JP2018156586A - Monitoring system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a monitoring system for detecting an object possessed by a person to be detected from a terahertz image without stopping the person to be detected at a predetermined position.SOLUTION: A monitoring system includes a first image acquisition unit (a terahertz image acquisition unit 101), a second image acquisition unit (a visualized image acquisition unit 102), a person detection unit 103, an object detection unit 105, an object classification unit 106, and a display unit 108. The first image acquisition unit receives a terahertz wave emitted from a person passing through a predetermined position and acquires a first image in which the predetermined position is imaged based on the reception result of the terahertz wave. The second image acquisition unit receives light reflected at the predetermined position and acquires a second image in which the predetermined position is imaged based on the light reception result of the light. The person detection unit detects a person from the second image. The object detection unit detects an object possessed by the person from a first position corresponding to the detected person in the first image.SELECTED DRAWING: Figure 1

Description

  Embodiments described herein relate generally to a monitoring system.

  The person is stopped at a predetermined position such as a gate that manages the entrance and exit of the person to the building and the exit of the escalator, the person is irradiated with the terahertz wave, and the person is reflected based on the result of receiving the terahertz wave reflected by the person. Technology has been developed to detect objects held by. In addition, a technique has been developed in which a person is stopped at a predetermined position and an object is detected using electromagnetic induction using a handheld device.

Japanese Patent No. 4751332 Japanese Patent No. 5072972 Special table 2009-519436

  However, in the above technique, it is necessary to temporarily stop a person at a predetermined position and detect an object possessed by the person, and the passage of the person at the predetermined position may be delayed.

  The monitoring system of the embodiment includes a first image acquisition unit, a second image acquisition unit, a person detection unit, an object detection unit, a classification unit, and a display unit. The first image acquisition unit receives a terahertz wave emitted from a person passing through a predetermined position, and acquires a first image obtained by imaging the predetermined position based on a reception result of the terahertz wave. The second image acquisition unit receives the light reflected at the predetermined position, and acquires a second image obtained by imaging the predetermined position based on the light reception result of the light. The person detection unit detects a person from the second image. The object detection unit detects an object possessed by the person from a first position corresponding to the detected person in the first image. The classification unit executes a classification process for determining whether or not the detected object is a predetermined object. The display unit displays the execution result of the classification process and the alarm level of the detected object based on the execution result.

FIG. 1 is a diagram illustrating an example of a configuration of a monitoring system according to the first embodiment. FIG. 2 is a flowchart illustrating an example of a flow of classification processing of an object possessed by a detection target person in the monitoring system according to the first embodiment. FIG. 3 is a flowchart illustrating an example of a flow of processing for displaying an alarm level in the monitoring system according to the first embodiment. FIG. 4 is a diagram illustrating a display example of an alarm level by the monitoring system according to the first embodiment. FIG. 5 is a diagram illustrating an example of a configuration of a monitoring system according to the second embodiment. FIG. 6 is a flowchart illustrating an example of a flow of classification processing of an object possessed by a person to be detected in the monitoring system according to the second embodiment. FIG. 7 is a diagram illustrating an example of a configuration of a monitoring system according to the third embodiment. FIG. 8 is a view of an example of the arrangement positions of the terahertz image acquisition unit and the visualized image acquisition unit of the monitoring system according to the third embodiment as viewed from above. FIG. 9 is a side view of an example of the arrangement positions of the terahertz image acquisition unit and the visualization image acquisition unit of the monitoring system according to the third embodiment. FIG. 10 is a flowchart illustrating an example of a flow of classification processing of an object possessed by a detection target person in the monitoring system according to the third embodiment. FIG. 11 is a flowchart illustrating an example of a flow of processing for displaying an alarm level in the monitoring system according to the third embodiment. FIG. 12 is a diagram illustrating a display example of an alarm level by the monitoring system according to the third embodiment. FIG. 13 is a diagram illustrating a display example of an alarm level by the monitoring system according to the third embodiment.

  Hereinafter, the monitoring system according to the present embodiment will be described with reference to the accompanying drawings.

(First embodiment)
FIG. 1 is a diagram illustrating an example of a configuration of a monitoring system according to the first embodiment. As shown in FIG. 1, the monitoring system according to the present embodiment includes a terahertz image acquisition unit 101, a visualized image acquisition unit 102, a person detection unit 103, an image processing unit 104, an object detection unit 105, and an object classification. Unit 106, detected object database 107, and display unit 108.

