  In the above-described embodiment, the case where a camera capable of shooting a moving image in color is applied as the shooting unit is described. However, the present invention is not limited to this. For example, another camera such as a camera capable of shooting a moving image in monochrome or an infrared camera may be applied. In this case as well, the same effects as in the above embodiment can be obtained.

  Moreover, although the case where a park was applied as the target area of the present invention has been described in the above embodiment, the present invention is not limited to this, for example, other open spaces, school routes, commuting routes, and the like. In addition to open spaces, closed spaces such as the interior of buildings can also be applied. In this case as well, the same effects as in the above embodiment can be obtained.

  In addition, the configuration of the crime prevention system 10 described in the above embodiment (see FIGS. 1 to 4) is merely an example, and unnecessary components may be deleted or newly added without departing from the gist of the present invention. It goes without saying that various components can be added.

  The processing flow of each program shown in the above embodiment (see FIGS. 13 and 14) is also an example, and unnecessary processing steps can be deleted or newly added within the scope not departing from the gist of the present invention. It goes without saying that various processing steps can be added and the order of the processing steps can be changed.

  The configurations of various databases shown in the above embodiment (see FIGS. 5 to 7) are also examples, and some information may be deleted or new information may be deleted without departing from the gist of the present invention. Needless to say, can be added or the storage position can be changed.

  Furthermore, the various arithmetic expressions shown in the above embodiment (see Expressions (1) to (10)) are also examples, and unnecessary parameters may be deleted or newly added within the scope not departing from the gist of the present invention. It goes without saying that additional parameters can be added and the order of operations can be changed.

10 Crime Prevention System 20 Crime Prevention Device 20A CPU
20D Secondary storage unit 20E Keyboard 20F Display 40 Camera DB1 Location information database DB2 Obstacle information database DB3 Crime risk value database

Claims (6)

Deriving means for deriving a physical quantity indicating a behavior state of a person existing in the target area based on a plurality of captured images obtained by a shooting means for continuously shooting the target area;
A computing means for computing an average value of the physical quantity derived by the deriving means within a predetermined period or with a predetermined number of people;
First identification means for identifying a suspicious degree of a person existing in the target area based on a difference from an average value calculated by the calculation means of the physical quantity of the person existing in the target area;
Detecting means for detecting a suspicious person from people existing in the target area based on the suspicious degree specified by the first specifying means;
Among the persons present in the photographed image, those who are excluded from the suspicious person detected by the detection means, and who are closest to the suspicious person from the closest position to a predetermined order are defined as other persons. A second identification means for identifying as a defender;
Suspicious person detection device comprising: 2. The suspicious device according to claim 1, wherein the calculation unit calculates the average value using a physical quantity excluding a physical quantity derived by the deriving unit based on the captured image obtained at a timing when a crime occurs in the target area. Person detection device. In addition to the average value, the calculation means further calculates a standard deviation of the physical quantity used for calculation of the average value,
The suspicious person detection device according to claim 1, wherein the detection means detects a person whose suspicious degree exceeds a threshold value based on the standard deviation as a suspicious person. The physical quantity is at least one of movement-related information related to a movement state of a person existing in the target area and stay-related information related to a stay state of a person existing in the target area in the target area. 4. The suspicious person detection device according to any one of 3 above.   The detection means detects a suspicious person from a person existing in the target region using a fault tree using a plurality of types of suspicious degrees specified by the first specifying means.
  The suspicious person detection device according to any one of claims 1 to 4. Computer
Deriving means for deriving a physical quantity indicating a behavior state of a person existing in the target area based on a plurality of captured images obtained by a shooting means for continuously shooting the target area;
A computing means for computing an average value of the physical quantity derived by the deriving means within a predetermined period of time or a predetermined number of persons;
First identification means for identifying a suspicious degree of a person existing in the target area based on a difference from an average value calculated by the calculation means of the physical quantity of the person existing in the target area;
Detecting means for detecting a suspicious person from people existing in the target area based on the suspicious degree specified by the first specifying means;
Among the persons present in the photographed image, those who are excluded from the suspicious person detected by the detection means, and who are closest to the suspicious person from the closest position to a predetermined order are defined as other persons. A second identification means for identifying as a defender;
Program to function as.