JP7009021B2 - Abnormal person prediction system, abnormal person prediction method, and program - Google Patents

Description

The present invention relates to an abnormal person prediction system, an abnormal person prediction method, and a program capable of predicting an abnormal person in a specific area such as a facility.

A suspicious person detection system has been proposed that can monitor a wide range and detect all of them even if a plurality of suspicious persons invade (Patent Document 1).

Japanese Patent Application Laid-Open No. 4-264522

Further, supervised learning is a well-known method as a machine learning method for artificial intelligence to perform image analysis.

However, in the method of Patent Document 1, since the collation with the database for registering the characteristics of human beings entering and exiting the monitored area is performed, it takes time and effort to prepare the database in advance. In addition, although it is effective when only specific people enter and leave the monitored area, there is a problem that it cannot be used in areas where an unspecified number of people enter and exit, such as inside Shinkansen trains and commercial facilities. ..

In addition, when supervised learning is performed for image analysis, a large number of images must be prepared for learning, and when detecting a suspicious person or an abnormal person, a sufficient suspicious person for the learning is performed. It is difficult to prepare a sample image of an abnormal person.

For this problem, the present inventor is relatively easy to prepare a normal behavior pattern for learning, and is suspicious if either a normal behavior pattern or an abnormal behavior pattern is memorized. We focused on the fact that we can predict people and abnormal people.

The present invention is an abnormal person prediction system and an abnormal person prediction method capable of predicting an abnormal person in a specific area such as a facility by memorizing either a normal behavior pattern or an abnormal behavior pattern. , And the purpose of providing the program.

The present invention provides the following solutions.

The invention according to the first feature is
It is an abnormal person prediction system that predicts whether the monitored person in a predetermined area is an abnormal person.
A camera means provided in the predetermined area and for photographing the monitored person,
Behavioral pattern storage means for memorizing normal or abnormal behavioral patterns, and
An abnormal person prediction means for predicting an abnormal person based on the above-mentioned behavior pattern,
Equipped with
The abnormal person prediction means calculates the probability of being an abnormal person according to the degree of deviation from the normal behavior pattern or the degree of corresponding to the abnormal behavior pattern, and the importance according to the calculated probability. Provides an abnormal person prediction system that changes and warns .

According to the invention according to the first feature, in an abnormal person prediction system for predicting whether a monitored person in a predetermined area is an abnormal person, the abnormal person is provided in the predetermined area, and the monitored person is photographed and normally performed. Abnormal behavior pattern or abnormal behavior pattern is memorized , an abnormal person is predicted based on the behavior pattern, and the abnormality is determined according to the degree of deviation from the normal behavior pattern or the degree of corresponding to the abnormal behavior pattern. The probability of being a person is calculated, and the importance is changed according to the calculated probability to warn .

The invention according to the first feature is in the category of the abnormal person prediction system, but the same operation and effect can be obtained even with the abnormal person prediction method and the program.

The invention according to the second feature is the anomalous person prediction system which is the invention according to the first feature.
The behavior pattern is a behavior pattern in which the monitored person performs a predetermined action a predetermined number of times.
The abnormal person prediction means provides an abnormal person prediction system for predicting an abnormal person based on the behavior pattern.

According to the invention according to the second feature, in the abnormal person prediction system according to the first feature, the behavior pattern is a behavior pattern in which the monitored person performs a predetermined action a predetermined number of times. , The abnormal person prediction means predicts an abnormal person based on the behavior pattern.

The invention according to the third feature is an abnormal person prediction system which is an invention according to the first feature or the second feature.
The behavior pattern is a behavior pattern in which the monitored person stays at a predetermined place for a predetermined time.
The abnormal person prediction means provides an abnormal person prediction system for predicting an abnormal person based on the behavior pattern.

According to the invention according to the third feature, in the abnormal person prediction system according to the first feature or the second feature, the behavior pattern means that the monitored person is at a predetermined place for a predetermined time. It is a behavior pattern, and the abnormal person prediction means predicts an abnormal person based on the behavior pattern.

The invention according to the fourth feature is an abnormal person prediction system which is an invention according to any one of the first feature to the third feature.
The abnormal person prediction means provides an abnormal person prediction system for predicting an abnormal person based on the baggage, clothes, and wears owned by the monitored person in addition to the behavior pattern.

According to the invention according to the fourth feature, in the abnormal person prediction system which is the invention according to any one of the first feature to the third feature, the abnormal person prediction means is added to the behavior pattern and the monitoring target. Predict an abnormal person based on the luggage, clothes, and wears owned by the person.

The invention according to the fifth feature is an abnormal person prediction system which is an invention according to any one of the first to fourth features.
Based on the data on whether or not the monitored person predicted by the abnormal person predicting means was an abnormal person as a result.
Provided is an abnormal person prediction system further provided with an abnormal person learning means for learning the prediction of an abnormal person.