  The terahertz image acquisition unit 101 (an example of a first image acquisition unit) emits from a person passing through a predetermined position (hereinafter referred to as a detection target person) such as a gate that manages entry / exit of a person to / from a building and an exit of an escalator. And a sensor for receiving terahertz waves. Then, the terahertz image acquisition unit 101 acquires an image obtained by imaging a predetermined position (hereinafter referred to as a terahertz image, an example of a first image) based on the reception result of the terahertz wave. This eliminates the need for the terahertz image acquisition unit 101 to irradiate the detection target person with the terahertz wave, thereby reducing the number of components constituting the monitoring system.

  In the present embodiment, the terahertz image acquisition unit 101 uses a passive system that receives a terahertz wave emitted from a detection target person. However, the terahertz wave is irradiated to the detection target person and reflected by the detection target person. It is also possible to use an active system that receives terahertz waves. In this embodiment, the monitoring system keeps the outside air temperature at the installation position of the terahertz image acquisition unit 101 within a predetermined temperature range (for example, around 20 ° C.) in which the reception accuracy of the terahertz wave emitted from the person is high. You may have an air-conditioning installation (for example, cooling element).

  The visualized image acquisition unit 102 (an example of a second image acquisition unit) is configured by a camera or the like, receives light reflected at a predetermined position, and images an image of the predetermined position based on the light reception result (hereinafter referred to as “light”). This is called a visualized image, which is an example of a second image. The person detection unit 103 detects a detection target person from the visualized image.

  The image processing unit 104 performs image processing such as filtering processing on the terahertz image so that an object held by the detection target person can be easily detected from the terahertz image.

  The object detection unit 105 detects an object possessed by the detection target person from a position corresponding to the detected detection target person (hereinafter referred to as a person position; an example of a first position) in the terahertz image subjected to image processing. . As a result, even if the detection target person is moving, the position of the detection target person in the terahertz image can be specified with high accuracy, so the detection target person possesses from the terahertz image without stopping the detection target person at the predetermined position. An object can be detected.

  The image processing unit 104 performs image processing on the terahertz image so that a relatively small object (for example, a folding knife) possessed by the person to be detected is easily detected. Thereby, a relatively small object possessed by the person to be detected can also be detected.

  In the present embodiment, the object detection unit 105 detects an object from a terahertz image that has been subjected to image processing. However, the present invention is not limited to this, and a terahertz image that has not been subjected to image processing (in other words, a terahertz image). You may detect an object from the terahertz image as it was acquired by the image acquisition part 101).

  The object classification unit 106 (an example of a classification unit) executes a classification process for determining whether or not the detected object is a predetermined object. Here, the predetermined object is an object that is prohibited from passing through a predetermined position. The detected object database 107 stores images (hereinafter referred to as model images) for each predetermined object classification. In the present embodiment, the object classification unit 106 determines that the detected object is a predetermined object when the detected object is similar to or matches the model image in the detected object database 107. In the present embodiment, the object classification unit 106 identifies the classification of the model image that is similar or coincident with the detected object as the classification of the detected object.

  The display unit 108 displays the execution result of the classification process by the object classification unit 106 and the alarm level of the detected object based on the execution result. In the present embodiment, when the display unit 108 determines that the detected object is a predetermined object, the display unit 108 changes the alarm level according to the classification of the detected object. In the present embodiment, the display unit 108 displays the terahertz image and the visualized image in addition to the classification process execution result and the alarm level.

  Next, with reference to FIG. 2, an example of the flow of the object classification process carried by the person to be detected in the monitoring system according to the present embodiment will be described. FIG. 2 is a flowchart illustrating an example of a flow of classification processing of an object possessed by a detection target person in the monitoring system according to the first embodiment.

  The terahertz image acquisition unit 101 receives a terahertz wave emitted from the detection target person at a preset cycle, and acquires a terahertz image based on the reception result of the terahertz wave (step S201). The visualized image acquisition unit 102 receives light reflected by the detection target person in synchronization with reception of the terahertz wave by the terahertz image acquisition unit 101, and acquires a visualized image based on the light reception result of the light (step S201). ).

