JP5594879B2 - Image monitoring device - Google Patents

Description

  The present invention relates to an image monitoring device that detects the occurrence of an abnormal behavior of a person by image processing, and more particularly to an image monitoring device that detects abnormal behavior around a storage such as a safe.

  Until now, the mechanical security system has mainly been responsible for security after closing and absence, but nowadays there is an increasing demand for security in manned environments where employees and householders are present. One of the guards in manned environment is a countermeasure against intrusion robbery aimed at safes. In order to realize security in manned environments, it is necessary to detect abnormalities by distinguishing bandits from people other than bandits.

  In Patent Document 1, in order to detect an anomaly that an intruder targeting a safe or the like exists and distinguish it from a person who simply passes near the safe or the like, the image of the moving object is erased in the monitoring image. Image processing that extracts only the image of the stopped object that has been stopped for a time is performed, and an area (monitoring area) for detecting an abnormality is set in advance in the vicinity of the safe in the monitoring image, and the image of the stopped object is extracted in the area. It is described that it is judged as abnormal when

JP-A-4-253297

  For example, when an employee or a family member uses a safe or a safe, it stops near the safe or the like. In other words, in the situation where there are employees or householders, if an abnormality is detected based on the stoppage of a person in the vicinity of a safe or the like, the normal state in which the employee or householder uses the safe or the like, and when a burglary occurs A bandit or an employee forced by the bandit opens a safe or the like, or the bandit steals money or the like from the bank or the like, thus making a false detection.

  In addition, in the conventional technology for detecting an abnormality based on the stoppage of a person, when an area for detecting an abnormality is set in advance, if the area is excessively wide, a flow line around a safe such as an employee is included and a false alarm is generated. It tends to occur, and if it is too narrow, the possibility of failing to detect a bandit increases. Thus, in the prior art, the quality of the monitoring area setting greatly affects the detection accuracy. However, the appropriate area setting differs depending on the property depending on the camera installation status (camera installation height and angle), the size of the safe, the flow line of employees, and the like. For this reason, a monitoring area suitable for each property is set each time, and skill is required to realize area setting that gives a detection accuracy of a certain level or more in all properties, and the load of area setting work is high.

  The present invention has been made in view of the above problems, and is frequently caused by intrusion robbery or the like in the situation where there are employees or householders, and abnormalities in which a bandit forces the employee or the like to open the storage An object of the present invention is to provide an image monitoring apparatus capable of detecting an action with high accuracy.

The image monitoring apparatus according to the present invention monitors a monitoring area for a locked storage, and acquires an image acquisition unit that acquires a monitoring image obtained by capturing the monitoring area at a predetermined cycle, and the monitoring image includes the monitoring A person tracking unit that tracks each person in the area,
A lock state detection unit that detects whether the storage is locked or unlocked, and is tracked at a position closest to the storage when the storage changes to the unlocked state. The unlocker identification unit that identifies the person as an unlocker, and the position with respect to the storage during the unlocking period from when the unlocker is specified until the storage changes to the locked state. And an abnormality determination unit that determines an abnormality when the person other than the unlocker is replaced with the unlocker.

  In a preferred aspect of the present invention, the abnormality determination unit detects the change when the person being tracked at a position closest to the storage during the unlocking period is not the same person as the unlocker. An image monitoring device.

  In another preferred aspect of the present invention, the abnormality determining unit is configured to perform the replacement when the person being tracked is located closer to the storage than the unlocker during the unlocking period. It is an image monitoring device that detects

  In another image monitoring apparatus according to the present invention, when the abnormality is determined, the abnormality determination unit sets an additional determination condition that the unlocker is in a stationary state during the unlocking period.

  In the image monitoring apparatus according to another aspect of the present invention, the abnormality determination unit includes an additional determination condition that a substitute for the unlocker has a predetermined appearance characteristic when determining the abnormality. To do.

  The image monitoring apparatus according to another aspect of the present invention further includes a situation image generation unit that generates a situation image in which the distinction between the unlocker and the substitute with the unlocker is displayed on the monitoring image.

  According to the present invention, it is difficult to erroneously detect the use of the storage by employees or householders during normal times as abnormal, and it is possible to detect abnormal actions taken by a thief who has been pushed in and a threatened employee. In addition, the setting of the detection area for abnormal behavior is less likely to affect the detection accuracy, and the area setting load is reduced.