According to the invention according to the fifth feature, in the abnormal person prediction system which is the invention according to any one of the first feature to the fourth feature, the monitored person predicted by the abnormal person prediction means is eventually found. Further provided with an abnormal person learning means for learning the prediction of an abnormal person based on the data of whether or not the person was an abnormal person.

The invention according to the sixth feature is an abnormal person prediction system which is an invention according to any one of the first to fifth features.
The behavior pattern storage means stores a normal behavior pattern or an abnormal behavior pattern for each type of the predetermined region, and stores the normal behavior pattern or the abnormal behavior pattern.
The abnormal person prediction means provides an abnormal person prediction system for predicting an abnormal person based on an action pattern according to the type of the predetermined area.

According to the invention according to the sixth feature, in the abnormal person prediction system which is the invention according to any one of the first feature to the fifth feature, the behavior pattern storage means is normal for each type of the predetermined region. The behavior pattern or the abnormal behavior pattern is memorized, and the abnormal person prediction means predicts the abnormal person based on the behavior pattern corresponding to the type of the predetermined area.

The invention according to the seventh feature is
It is an abnormal person prediction method executed by an abnormal person prediction system that predicts whether the monitored person in a predetermined area is an abnormal person.
A step of photographing the monitored person provided in the predetermined area, and
Steps to memorize normal or abnormal behavior patterns,
Steps to predict an abnormal person based on the above behavior pattern,
When predicting an abnormal person, the probability of being an abnormal person is calculated according to the degree of deviation from the normal behavior pattern or the degree of corresponding to the abnormal behavior pattern, and the importance is calculated according to the calculated probability. Steps to change and warn, and
Provide a method for predicting an abnormal person.

The invention according to the eighth feature is
In the abnormal person prediction system that predicts whether the monitored person in the predetermined area is an abnormal person,
A step provided in the predetermined area to photograph the monitored person,
Steps to memorize normal or abnormal behavior patterns,
Steps to predict an abnormal person based on the behavior pattern,
When predicting an abnormal person, the probability of being an abnormal person is calculated according to the degree of deviation from the normal behavior pattern or the degree of corresponding to the abnormal behavior pattern, and the importance is calculated according to the calculated probability. Steps to change and warn,
Provides a computer-readable program for running.

According to the present invention, an abnormal person prediction system and an abnormal person who can predict an abnormal person in a specific area such as a facility by memorizing either a normal behavior pattern or an abnormal behavior pattern. It becomes possible to provide a prediction method and a program.

FIG. 1 is a schematic diagram of a preferred embodiment of the present invention. FIG. 2 is a diagram showing the relationship between the functional blocks of the camera 100 and the computer 200 and each function. FIG. 3 is a flowchart of the abnormal person prediction process. FIG. 4 is a diagram showing the relationship between the functional blocks of the camera 100 and the computer 200 and each function when learning an abnormal person is performed. FIG. 5 is a flowchart of an abnormal person prediction process when learning an abnormal person. FIG. 6 is a detailed flowchart of abnormal person prediction. FIG. 7 is an example of a screen of an abnormal person prediction system when the type of a predetermined area is the Shinkansen. FIG. 8 is an example of a screen of an abnormal person prediction system when a predetermined area is an office. FIG. 9 is an example of the configuration of a database of normal behavior patterns. FIG. 10 is an example of the configuration of a database of normal behavior patterns.

Hereinafter, the best mode for carrying out the present invention will be described with reference to the drawings. It should be noted that this is only an example, and the technical scope of the present invention is not limited to this.

[Overview of abnormal person prediction system]
FIG. 1 is a schematic diagram of a preferred embodiment of the present invention. The outline of the present invention will be described with reference to FIG. The abnormal person prediction system includes a camera 100, a computer 200, and a communication network 300.

In FIG. 1, the number of cameras 100 is not limited to one, and may be plural. Further, the computer 200 is not limited to an existing device, but may be a virtual device.

As shown in FIG. 2, the camera 100 includes a camera unit 10, a control unit 110, a communication unit 120, and a storage unit 130. Further, as shown in FIG. 2, the computer 200 is also composed of a control unit 210, a communication unit 220, a storage unit 230, and an input / output unit 240. The control unit 210 cooperates with the communication unit 220, the storage unit 230, and the input / output unit 240 to realize the abnormal person prediction module 211. Further, the storage unit 230 cooperates with the control unit 210, the communication unit 220, and the input / output unit 240 to realize the action pattern storage module 231. The communication network 300 may be a public communication network such as the Internet or a dedicated communication network, and enables communication between the camera 100 and the computer 200.