  The image processing unit 104 performs image processing such as smoothing processing, binarization processing, and expansion / contraction processing on the terahertz image, and converts the terahertz image into an image in which an object included in the terahertz image is emphasized. (Step S202). The person detection unit 103 detects a detection target person from the visualized image (step S203). In the present embodiment, the person detection unit 103 creates an image obtained by averaging rectangular images in a plurality of person images as a template (hereinafter referred to as a person template). Then, the person detection unit 103 detects, as a detection target person, the position where the difference in luminance value with the person template is the smallest in the visualized image while the person template is translated, rotated, or enlarged / reduced in the visualized image. To do.

  The object detection unit 106 detects an object possessed by the person to be detected from the position of the person in the terahertz image that has been subjected to the image processing (step S204). As a result, even if the detection target person is moving, the position of the detection target person in the terahertz image can be specified with high accuracy, so the detection target person possesses from the terahertz image without stopping the detection target person at the predetermined position. An object can be detected. Here, the person position is a position corresponding to the detected person to be detected in the terahertz image. In the present embodiment, the object detection unit 105 sets the same label for consecutive pixels having the same pixel value at the person position of the terahertz image that has been subjected to the binarization process. Then, the object detection unit 105 detects, as an object possessed by the person to be detected, an isolated area smaller than a preset size among areas where labels are set at a person position in the terahertz image.

  The object classification unit 106 determines whether the object detection unit 105 has detected an object possessed by the person to be detected (step S205). When an object possessed by the person to be detected is not detected (step S205: No), the process returns to step S201, and the terahertz image acquisition unit 101 and the visualized image acquisition unit 102 acquire a new terahertz image and a visualized image.

  On the other hand, when an object possessed by the person to be detected is detected (step S205: Yes), the object classification unit 106 executes a classification process for determining whether the object is a predetermined object for each detected object. (Step S206). In the present embodiment, the object classification unit 106 repeats the classification process for all detected objects. The display unit 108 also displays the execution result of the classification process by the object classification unit 106 and the alarm level of the detected object based on the execution result (step S206).

  Specifically, the object classification unit 106 calculates the degree of similarity between the detected object and each model image stored in the detected object database 107. Then, the object classification unit 106 determines that the detected object is a predetermined object when the similarity between the detected object and at least one model image is equal to or higher than a preset similarity. When the similarity between the detected object and the plurality of model images is equal to or higher than a preset similarity, the object classification unit 106 has the highest similarity with the detected object among the plurality of model images. Is identified as the classification of the detected object. Alternatively, the object classification unit 106 uses a deep learning or the like to construct a classification method that receives a terahertz image and outputs a classification of the detected object, and identifies the classification of the detected object using the classification method. May be.

  Next, an example of a processing flow for displaying an alarm level in the monitoring system according to the present embodiment will be described with reference to FIG. FIG. 3 is a flowchart illustrating an example of a flow of processing for displaying an alarm level in the monitoring system according to the first embodiment.

  The display unit 108 acquires the reliability of classification of the detected object (step S301). Here, the reliability is the degree of reliability of the detected object classification. In this embodiment, the reliability increases as the similarity between the detected object and the model image increases. Next, the display unit 108 determines whether or not the acquired reliability is greater than a preset threshold value R (step S302). If the acquired reliability is greater than the threshold value R (step S302: Yes), the display unit 108 sets the alarm level of the detected object to alarm level: “3” and displays the alarm level: “3”. (Step S303). In the present embodiment, the alarm level increases as the degree of danger of the detected object increases.

  On the other hand, when the acquired reliability is equal to or less than the threshold value R (step S302: No), the display unit 108 indicates that the position of the detected object in the person position is a predetermined position such as a pocket region of clothes worn by the detection target person. It is determined whether or not (step S304). When it is determined that the position of the detected object is a predetermined position (step S304: Yes), the display unit 108 sets the alarm level of the detected object to alarm level: “1”, and sets the alarm level: “1”. It is displayed (step S305). On the other hand, when it is determined that the position of the detected object is other than the predetermined position (step S304: No), the display unit 108 sets the alarm level of the detected object to alarm level: “2”, and the alarm level: “ 2 "is displayed (step S306). That is, the display unit 108 changes the alarm level based on the position of the detected object (an example of the second position) in the person position.

  The alarm level when it is determined that the position of the detected object is other than the predetermined position is higher than the alarm level when it is determined that the position of the detected object is the predetermined position. This is because it can be assumed that there is a high possibility that the object is intentionally concealed. Further, the display unit 108 may display the alarm level in a subdivided manner based on at least one of the detected object classification and the detected object position (for example, a pocket area).