1 is a schematic diagram illustrating an overall configuration of an image monitoring apparatus according to an embodiment of the present invention. 1 is a schematic block diagram of an image sensor constituting an image monitoring apparatus according to an embodiment of the present invention. It is a schematic flowchart of the monitoring operation | movement of the image monitoring apparatus which concerns on embodiment of this invention. It is a general | schematic flowchart of an abnormality determination process. It is a schematic diagram which shows the specific example of replacement. It is a schematic diagram which shows the example of a condition image.

  Hereinafter, an image monitoring apparatus 1 according to an embodiment of the present invention (hereinafter referred to as an embodiment) will be described with reference to the drawings.

  The image monitoring apparatus 1 uses a safe room in which a safe (storage) is installed as a monitoring space, processes a monitoring image obtained by imaging the monitoring space, and is taken by an employee who is threatened by a bandit or a bandit when an intrusion robbery occurs. Detect abnormal behavior. In particular, the image monitoring apparatus 1 detects, as an abnormal behavior, a replacement action (substitution) in which a bandit replaces an employee who has unlocked a safe and takes out a gold item by himself.

[Configuration of image monitoring device]
FIG. 1 is a schematic diagram showing an overall configuration of an image monitoring apparatus 1 that uses a safe room in which a safe 10 is placed as a monitoring space. The image monitoring device 1 includes an image sensor 2, a controller 3, a setting input unit 4, a safe lock sensor 5, a threatening device 6, and a security center device 7. The image sensor 2, setting input unit 4, safe lock sensor 5, threatening device 6 and security center device 7 are connected to the controller 3.

  The controller 3 is installed in an office room in a building equipped with a safe room. Each unit connected to the controller 3 transmits and receives various signals via the controller 3. Each unit connected to the controller 3 is controlled by the controller 3 as necessary.

  The image sensor 2 is installed, for example, on a ceiling portion directly above the safe 10 so as to overlook the safe room, and performs imaging of a monitoring image and detection of abnormal behavior using the monitoring image. The image sensor 2 outputs a detection signal to the controller 3 when detecting abnormal behavior.

  The imaging range of the monitoring image includes at least the area immediately before the door of the safe 10 and its peripheral area. The peripheral area is wider than the position range in which the unlocker of the safe 10 stays during the unlocking operation, and a radius of about 5 m with the safe 10 as the center is preferable so that the person around the unlocker can be easily tracked. is there. The image sensor 2 is installed at a position where the peripheral area can be imaged.

  The setting input unit 4 is an interface device through which the operator of the image monitoring apparatus 1 inputs settings such as a storage position 210 that represents the position of the safe 10 in the monitoring image, an auxiliary evaluation item 215, and the like. The setting input unit 4 is temporarily connected to the controller 3 before the operation of the image monitoring apparatus 1. At the time of connection, the monitoring image from the image sensor 2 is relayed by the controller 3 and displayed on the setting input unit 4. When the storage location 210, the incidental evaluation item 215, and the like in the displayed monitoring image are set and input, the settings are input to the image sensor 2 via the controller 3, and the image sensor 2 stores the settings.

  The safe lock sensor 5 is installed in the safe 10 and outputs a lock state signal indicating whether the lock of the safe 10 is locked or unlocked to the controller 3. The lock state signal is relayed by the controller 3 and input to the image sensor 2.

  The threatening device 6 is an acoustic device that is installed in a safe room and reproduces warning voices and warnings for bandits. When the detection signal is input from the image sensor 2, the controller 3 outputs a start signal to the threatening device 6, and the threatening device 6 operates to threaten the bandits when the start signal is input.

  The security center device 7 is installed in a security center at a remote location, for example, and is connected to the controller 3 via a wide area communication network such as a public telephone line. When the detection signal is input from the image sensor 2, the controller 3 transmits an abnormality notification signal to the security center device 7, and when the security center device 7 receives the abnormality notification signal, an abnormality occurs to a security guard resident in the security center. Is notified.

  FIG. 2 is a schematic block diagram of the image sensor 2. The image sensor 2 includes an imaging unit 20, a storage unit 21, a control unit 22, and a communication unit 23. The imaging unit 20, the storage unit 21, and the communication unit 23 are connected to the control unit 22.

  The imaging unit 20 is a monitoring camera, images the monitoring space every predetermined time, and sequentially outputs monitoring images to the control unit 22. The time interval at which the monitoring image is captured is 1/5 second, for example. Hereinafter, the unit of time recorded at the time interval of imaging is referred to as time.