The camera 100 is an image pickup device provided with an image pickup device such as an image pickup element and a lens capable of data communication with the computer 200, and captures an image in a predetermined area to be monitored. Here, a security camera is shown as an example, but it is an image pickup device equipped with necessary functions such as a digital camera, a digital video, a camera mounted on a drone, a wearable device camera, a WEB camera, an in-vehicle camera, and a 360-degree camera. It may be there. Further, the captured image may be stored in the storage unit 130.

The computer 200 is a computing device capable of data communication with the camera 100. Here, a desktop computer is illustrated as an example, but in addition to a mobile phone, a mobile information terminal, a tablet terminal, and a personal computer, an electric appliance, a wearable terminal, or the like may be used. Further, the device is not limited to an existing device, and may be a virtual device.

In the abnormal person prediction system of FIG. 1, first, the behavior pattern storage module 231 stores a normal behavior pattern or an abnormal behavior pattern in the storage unit 230 of the computer 200 (step S01). In the present invention, the behavior pattern means, for example, that a person to be monitored performs a predetermined action a predetermined number of times or stays at a predetermined place for a predetermined time. In addition, the behavior pattern shall include the luggage and clothing owned by the person to be monitored, and the wear such as hats, sunglasses and masks. These behavior patterns may be acquired from another computer or a storage medium, or the behavior patterns may be extracted based on an image taken by the camera 100 and created by the computer 200. This step S01 can be omitted if a sufficient behavior pattern is already stored.

9 and 10 are examples of the configuration of a database of normal behavior patterns. Here, for convenience of drawing, the description is divided into FIGS. 9 and 10, but all the items of behavior, predetermined place, and appearance in FIGS. 9 and 10 are retained for each type of predetermined area. It may be a configuration, or a configuration in which only one of the items is held for each type of a predetermined area may be used. The predetermined area is a place to be monitored, and by preparing a normal behavior pattern or an abnormal behavior pattern for each type, it becomes possible to more appropriately predict an abnormal person. In the examples of FIGS. 9 and 10, the types of predetermined areas are Shinkansen (unreserved seats), Shinkansen (reserved seats), stations (home), stations (ticket gates), offices (desk work), offices (sales / planning), and offices. (Development), office (meeting room), office (entrance), office (elevator), hospital (sickroom), hospital (entrance), nursing facility (living room), nursing facility (entrance) are described. In this way, by separating unreserved seats and reserved seats on the same Shinkansen, or separating platforms and ticket gates at stations, it is possible to memorize normal behavior patterns or abnormal behavior patterns suitable for the type of predetermined area. As a result, the accuracy of predicting abnormal persons can be improved. Further, although FIGS. 9 and 10 show examples of normal behavior patterns, it is normal that there are no abnormal persons photographed by the camera 100 according to the items such as behavior, predetermined place, and appearance prepared in advance. The image of the case may be analyzed to extract a normal behavior pattern. By regarding the case of deviating from the normal behavior pattern as abnormal, it is possible to deal with the problem that the sample image of the suspicious person is small and it is difficult to create the abnormal behavior pattern. .. When creating a database of abnormal behavior patterns, if an image with an abnormal person can be sufficiently prepared, the image may be analyzed to extract an abnormal behavior pattern.

Returning to FIG. 1, the abnormal person prediction module 211 of the computer 200 acquires an image of a predetermined area to be monitored from the camera 100 (step S02). In FIG. 1, only one camera 100 is shown, but when there are a plurality of cameras 100, an image is acquired from each of them. The image acquisition process shall be continuously performed while the abnormal person prediction system is being executed.

Next, the abnormal person prediction module 211 analyzes the acquired image and predicts the abnormal person (step S03). The abnormal person prediction module 211 detects a person from the acquired image, and tracks and monitors each of the detected people. That is, when a person X is detected at a certain point in time, the behavior patterns are continuously compared by regarding the person X as the same person until the person X disappears from the predetermined area. Here, an example in which the type of the predetermined area is the Shinkansen (unreserved seat), and a case where an abnormal person is predicted by comparing with the normal behavior patterns of FIGS. 9 and 10 will be described. When the type of the predetermined area is the Shinkansen (unreserved seat), the normal behavior pattern is whether the number of times to stand up is within 2 times per hour, the staying time in the predetermined area is within the one-way boarding time of the Shinkansen, and the posture stands. Sit down, phone time is less than 5 minutes. Also, if it is possible to continue monitoring even if you move from a predetermined area under monitoring to another place, the time spent in another place is, for example, 5 minutes or less for the deck, 5 minutes or less for the toilet, and a washroom. Is within 5 minutes. As for the appearance, the number of luggage is 3 or less, the size of the luggage is 80 cm or less, the clothes are seasonal, the sunglasses, hat, mask are specially treated, and the whole body is 2 or more colors. ,. The * mark in FIG. 10 indicates that special processing related to other items is required. It is determined that the special treatment for sunglasses, hats, and masks in FIG. 10 is normal when not all of them are "yes". In step S03, for example, if the number of times the person stands up is three times within one hour, the person is predicted as an abnormal person. Or, if you move to the bathroom for more than 5 minutes, if you wear sunglasses, a hat, and a mask, or if you wear only one color, you will be predicted as an abnormal person.