  Next, a display example of an alarm level by the monitoring system according to the present embodiment will be described with reference to FIG. FIG. 4 is a diagram illustrating a display example of an alarm level by the monitoring system according to the first embodiment.

  As illustrated in FIG. 4, the display unit 108 includes a visualized image 401, a pocket area image 402 representing a predetermined position (for example, a pocket area), a terahertz image 403, an object image 404 that is an image of a detected object, and a detected object. Judgment result 405 of whether or not the position is in the pocket area (when the position of the detected object is in the pocket: “◯”, when the position of the detected object is other than the pocket: “×”), detected A screen including the classification result 406 of the object, the reliability 407 of the classification result, and the alarm level 408 is displayed.

  In the present embodiment, the pocket area image 402 is an image obtained by superimposing a pocket image on the terahertz image subjected to image processing by the image processing unit 104. Further, the display unit 108 may change the display mode (for example, color, animation) of the character representing the alarm level 408 according to the alarm level 408. Further, the display unit 108 may output a sound for notifying the alarm level 408 according to the alarm level 408.

  As described above, according to the monitoring system according to the first embodiment, the detection target person is detected by detecting an object possessed by the detection target person from the person position specified based on the visualized image in the terahertz image. Even if the object is moving, the position of the person to be detected in the terahertz image can be specified with high accuracy. Therefore, the object possessed by the person to be detected can be detected from the terahertz image without stopping the person to be detected at the predetermined position.

(Second Embodiment)
The present embodiment is an example in which a detection target person is tracked between visualized images as frames constituting a moving image, and classification processing is executed on an object detected from the person position of the same detection target person being tracked. In the following description, description of the same configuration as that of the first embodiment is omitted.

  FIG. 5 is a diagram illustrating an example of a configuration of a monitoring system according to the second embodiment. As shown in FIG. 5, the monitoring system according to the present embodiment includes a terahertz image acquisition unit 101, a visualized image acquisition unit 102, a person detection unit 103, an image processing unit 104, an object detection unit 105, an object classification unit 106, and a detected object. In addition to the database 107 and the display unit 108, a person tracking unit 501 is provided.

  The person tracking unit 501 tracks a detection target person detected by the person detection unit 103 between frames constituting a visualized image of a moving image. Then, the object classification unit 106 performs a classification process on an object detected from the person position of the same detection target person being tracked. As a result, it is possible to prevent the classification process from being performed on an object detected from a person position of a different detection target person, so that the accuracy of classification of the detected object can be improved. At that time, the object classification unit 106 compares the objects detected from the person positions of the same detection target person being tracked, and removes the noise included in the object and integrates the characteristics included in the object. At least one of them may be executed.

  Further, the object detection unit 105 may track the same object from the person position between frames constituting the terahertz image of the moving image. In this case, the object classification unit 106 performs the classification process for each object tracked as the same object among the objects detected from the person positions of the same detection target person being tracked. As a result, it is possible to prevent different objects from being classified as the same object, so that the accuracy of classification of detected objects can be improved. Specifically, the object detection unit 105 has an object whose amount of movement of a rectangular area in which an object is detected between frames is equal to or smaller than a predetermined amount and whose similarity between rectangular areas between terahertz images is equal to or higher than a preset similarity. Are tracked as the same object.

  Next, with reference to FIG. 6, an example of a flow of classification processing of an object possessed by a person to be detected in the monitoring system according to the present embodiment will be described. FIG. 6 is a flowchart illustrating an example of a flow of classification processing of an object possessed by a person to be detected in the monitoring system according to the second embodiment.

  In the present embodiment, the person tracking unit 501 tracks a detection target person detected by the person detection unit 103 between frames constituting a visualized image of a moving image (step S601). Specifically, the person tracking unit 501 has a movement amount of a rectangular area in which a detection target person is detected between frames is equal to or less than a predetermined amount, and the similarity of the rectangular areas between frames is equal to or higher than a preset similarity. The detection target person is tracked as the same detection target person.

  The object classification unit 106 determines whether or not the tracking of the same person to be detected has been completed (step S602). When it is determined that tracking of the same person to be detected has not been completed (step S602: No), the process returns to step S201, and the terahertz image acquisition unit 101 and the visualized image acquisition unit 102 continue to acquire the terahertz image and the visualized image. To do.