  The storage unit 21 is a storage device such as a ROM (Read Only Memory) or a RAM (Random Access Memory). The storage unit 21 stores various programs and various data, and inputs / outputs such information to / from the control unit 22. The various data includes a storage location 210, a lock state 211, tracking information 212, a person position 213, unlocked person information 214, and incidental evaluation items 215.

  The storage location 210 is information on the position of the safe 10 in the monitoring image, and is set in advance via the setting input unit 4. Specifically, the storage location 210 is one XY coordinate expressed in the coordinate system of the monitoring image.

  The lock state 211 is a lock state obtained from the lock state signal. In order to determine the change in state, the state at the current time and the state one hour before are stored.

  The tracking information 212 is image information generated and referred to for tracking a person who moves in the monitoring space. Specifically, the tracking information 212 is a background image generated based on a monitoring image captured when there is no person and a feature amount extracted from each person's image. As the feature amount, a luminance histogram of a human image, texture information, or the like can be used. The tracking is a process for obtaining the position of each person at each time by identifying the persons having similar feature quantities among the persons detected at the preceding and following times as the same person. A common identifier (person ID) is assigned to the identified person, and the feature amount is stored in association with the person ID for each identified person.

  The person position 213 is a history of the position of the person in the monitoring image, and is generated as a tracking result. The person position 213 is stored for each identified person in association with the person ID.

  The unlocker information 214 is information for identifying a person (unlocker) who unlocked the storage. The unlocking person is specified by the unlocking person specifying unit 221 described later, and the person ID of the specified unlocking person is stored as the unlocking person information 214. When the storage is locked, the unlocker information 214 is cleared to “no unlocker”.

  The incidental evaluation item 215 is information of an item to be additionally evaluated as an additional determination condition in addition to the replacement in order to increase the determination accuracy of abnormality. Specifically, these items are a movement evaluation setting for evaluating the movement of the unlocker, and an appearance evaluation setting for evaluating the appearance characteristics of the worker (alternator) who has been replaced with the unlocking person. It is set in advance according to the operation of the property.

  The movement evaluation setting is set to either valid / invalid. If it is valid, a change detection signal is output when the unlocker is in a stationary state after a change is detected. Whether or not to output an abnormality detection signal is determined regardless of the movement of the unlocker. For example, in a property that requires the presence of a supervisor to operate a safe, the supervisor supervises the safe depositing / unloading of safe deposits after unlocking the safe. In this case, since the supervisor may not move during the work in / out, it may be preferable to set the motion evaluation setting to invalid. On the other hand, in a property where unlocking and carrying-in / out work are carried out independently, the dynamic evaluation setting can be made effective.

  Appearance evaluation settings are further subdivided into three sub-items: (1) Face concealment, (2) Uniform presence / absence, and (3) Face authentication. If the employee's uniform pattern data and face authentication are valid, the face data of each employee is set as registered face data.

  If “face concealment” is enabled, it is determined whether the face of the substitute is concealed by sunglasses or a mask when a change is detected, and a detection signal is output when the concealment is determined and is invalid. If so, it is determined whether or not to output a detection signal regardless of the concealment determination. For example, “face concealment” can be set to be invalid when the number of mask wearers increases due to a cold or hay fever, and can be set to be valid at other times. In addition, it is preferably invalidated at work places where employees wear masks or the like for protection purposes or in special storage.

  If “Uniform presence / absence” is valid, it is determined whether or not a replacement image is detected when the replacement image contains a pattern that matches the uniform pattern data. If it is output and invalid, it is determined whether or not to output a detection signal regardless of the uniform determination. For example, in a property that requires uniforms to be worn or a property that can extract the type of employee's clothes, “While uniform” can be set effective, but in a property that does not have a uniform, you can select invalid. it can.

  If “face authentication” is enabled, it is determined whether or not a replacement image is detected and the replacement image contains a pattern that matches the registered face data. If the signal is output and is invalid, it is determined whether to output the detection signal regardless of face authentication. For example, “face authentication” can be selected when the safe operator is limited to a specific person. In order to realize high-accuracy face authentication processing, the control unit 22 needs higher processing capacity than the processing of other incidental evaluation items, and the capacity of the storage unit 21 is required for registration of face data. The cost of the image sensor 2 may increase. In addition, it is necessary to update the registered face data in a timely manner, which may increase the operation cost. In these respects, the appearance evaluation function based on “face authentication” may be installed as a user option.