Returning to FIG. 1, finally, the abnormal person prediction module 211 of the computer 200 warns of the abnormal person based on the abnormal person prediction result in step S03 (step S04). As a warning method here, it may be output to the input / output unit 240 of the computer 200, or may be output to the input / output unit of another terminal or the like (not shown) designated by the user.

FIG. 7 is an example of a screen of an abnormal person prediction system when the type of a predetermined area is the Shinkansen. This screen may be displayed on the input / output unit 240 of the computer 200, or may be displayed on the input / output unit of another terminal (not shown) used by the user. The screen 701 of the abnormal person prediction system displays the monitoring location as the display 702, the image acquired as the display 703, the prediction history as the display 704, and the information of the abnormal person predicted as the display 705. In the screen of the display 703, the information of the predicted abnormal person is displayed as a balloon 708. Compared with the normal behavior patterns shown in FIGS. 9 and 10, in FIG. 7, the possibility of being an abnormal person and its main character are described as "possibility of abnormal person: 65%, standing 12 times in the seat". An example of displaying the reason is shown. The balloon 708 may be systematically configurable as to what percentage or more the probability of an abnormal person is displayed. In addition, it is desirable to display the color and shape of the balloon, the color, thickness, size, etc. of the characters so that the importance is easy to understand, as the possibility of being an abnormal person is high. Further, not only the display of the balloon 708 but also the warning sound may be output at the same time. The display and output method by the balloon 708 is not limited to the invention of the present application, and it is only necessary to confirm which person the abnormal person is predicted for and the content thereof. By selecting the information portion of the predicted abnormal person on the display 705 or the balloon 708, the behavior history that causes the problem of the person, for example, the time of leaving the seat or the length thereof may be displayed. Further, in the monitoring place of the display 702, in addition to the name and address of the specific monitoring destination, the type of the monitoring place applied for detecting the abnormal person may be displayed. In FIG. 7, the “Shinkansen (reserved seat)” is applied as the type of monitoring location. Here, only the display is performed, but the type of monitoring location may be switched by the user on the screen 701. Further, a switching button 706 is provided on the screen 701 of the abnormal person prediction system, and a camera 100 is installed in another vehicle, a deck, a toilet, a washroom, etc. in the same Shinkansen for monitoring. If it is possible, the screen display of the display 703 may be switched. In addition, it is possible to check the history of what kind of behavior the person A, who was predicted to have a 65% chance of being abnormal here, was taking within the range that can be monitored by the other camera 100. May be good. Further, the screen 701 of the abnormal person prediction system may be provided with a notification button 707 so that the abnormal person can be notified to the conductor or crew of the Shinkansen, the Shinkansen general command room, or the like. When making a report, the position and seat of the abnormal person, the possibility and reason for the abnormal person, the gender and appearance characteristics of the abnormal person, the image acquired from the camera 100, etc. may be notified at the same time. And.

As described above, according to the present invention, it is possible to predict an abnormal person in a specific area such as a facility by memorizing either a normal behavior pattern or an abnormal behavior pattern. It becomes possible to provide a system, an abnormal person prediction method, and a program.

[Explanation of each function]
FIG. 2 is a diagram showing the relationship between the functional blocks of the camera 100 and the computer 200 and each function. The camera 100 includes a camera unit 10, a control unit 110, a communication unit 120, and a storage unit 130. Further, the computer 200 is composed of a control unit 210, a communication unit 220, a storage unit 230, and an input / output unit 240. The control unit 210 cooperates with the communication unit 220, the storage unit 230, and the input / output unit 240 to realize the abnormal person prediction module 211. Further, the storage unit 230 cooperates with the control unit 210, the communication unit 220, and the input / output unit 240 to realize the action pattern storage module 231. The communication network 300 may be a public communication network such as the Internet or a dedicated communication network, and enables communication between the camera 100 and the computer 200.

The camera 100 is an image pickup device provided with an image pickup device such as an image pickup element and a lens capable of data communication with the computer 200, and captures an image in a predetermined area to be monitored. Here, a security camera is shown as an example, but it is an image pickup device equipped with necessary functions such as a digital camera, a digital video, a camera mounted on a drone, a wearable device camera, a WEB camera, an in-vehicle camera, and a 360-degree camera. It may be there. Further, the captured image may be stored in the storage unit 130.

The camera 100 includes a lens, an image sensor, various buttons, an image pickup device such as a flash, and the like as the camera unit 10, and captures an image as an image captured image such as a moving image or a still image. Further, the image obtained by imaging is a precise image having an amount of information necessary for image analysis.

The control unit 110 includes a CPU (Central Processing Unit), a RAM (Random Access Memory), a ROM (Read Only Memory), and the like.

As the communication unit 120, a device for enabling communication with other devices, for example, a WiFi (Wireless WiFi) compatible device compliant with IEEE802.11 or an IMT-2000 standard such as a 3rd generation or 4th generation mobile communication system. Equipped with compliant wireless devices, etc. It may be a wired LAN connection.

The storage unit 130 includes a data storage unit using a hard disk or a semiconductor memory, and stores necessary data such as a captured image. The purpose of image analysis and the shooting environment of the image may also be stored.

The computer 200 is a computing device capable of data communication with the camera 100. Here, a desktop computer is illustrated as an example, but in addition to a mobile phone, a mobile information terminal, a tablet terminal, and a personal computer, an electric appliance, a wearable terminal, or the like may be used. Further, the device is not limited to an existing device, and may be a virtual device.

The control unit 210 includes a CPU, RAM, ROM, and the like. The control unit 210 cooperates with the communication unit 220, the storage unit 230, and the input / output unit 240 to realize the abnormal person prediction module 211.

As the communication unit 220, a device for enabling communication with other devices, for example, a WiFi compatible device compliant with IEEE802.11 or a wireless device compliant with the IMT-2000 standard such as a 3rd generation or 4th generation mobile communication system. Etc. are provided. It may be a wired LAN connection. Further, it is assumed that communication is performed with another terminal (not shown) used by the user as necessary via the communication unit 220.

The storage unit 230 includes a data storage unit using a hard disk or a semiconductor memory, and stores captured images and data necessary for processing such as normal behavior patterns or abnormal behavior patterns. The storage unit 230 cooperates with the control unit 210, the communication unit 220, and the input / output unit 240 to realize the action pattern storage module 231.

The input / output unit 240 is provided with functions necessary for using the abnormal person prediction system. As an example for realizing input, it is possible to provide a liquid crystal display that realizes a touch panel function, a keyboard, a mouse, a pen tablet, a hardware button on the device, a microphone for performing voice recognition, and the like. Further, as an example for realizing the output, a form such as a display such as a liquid crystal display, a display of a PC, a projection on a projector, and an audio output can be considered. The function of the present invention is not particularly limited by the input / output method. Further, input / output can be performed from an input / output unit of another terminal (not shown) used by the user as needed.

[Abnormal person prediction processing]
FIG. 3 is a flowchart of the abnormal person prediction process. The process executed by each of the above-mentioned modules will be described in accordance with this process.

The behavior pattern storage module 231 of the computer 200 stores a normal behavior pattern or an abnormal behavior pattern in the storage unit 230 of the computer 200 (step S301). In the present invention, the behavior pattern means, for example, that a person to be monitored performs a predetermined action a predetermined number of times or stays at a predetermined place for a predetermined time. In addition, the behavior pattern shall include the luggage and clothing owned by the person to be monitored, and the wear such as hats, sunglasses and masks. The behavior pattern may be acquired from another computer or a storage medium, or the behavior pattern may be extracted based on the image taken by the camera 100 and created by the computer 200. This step S301 can be omitted if a sufficient behavior pattern is already stored.

9 and 10 are examples of the configuration of a database of normal behavior patterns. Here, for convenience of drawing, the description is divided into FIGS. 9 and 10, but all the items of behavior, predetermined place, and appearance in FIGS. 9 and 10 are retained for each type of predetermined area. It may be a configuration, or a configuration in which only one of the items is held for each type of a predetermined area may be used. The predetermined area is a place to be monitored, and by preparing a normal behavior pattern or an abnormal behavior pattern for each type, it becomes possible to more appropriately predict an abnormal person. In the examples of FIGS. 9 and 10, the types of predetermined areas are Shinkansen (unreserved seats), Shinkansen (reserved seats), stations (home), stations (ticket gates), offices (desk work), offices (sales / planning), and offices. (Development), office (meeting room), office (entrance), office (elevator), hospital (sickroom), hospital (entrance), nursing facility (living room), nursing facility (entrance) are described. In this way, by separating unreserved seats and reserved seats on the same Shinkansen, or separating platforms and ticket gates at stations, it is possible to memorize normal behavior patterns or abnormal behavior patterns suitable for the type of predetermined area. As a result, the accuracy of predicting abnormal persons can be improved. Further, although FIGS. 9 and 10 show examples of normal behavior patterns, it is normal that there are no abnormal persons photographed by the camera 100 according to the items such as behavior, predetermined place, and appearance prepared in advance. The image of the case may be analyzed to extract a normal behavior pattern. By regarding the case of deviating from the normal behavior pattern as abnormal, it is possible to deal with the problem that the sample image of the suspicious person is small and it is difficult to create the abnormal behavior pattern. .. When creating a database of abnormal behavior patterns, if an image with an abnormal person can be sufficiently prepared, the image may be analyzed to extract an abnormal behavior pattern.