  The control unit 22 is an arithmetic device such as a CPU (Central Processing Unit), a DSP (Digital Signal Processor), or an MCU (Micro Control Unit). The control unit 22 functions as these units by reading out from the storage unit 21 and executing programs describing the operations of the person tracking unit 220, the unlocking person specifying unit 221, the abnormality determining unit 222, the situation image generating unit 223, and the like. .

  The person tracking unit 220 tracks a person in the monitoring image and stores the person position 213 of each person in the storage unit 21. Therefore, the person tracking unit 220 repeats a person image extraction process for extracting a person image (person image) from the monitoring image at each time and an identification process for identifying a person image at the preceding and following times, for each identified person image. The position of the center of gravity is stored as the person position 213.

  The person extraction process will be specifically described. The person tracking unit 220 stores a monitoring image captured when there is no person in the monitoring space as a background image of the tracking information 212. Then, background difference processing is performed between the new monitoring image and the background image to detect the background difference region, label the background difference region, and extract a portion corresponding to each label from the monitoring image as a person image. . After the extraction, the person tracking unit 220 updates the background image by adding and synthesizing the portion of the monitoring image where the person image has not been extracted to the background image. Note that, instead of the background difference processing, correlation processing is performed between the monitoring image and the background image, and a person image can be extracted by labeling an image of a region having a low correlation.

  The identification process will be specifically described. The person tracking unit 220 analyzes each person image extracted from the new monitoring image and extracts a feature amount for each person image, and extracts the feature amount for each person ID stored as the extracted feature amount and the tracking information 212. The similarity with the feature amount is calculated and compared with an identification threshold set in advance, and the person image and person ID whose similarity exceeds the identification threshold are identified. Then, the person tracking unit 220 calculates the gravity center position of the person image and adds the gravity center position to the person position 213 of the person ID identified as the person image.

  If there is a person image for which identification has not been obtained at this time, the person tracking unit 220 assigns a new person ID to the feature amount of the person image and appends it to the tracking information 212, assuming that a new person has appeared. With the assignment of a new person ID, the person tracking unit 220 newly creates a person position 213 corresponding to the person ID.

  If there is a person ID for which identification has not been obtained, the person tracking unit 220 deletes the feature quantity corresponding to the person ID from the storage unit 21 assuming that the person image represented by the person ID has disappeared.

  The unlocker identifying unit 221 identifies the person who is being tracked at the position closest to the storage when the storage changes from the locked state to the unlocked state, and the person ID of the unlocker Is stored in the storage unit 21 as unlocker information 214. In addition, the unlocker identification unit 221 clears the unlocker information 214 when the storage changes from the unlocked state to the locked state.

  Specifically, when the unlocking person specifying unit 221 detects a change to the unlocked state, the current position of each person stored as the person position 213 and the position of the storage stored as the storage position 210 The person with the smallest distance calculated by calculating the distance between and is determined as the unlocker.

  The change in the lock state is detected by referring to the lock state 211 at the current time and the lock state 211 one hour before.

  The abnormality determining unit 222 refers to the person position 213 in the unlocking period from when the unlocking person is determined by the unlocking person specifying unit 221 until the storage is changed to the locked state, and the position of the unlocking person is switched. The presence / absence of a person is detected, and if a replaced person (substitute) is detected, an abnormality is determined. When the abnormality determination unit 222 determines an abnormality, the abnormality determination unit 222 generates abnormality information related to the abnormality and outputs the abnormality information to the communication unit 23.

  Specifically, the abnormality determination unit 222 calculates a distance between the current position of each person and the position of the storage during the unlocking period, identifies the person with the smallest calculated distance as an operator, If the worker does not match the unlocker, the worker is detected as a substitute.

  In another embodiment, the abnormality determination unit 222 calculates a distance between the current position of each person and the position of the storage during the unlocking period, so that the person other than the unlocker is more than the unlocker. A person whose small distance is calculated is detected as a substitute. In this case, all the substitutes can be detected when there are a plurality of substitutes.

  Further, the abnormality determination unit 222 determines whether or not the unlocking person is in a stationary state from the person position 213 during the unlocking period if the moving and static evaluation setting is valid in the incidental evaluation item 215. Is determined to be abnormal. Specifically, the abnormality determination unit 222 extracts a time length without movement from the person position 213 of the unlocker, and determines that the state is still if the extracted time length exceeds a preset stillness determination time. . If the position difference between successive times at the person position 213 is less than or equal to the extraction error of the person extraction process, no movement is assumed.