Returning to FIG. 3, the abnormal person prediction module 211 of the computer 200 makes an image transmission request to the camera 100 in a predetermined area to be monitored (step S302). In FIG. 2, only one camera 100 is shown, but when there are a plurality of cameras 100, an image is acquired from each of them. The image transmission request in step S302 may be made once when the abnormal person prediction system is executed, and then the images may be continuously acquired until the end instruction is given.

Upon receiving the image transmission request from the computer 200, the camera 100 transmits an image of a predetermined area captured by the camera unit 10 to the computer 200 via the communication unit 120 (step S303).

The abnormal person prediction module 211 of the computer 200 acquires an image from the camera 100 (step S304). The image transmission process by the camera 100 in step S303 and the image acquisition process of the computer 200 in step S304 shall be continuously performed while the abnormal person prediction system is being executed.

Next, the abnormal person prediction module 211 analyzes the acquired image and predicts the abnormal person (step S305). The details of the abnormal person prediction will be described with reference to FIG.

FIG. 6 is a detailed flowchart of abnormal person prediction. First, the abnormal person prediction module 211 detects a person from the acquired image (step S601). The number of people to be detected is not limited to one, and may be multiple or zero.

Next, the abnormal person prediction module 211 tracks the detected person (step S602). Specifically, tracking is to consider the continuity with the next acquired image. In the continuous acquired images, it is determined whether or not they are the same person, and if they are the same person, the actions of that person are accumulated as a history. By doing so, it is possible to determine the number of times the person leaves the seat, the staying time, the telephone time, and the like in the behavior pattern shown in FIG. In addition, even when the appearance of the same person changes, it is possible to predict that the person is an abnormal person. That is, when a person X is detected at a certain point in time, the behavior patterns are continuously compared by regarding the person X as the same person until the person X disappears from the predetermined area.

Next, the abnormal person prediction module 211 determines whether to use a normal behavior pattern or an abnormal behavior pattern (step S603). As the determination method here, when either the normal behavior pattern or the abnormal behavior pattern is present in the storage unit 230, the existing behavior pattern is used. If both normal and abnormal behavior patterns exist, the one with the behavior pattern suitable for the type of predetermined area to be monitored may be used, or when the behavior pattern is memorized. The one with a large amount of data based on the above may be used, or the one selected by the user may be used.

When the normal behavior pattern is decided to be used, the abnormal person prediction module 211 determines whether or not the behavior of the person of interest is within the range of the normal behavior pattern (step S604). If it is within the range of the normal behavior pattern, the process returns to step S601 and the process is continued. If it is out of the range of the normal behavior pattern, the process proceeds to the next step S606.

When the abnormal behavior pattern is decided to be used, the abnormal person prediction module 211 determines whether or not the behavior of the person of interest is within the range of the abnormal behavior pattern (step S605). If it is within the range of the abnormal behavior pattern, the process proceeds to the next step S606. If it is out of the range of the abnormal behavior pattern, the process returns to step S601 to continue the process.

Next, the abnormal person prediction module 211 predicts the person of interest as an abnormal person based on the determination result in step S604 or step S605 (step S606). At this time, the probability of being an abnormal person may be calculated depending on how much the person of interest deviates from the normal behavior pattern or how much the person corresponds to the abnormal behavior pattern. The probability may be determined from the number of items that deviate from the normal behavior pattern or the number of items that correspond to the abnormal behavior pattern. Further, when the number of times and the time of the behavior pattern deviate greatly from the predetermined value, the probability of being an abnormal person may be calculated as a higher probability. The above is the detailed flow of prediction of abnormal persons.

Returning to FIG. 3, finally, the abnormal person prediction module 211 of the computer 200 warns the abnormal person based on the abnormal person prediction result in step 305 (step S306). As a warning method here, it may be output to the input / output unit 240 of the computer 200, or may be output to the input / output unit of another terminal or the like (not shown) designated by the user.