  Further, the abnormality determination unit 222 determines whether the face is concealed from the substitute person's image if the small item of “face concealment” in the appearance evaluation setting is valid in the incidental evaluation item 215. Is determined to be abnormal. Specifically, the abnormality determination unit 222 detects an oval shape from the substitute person's image, extracts skin color pixels within the ellipse, and determines face concealment if the skin color area relative to the area of the ellipse is less than a predetermined value. To do.

  In addition, the abnormality determination unit 222 searches for the uniform pattern in the substitute person's image if the minor item “appearance of uniform” in the appearance evaluation setting in the incidental evaluation item 215 is valid, and if the uniform pattern is not detected, the appearance is abnormal. It is determined that Specifically, the abnormality determination unit 222 scans uniform pattern data set in advance on the person image of the substitute to detect a portion that matches the uniform pattern data.

  Also, the abnormality determination unit 222 determines whether or not the face portion of the substitute person's image matches each registered face data if the sub-item “face authentication” of the appearance evaluation setting is valid in the incidental evaluation item 215. If it does not match any registered face data, it is determined that the appearance is abnormal. Specifically, the abnormality determination unit 222 detects an ellipse shape from the person image of the substitute, compares the image in the ellipse with the registered face data, and determines whether or not it matches the registered face data.

  When two or more substitutes are detected, the abnormality determination unit 222 performs an appearance evaluation process according to the appearance evaluation setting for each substitute.

  The abnormality determination unit 222 that performs the replacement determination and various determinations according to the incidental evaluation item 215 determines whether to output an abnormality detection signal to the controller 3 according to the determination result and the incidental evaluation item 215.

  The situation image generation unit 223 generates a situation image by adding information that can identify the unlocking person and the substitute who has replaced the unlocking person to the monitoring image that has been determined to be abnormal by the abnormality determination part 222. The generated situation image is added to the abnormality information by the abnormality determination unit 222.

  The communication unit 23 is an interface circuit for transmitting and receiving various signals between the controller 3 and the control unit 22 and its driver software. For example, a lock state signal from the safe lock sensor 5 is received by the communication unit 23 and read by the control unit 22. A detection signal from the control unit 22 is transmitted to the controller 3 via the communication unit 23.

[Operation of image monitoring device]
As an advance preparation for image monitoring, the operator of the apparatus connects the setting input unit 4 to the controller 3 to set the storage location 210 and the accompanying evaluation items 215.

  When the monitoring image from the imaging unit 20 is displayed on the setting input unit 4 and the operator designates one point in the area where the storage is imaged on the monitoring image, the coordinate value of the designated point is the image. It is transmitted to the sensor 2 and stored as the storage location 210 in the storage unit 21. The setting of the storage location 210 is different from the setting of the area for detecting an abnormality performed in the prior art, because conversion from the actual size to the area size on the image is not necessary, and a simple point can be specified. Setting workload is extremely small.

  After the initial setting, the monitoring operation starts when an operator who confirms that the monitoring space is unattended and the storage is locked is turned on.

  Hereinafter, the monitoring operation of the image monitoring apparatus 1 will be described. FIG. 3 is a schematic flowchart of the monitoring operation of the image monitoring apparatus 1.

  First, the control unit 22 performs initialization (S1). That is, the person tracking unit 220 of the control unit 22 initializes the background image of the tracking information 212 of the storage unit 21 with the monitoring image captured immediately after the operation starts, and the unlocking person specifying unit 221 of the control unit 22 is stored in the storage unit 21. And the control unit 22 initializes the lock state 211 at the current time to “locked”.

  After the initialization, the processing of steps S2 to S12 is repeated each time a new monitoring image is input from the imaging unit 20 to the control unit 22.

  When a new monitoring image is input (S2), first, the person tracking unit 220 compares the monitoring image with the tracking information 212 and advances the process of tracking a person in the monitoring image for one hour (S3). . As a result, if a person exists in the monitoring space, the person tracking unit 220 accumulates the position of each person in the monitoring image in the person position 213 of the storage unit 21, and the person tracking unit 220 uses the tracking unit 21 for tracking. A feature amount of each person's image is generated and updated in the information 212.