FIG. 8 is an example of a screen of an abnormal person prediction system when a predetermined area is an office. This screen may be displayed on the input / output unit 240 of the computer 200, or may be displayed on the input / output unit of another terminal (not shown) used by the user. The screen 801 of the abnormal person prediction system displays the monitoring location as the display 802, the image acquired as the display 803, the prediction history as the display 804, and the information of the abnormal person predicted as the display 805. In the screen of the display 703, the information of the predicted abnormal person is displayed as a balloon 808. Compared with the normal behavior patterns shown in FIGS. 9 and 10, in FIG. 8, it is possible to be an abnormal person as "possibility of abnormal person: 85%, sunglasses: yes, clothes: out of range of office worker". An example of displaying gender and its main reasons is shown. The balloon 808 may be systematically configurable as to what percentage or more the probability of an abnormal person is displayed. In addition, it is desirable to display the color and shape of the balloon, the color, thickness, size, etc. of the characters so that the importance is easy to understand, as the possibility of being an abnormal person is high. For example, the possibility of the abnormal person in the balloon 708 in FIG. 7 is 65%, whereas the possibility of the abnormal person in the balloon 808 in FIG. 8 is 85%, so that attention is paid more. It has a shape. Further, not only the display of the balloon 808 but also the warning sound may be output at the same time. The display and output method by the balloon 808 is not limited to the present invention, and it is only necessary to confirm which person the abnormal person is predicted for and the content thereof. By selecting the information portion of the predicted abnormal person on the display 805 or the balloon 808, the behavior history that causes the problem of the person, for example, the time and the length of the appearance in the predetermined monitoring target area may be displayed. Further, in the monitoring place of the display 802, in addition to the name and address of the specific monitoring destination, the type of the monitoring place applied for detecting the abnormal person may be displayed. In FIG. 8, “office (desk work)” is applied as the type of monitoring location. Here, only the display is performed, but the type of monitoring location may be switched by the user on the screen 801. Further, a tracking button 806 may be provided on the screen 801 of the abnormal person prediction system so that the camera 100 may follow and monitor the selected abnormal person. In addition, it is possible to check the history of what kind of behavior the person B, who was predicted to have an abnormal person with an abnormality of 85%, took in the range that can be monitored by the other camera 100. May be good. In addition, a report button 807 is provided on the screen 801 of the abnormal person prediction system, and the security guard room, the security company, the person in charge of the office X as the monitoring place, the person designated as the target of the report, and the entire office X are provided. For example, it may be possible to report an abnormal person. When making a report, the position and seat of the abnormal person, the possibility and reason for the abnormal person, the gender and appearance characteristics of the abnormal person, the image acquired from the camera 100, etc. may be notified at the same time. And. Here, when the report button 807 is selected from the display on the screen 801, the security guard room, the security company, the person in charge of the office X as the monitoring place, the person designated as the target of the report, the entire office X, etc. Although an example of reporting an abnormal person has been described in the above, the report may be made without displaying the screen 801 in step S306.

As described above, according to the present invention, it is possible to predict an abnormal person in a specific area such as a facility by memorizing either a normal behavior pattern or an abnormal behavior pattern. It becomes possible to provide a system, an abnormal person prediction method, and a program.

[Abnormal person learning process]
FIG. 4 is a diagram showing the relationship between the functional blocks of the camera 100 and the computer 200 and each function when learning an abnormal person is performed. In addition to the configuration of FIG. 2, the control unit 210 of the computer 200 cooperates with the communication unit 220, the storage unit 230, and the input / output unit 240 to realize the abnormal person learning module 212. FIG. 5 is a flowchart of an abnormal person prediction process when learning an abnormal person. The process executed by each of the above-mentioned modules will be described in accordance with this process. Since the process from step S501 to step S506 in FIG. 5 corresponds to the process from step S301 to step S306 in FIG. 3, steps S507 and subsequent steps will be described. It should be noted that the process of step S501 can be omitted if a sufficient action pattern is already stored, as in step S301.

The abnormal person learning module 212 of the computer 200 receives feedback data for the abnormal person warning in step S506 (step S507). As a method for receiving the feedback data, for example, the screen 701 of FIG. 7 or the screen 801 of FIG. 8 may be provided with an interface for the user to transmit the feedback data. For a person who has been shown the possibility of an abnormal person, display the question "Did this person need to be vigilant as an abnormal person?" And get a "yes" or "no" answer to the feedback data. You will be able to receive it.

Next, the abnormal person learning module 212 learns using the abnormal person prediction in step S505 and the feedback data in step S507 (step S508). For example, in the example of FIG. 7, if the answer to the question "Did this person need to be vigilant as an abnormal person?" To the balloon 708 is "No", the number of times of leaving the seat is not so important. It is possible. Also, in the example of FIG. 8, if the answer to the question "Did this person need to be vigilant as an abnormal person?" To the balloon 808 is "Yes", the person whose clothes are outside the range of the office worker, or It is possible to learn that it is better to increase the possibility of being an abnormal person for a person wearing sunglasses.