  Next, the control unit 22 acquires the lock state 211 (S4). That is, the control unit 22 copies the lock state 211 at the current time to the lock state 211 one hour ago, controls the communication unit 23 to acquire the lock state signal from the safe lock sensor 5, and displays the state represented by the signal. The lock state 211 at the current time is overwritten (S4).

  When the lock state 211 one hour before is the locked state and the lock state 211 at the current time is the unlocked state (YES in S5), the control unit 22 identifies the unlocker by the unlocker identifying unit 221. (S6). Specifically, the unlocking person specifying unit 221 calculates the distance between the storage location 210 and the person position 213 for each person being tracked, and calculates the person ID of the person with the smallest calculated distance. The information 214 is stored in the storage unit 21. When the lock of the safe 10 changes from the locked state to the unlocked state, there is an unlocking person in the monitoring space, and the unlocking person can be estimated to be the person who is closest to the safe 10.

  When lock state 211 does not satisfy the condition of step S5 (NO in S5), step S6 is skipped.

  Subsequently, if both the lock state 211 one hour ago and the lock state 211 at the current time are both unlocked (YES in S7), the control unit 22 determines whether there is an abnormality by the abnormality determination unit 222. (S8). Incidentally, when the process proceeds to YES in step S7, the safe 10 is unlocked and the unlocking person is specified.

  FIG. 4 is a schematic flowchart of the abnormality determination process S8. With reference to the flowchart of FIG. 4, the abnormality determination process in step S8 will be described.

  In a state where the safe 10 is unlocked, the person located closest to the safe 10 can be estimated to be an operator who is taking out from the safe 10 or performing storage work in the safe 10. First, the abnormality determination unit 222 calculates the distance between the storage position 210 and the person position 213 for each person being tracked (S80), and specifies the person ID of the person with the smallest calculated distance as the worker. Then, it is determined whether or not the worker's person ID is the same as the unlocker's person ID (S81). If they are not the same, it is determined that there is a replacement, and if they are the same, it is determined that there is no replacement. If there is a change, the bandit may threaten the employee to unlock the safe 10 and then the bandit may have taken out the money from the bank 10. On the other hand, when there is no replacement, that is, when the worker and the unlocker are the same, it is estimated that the employee who has unlocked the safe 10 continues to work in front of the safe 10 in a normal state.

  FIG. 5 is a schematic diagram showing a specific example of replacement. The replacement determination process will be described with reference to FIG. The time advances to t1, t2, and t3, and the monitoring images 50-1, 50-2, and 50-3 are captured at these times. In the monitoring image 50, a storage 51, an employee 52, and a bandit 53 are captured. A person is tracked using these monitoring images 50. For example, the employee 52 is assigned a person ID “A” and is tracked as a person A, and the bandit 53 is assigned a person ID “B” and is tracked as a person B. Images 60-1, 60-2, and 60-3 are obtained by plotting the person positions 213 and storage positions 210 of the persons A and B on the images corresponding to the monitoring images 50-1, 50-2, and 50-3. .

  Also, the three tables at the bottom of FIG. 5 show the data of the lock state 211, the data of the unlocking person information 214 at each time, and the operator determination result 70 by the abnormality determination unit 222.

  At time t1, the bandits 53 approached the storage 51 while threatening the employees 52. At this time, the lock state 211-1 is a locked state, and the unlocking person information 214-1 is “none”. Since the worker is not specified while the locked state 211 is in the locked state, the worker determination result 70-1 is set to “-”.

  The employee 52 unlocked the storage 51 at time t2. The unlocked state 211-2 is rewritten as “unlocked”, the person A is identified as the unlocked person by the unlocking person specifying unit 221, and the person ID “A” is stored in the unlocking person information 214-2. At this timing, the operator is specified by the abnormality determination unit 222, but no replacement has yet occurred, and the worker determination result 70-2 is the person A.

  At time t3, bandits 53 started work in front of storage 51. The worker is the person B, and the worker determination result 70-3 is not the same as the person A in the unlocker information 214-3, and the replacement is determined.

  A process when there is a replacement as shown in FIG. 5 will be described. As shown in FIG. 4, when there is a replacement (YES in S82), the situation image generation unit 223 of the control unit 22 monitors the replacement image, the person position 213 of the unlocker at that time, and the current position A situation image is generated based on the person position 213 of the worker (substitute) (S83), and the abnormality determination unit 222 generates abnormality information including the replacement and the situation image and stores it in the storage unit 21 (S84). ).

  FIG. 6 is a schematic diagram illustrating an example of a situation image. The base image in FIG. 6 is the monitoring image 50-3 in FIG. 5, and the base person positions 213 are the person positions 213-A3 and 213-B3 in FIG. The situation image generation unit 223 duplicates the monitoring image 50-3 as the situation image 80, draws the character 81 of “unlocker” at the end of the situation image 80, and from the predetermined position near the character 81, the person position of the unlocker A straight line 82 directed to 213-A3 is drawn to the front of the person image 52-3, and a character 83 of “alternate” is drawn at a position that does not overlap the character 81 at the end of the situation image 80. A straight line 84 directed to the substitute person position 213-B3 is drawn up to the front of the person image 53-3.

  Further, when there is a replacement, as shown in FIG. 4, the abnormality determination unit 222 confirms the incidental evaluation item 215 and performs incidental evaluation according to the setting (S85 to S92).

  First, if any one or more of the suspicious appearance small items of “Face concealment”, “Uniform presence / absence”, and “Face authentication” in the appearance evaluation setting of the incidental evaluation item 215 are valid (YES in S85) The determination unit 222 evaluates the sub-items that are set to be effective for the substitute person image to determine whether the substitute person image has a suspicious appearance (S86). If it is determined (YES in S87), the fact that the suspicious appearance is determined is added to the abnormality information (S88). If the appearance evaluation setting is invalid (NO in S85), steps S86 to S88 are omitted. If the suspicious appearance is not determined (NO in S87), step S88 is omitted.

  Subsequently, if the movement evaluation setting of the incidental evaluation item 215 is valid (YES in S89), the abnormality determination unit 222 does not move the unlocker from the person position 213 of the unlocker during the unlocking period for more than the stationary determination time. It is determined whether or not it is stationary (S90). If it is stationary (YES in S91), the fact that it is stationary is added to the abnormality information (S92), and the process proceeds to step S9 in FIG. It is advanced.

  In unusual cases, such as intrusion robbery, it is unlikely that the thief will release the unlocked employee as it is, and the employee will remain constrained within the reach of the thief's eyes and remain stationary. Conversely, if a change occurs in a normal case where multiple employees work together in front of the vault, it is assumed that the unlocker will not be stationary because the unlocker will also perform the work unlike in the case of push-in robbery. The

  If the dynamic evaluation setting is invalid (NO in S89), steps S90 to S92 are omitted. If the stationary state is not determined (NO in S91), step S92 is omitted.

  The case where the determination result indicating that there is a replacement is obtained in the replacement determination S81 has been described above (YES in S82). On the other hand, when the determination result that there is no replacement is obtained (NO in S82), the worker and the unlocker are the same, and in this case, the employee who unlocked safe 10 continues. Thus, it is presumed that the normal state of continuing the work in front of the safe 10. That is, the abnormality determination unit 222 proceeds to step S9 as normal if the worker and the unlocker are the same.

  Returning to FIG. 3 again, the continuation of the process will be described. When abnormality information satisfying the setting of the incidental evaluation item 215 is generated in step S8 (YES in S9), the abnormality determination unit 222 outputs an abnormality detection signal including the abnormality information to the communication unit 23 ( S10). The communication unit 23 transmits a detection signal to the controller 3, and the controller 3 receives the detection signal and transmits an abnormality notification signal to the security center device 7. The security center device 7 that has received the abnormality notification signal performs display processing for displaying an abnormality image and an abnormal image. The monitor who has viewed the display can quickly grasp the relationship between the characters in the abnormal image and take appropriate coping actions.

  When unlocked as described above, the control unit 22 identifies the unlocker (S6) and performs replacement monitoring. On the other hand, if the control unit 22 is locked, that is, if the lock state 211 one hour ago is in the unlocked state and the lock state 211 at the current time is in the locked state (YES in S11), the controller 22 identifies the unlocker. The unlocker information 214 is cleared by the part 221 (S12).

  When lock state 211 does not satisfy the condition of step S11 (NO in S11), step S12 is skipped. That is, the unlocker information 214 specified at the time of unlocking the safe 10 in step S6 is held until the safe 10 is locked.

  When the above processing ends, the processing returns to step S2 again, and processing at the next time is performed.

  In the above embodiment, the image monitoring apparatus 1 applied when the storage is a safe has been described as an example. However, the storage is not limited to a safe, and the image monitoring apparatus of the present invention is used in the monitoring space in which various storages such as a chemical storage, a jewelry showcase, and a dangerous goods storage are installed. Similarly, abnormal behavior can be detected.