Finally, the abnormal person learning module 212 updates the behavior pattern database if necessary according to the learning result of step S508 (step S509). For example, in the example of FIG. 7 above, the learning result that the number of times of leaving the seat is not so important, as the answer to the question "Did this person need to be vigilant as an abnormal person?" Was "No". In the database of normal behavior patterns of the Shinkansen (reserved seats) in FIG. 9, when the number of seats is accumulated, the number of times of seating is increased from 3 times / hour and updated.

In this way, by receiving feedback data for abnormal person prediction and performing learning, it is possible to perform abnormal person prediction with higher accuracy according to the type of a predetermined area, an abnormal person prediction system, an abnormal person prediction method, and a program. Can be provided.

According to the present invention, an abnormal person prediction system capable of predicting an abnormal person in a specific area such as a facility with high accuracy by memorizing either a normal behavior pattern or an abnormal behavior pattern. , Abnormal person prediction methods, and programs can be provided.

The above-mentioned means and functions are realized by a computer (including a CPU, an information processing device, and various terminals) reading and executing a predetermined program. The program may be, for example, in the form of being provided from a computer via a network (SaaS: software as a service), or may be a flexible disk, a CD (CD-ROM, etc.), or a DVD (DVD-ROM, DVD). -RAM, etc.), provided in the form of being recorded on a computer-readable recording medium such as a compact memory. In this case, the computer reads the program from the recording medium, transfers it to the internal storage device or the external storage device, stores it, and executes it. Further, the program may be recorded in advance in a storage device (recording medium) such as a magnetic disk, an optical disk, or a magneto-optical disk, and the program may be provided from the storage device to a computer via a communication line.

Although the embodiments of the present invention have been described above, the present invention is not limited to these embodiments described above. Further, the effects described in the embodiments of the present invention merely list the most suitable effects arising from the present invention, and the effects according to the present invention are limited to those described in the embodiments of the present invention. is not it.

100 cameras, 200 computers, 300 communication networks

Claims (8)

It is an abnormal person prediction system that predicts whether the monitored person in a predetermined area is an abnormal person.
A camera means provided in the predetermined area and for photographing the monitored person,
Behavioral pattern storage means for memorizing normal or abnormal behavioral patterns, and
An abnormal person prediction means for predicting an abnormal person based on the above-mentioned behavior pattern,
Equipped with
The abnormal person prediction means calculates the probability of being an abnormal person according to the degree of deviation from the normal behavior pattern or the degree of corresponding to the abnormal behavior pattern, and the importance according to the calculated probability. Abnormal person prediction system that changes and warns . The behavior pattern is a behavior pattern in which the monitored person performs a predetermined action a predetermined number of times.
The abnormal person prediction system according to claim 1, wherein the abnormal person predicting means predicts an abnormal person based on the behavior pattern. The behavior pattern is a behavior pattern in which the monitored person stays at a predetermined place for a predetermined time.
The abnormal person prediction system according to claim 1 or 2, wherein the abnormal person predicting means predicts an abnormal person based on the behavior pattern. The invention according to any one of claims 1 to 3, wherein the abnormal person predicting means predicts an abnormal person based on the baggage, clothes, and wears owned by the monitored person in addition to the behavior pattern. Abnormal person prediction system. Based on the data on whether or not the monitored person predicted by the abnormal person predicting means was an abnormal person as a result.
The abnormal person prediction system according to any one of claims 1 to 4, further comprising an abnormal person learning means for learning the prediction of an abnormal person. The behavior pattern storage means stores a normal behavior pattern or an abnormal behavior pattern for each type of the predetermined region, and stores the normal behavior pattern or the abnormal behavior pattern.
The abnormal person prediction system according to any one of claims 1 to 5, wherein the abnormal person predicting means predicts an abnormal person based on an action pattern according to the type of the predetermined area. It is an abnormal person prediction method executed by an abnormal person prediction system that predicts whether the monitored person in a predetermined area is an abnormal person.
A step of photographing the monitored person provided in the predetermined area, and
Steps to memorize normal or abnormal behavior patterns,
Steps to predict an abnormal person based on the above behavior pattern,
When predicting an abnormal person, the probability of being an abnormal person is calculated according to the degree of deviation from the normal behavior pattern or the degree of corresponding to the abnormal behavior pattern, and the importance is calculated according to the calculated probability. Steps to change and warn , and
Abnormal person prediction method. In the abnormal person prediction system that predicts whether the monitored person in the predetermined area is an abnormal person,
A step provided in the predetermined area to photograph the monitored person,
Steps to memorize normal or abnormal behavior patterns,
Steps to predict an abnormal person based on the behavior pattern,
When predicting an abnormal person, the probability of being an abnormal person is calculated according to the degree of deviation from the normal behavior pattern or the degree of corresponding to the abnormal behavior pattern, and the importance is calculated according to the calculated probability. Steps to change and warn ,
A computer-readable program for running.

Images