  As already described, the abnormality determination unit 222 determines whether or not to output an abnormality detection signal via the controller 3 in accordance with the replacement determination result and the accompanying evaluation item 215. For example, when the determination result and the incidental evaluation item 215 are generated, the abnormality determination unit 222 can generate various abnormality detection signals depending on, for example, a required monitoring level or a difference in conditions for each property. It is. For example, an abnormality detection signal can be output only by replacement detection regardless of the result of incidental evaluation. This configuration increases the sensitivity of detecting abnormal behavior and can reduce false alarms.

  As described above, the incidental evaluation items can be selected according to the user's needs and usage environment. For example, four of the small items of the motion evaluation and the appearance evaluation can be set as additional requirements in parallel for the replacement determination. That is, in addition to the detection of replacement, if any one of the four associated evaluation items that is set to be valid is an abnormality determination, the abnormality determination unit 222 generates a detection signal. This configuration can be expected to have fewer false alarms than the configuration that generates the detection signal based only on the replacement determination described above.

  In addition, it is also possible to adopt a configuration in which the motion / static evaluation is set as a first additional requirement, and the three small items of appearance evaluation are set as second additional requirements parallel to each other for the replacement determination. In this configuration, if the change is detected, the motion evaluation is an abnormality determination, and one of the three sub-items of the appearance evaluation that is set valid is an abnormality determination, the abnormality determination unit 222 detects Generate a signal. This configuration can further suppress false alarms than the configuration in which the above-described four incidental evaluation items are handled in parallel.

  1 image monitoring device, 2 image sensor, 3 controller, 4 setting input unit, 5 safe lock sensor, 6 threatening device, 7 security center device, 10 safe, 20 imaging unit, 21 storage unit, 22 control unit, 23 communication unit, 50 monitoring image, 80 status image, 210 storage location, 211 lock state, 212 tracking information, 213 person position, 214 unlocking person information, 215 additional evaluation items, 220 person tracking unit, 221 unlocking person specifying unit, 222 An abnormality determination unit, 223 a situation image generation unit.

Claims (5)

An image monitoring device for monitoring a monitoring space where a locked storage is installed,
An image acquisition unit for acquiring a monitoring image obtained by photographing the monitoring space at a predetermined period;
A person tracking unit for tracking each person present in the monitoring space in the monitoring image;
A lock state detection unit that detects whether the storage is locked or unlocked; and
An unlocker identification unit that identifies the person being tracked as the unlocker when the storage is changed to the unlocked state;
During the unlocking period from when the unlocker is specified until the storage changes to the locked state, the person other than the unlocker is replaced with the unlocker with respect to the position with respect to the storage. An abnormality determination unit that determines abnormality when it occurs;
Have
The abnormality determination unit detects the change when the person being tracked at a position closest to the storage is not the same person as the unlocker during the unlocking period;
An image monitoring apparatus characterized by the above. An image monitoring device for monitoring a monitoring space where a locked storage is installed,
An image acquisition unit for acquiring a monitoring image obtained by photographing the monitoring space at a predetermined period;
A person tracking unit for tracking each person present in the monitoring space in the monitoring image;
A lock state detection unit that detects whether the storage is locked or unlocked; and
An unlocker identification unit that identifies the person being tracked as the unlocker when the storage is changed to the unlocked state;
During the unlocking period from when the unlocker is specified until the storage changes to the locked state, the person other than the unlocker is replaced with the unlocker with respect to the position with respect to the storage. An abnormality determination unit that determines abnormality when it occurs;
Have
The abnormality determination unit detects the alternation when the person being tracked is present at a position closer to the storage than the unlocker during the unlocking period. Monitoring device. In the image monitoring apparatus according to claim 1 or 2 ,
The image monitoring apparatus according to claim 1, wherein the abnormality determination unit sets an additional determination condition that the unlocker is in a stationary state during the unlocking period when determining the abnormality. The image monitoring apparatus according to any one of claims 1 to 3 ,
The image monitoring apparatus according to claim 1, wherein, when determining the abnormality, the abnormality determination unit sets an additional determination condition that a substitute for the unlocker has a predetermined appearance characteristic. In the image monitoring apparatus according to any one of claims 1 to 4, further
An image monitoring apparatus, comprising: a situation image generation unit configured to generate a situation image in which a distinction between the unlocker and a substitute with the unlocker is displayed on the monitoring image.

Images