  On the other hand, when it is determined that the tracking of the same detection target person has ended (step S602: Yes), the object classification unit 106 determines whether the object detection unit 105 has detected an object possessed by the detection target person. (Step S205). When an object possessed by the detection target person is detected (step S205: Yes), the object classification unit 106 performs a classification process on the object detected from the person position of the same detection target person (step S603). . As a result, it is possible to prevent the classification process from being performed on an object detected from a person position of a different detection target person, so that the accuracy of the classification process of the object detected from the terahertz image can be improved.

  At that time, the object classification unit 106 compares the objects detected from the person positions of the same person to be detected (for example, compares the size and shape of the image of the object), and compares the noise included in the object. After performing at least one of removal and integration of features included in the object, a classification process is performed on the object.

  At that time, the object classification unit 106 may compare objects tracked as the same object. As a result, it is possible to prevent the noise included in the object from being removed and the integration of the features included in the object from being executed based on different objects, so whether or not the object detected from the terahertz image is a predetermined object. The accuracy of the classification process for determining whether or not can be further improved.

  As described above, according to the monitoring system according to the second embodiment, the object detected from the person position of the same detection target person is detected from the person position of the different detection target person by executing the classification process. Since the classification process can be prevented from being performed on the object, the accuracy of the classification process of the object detected from the terahertz image can be improved.

(Third embodiment)
The present embodiment detects clothing worn by a person to be detected based on the visualized image, estimates the position of the pocket of the detected clothing, and the position of the detected object is the estimated pocket position in the person position. This is an example of raising the alarm level when the position is a suspicious position other than. In the following description, description of the same parts as those in the above-described embodiment will be omitted.

  FIG. 7 is a diagram illustrating an example of a configuration of a monitoring system according to the third embodiment. As shown in FIG. 7, the monitoring system according to the present embodiment includes a terahertz image acquisition unit 101, a visualized image acquisition unit 102, a person detection unit 103, an image processing unit 104, an object detection unit 105, an object classification unit 106, and a detected object. In addition to the database 107, the display unit 108, and the person tracking unit 501, a clothing classification unit 701 and a clothing database 702 are included.

  The clothing database 702 stores an image of clothing worn by a person (hereinafter referred to as a clothing model) and a pocket position of the clothing in association with each other. The clothing classification unit 701 (an example of a clothing detection unit) detects clothing worn by the detection target person based on the visualized image. In the present embodiment, the clothing classification unit 701 detects clothing from the detection target person every time tracking of the same detection target person ends. In addition, the clothing classification unit 701 estimates the position of the pocket of the detected clothing (an example of the third position). In the present embodiment, the clothing classification unit 701 estimates the position of the pocket associated with the clothing model of the detected clothing in the clothing database 702 as the position of the pocket of the detected clothing. The display unit 108 increases the alarm level when the detected position of the object is a position other than the pocket position. As a result, the alarm level can be increased when there is a high possibility that the detection target person intentionally hides the object, so that a more appropriate alarm level can be displayed according to the object possessed by the detection target person. it can.

  Next, an arrangement example of the terahertz image acquisition unit 101 and the visualized image acquisition unit 102 included in the monitoring system according to the present embodiment will be described with reference to FIGS. 8 and 9. FIG. 8 is a view of an example of the arrangement positions of the terahertz image acquisition unit and the visualized image acquisition unit of the monitoring system according to the third embodiment as viewed from above. FIG. 9 is a side view of an example of the arrangement positions of the terahertz image acquisition unit and the visualization image acquisition unit of the monitoring system according to the third embodiment.

  In the present embodiment, the monitoring system includes four terahertz image acquisition units 101 and four visualized image acquisition units 102, as shown in FIGS. Specifically, as illustrated in FIG. 8, the monitoring system includes two terahertz image acquisition units 101 and two visualization image acquisition units, one on each of the upstream side and the downstream side in the traveling direction of the detection target person passing through the predetermined position P. 102.

  As shown in FIGS. 8 and 9, the two upstream terahertz image acquisition units 101 are provided so as to be spaced apart in the horizontal direction across the predetermined position P, and acquire a terahertz image when the detection target person is viewed from the front. To be provided. As shown in FIGS. 8 and 9, the two terahertz image acquisition units 101 on the downstream side are provided to be spaced apart in the horizontal direction across the predetermined position P, and acquire a terahertz image when the detection target person is viewed from the back side. To be provided.

  As shown in FIGS. 8 and 9, the two upstream visualized image acquisition units 102 are provided so as to be spaced apart in the horizontal direction across the predetermined position P, and acquire a visualized image when the detection target person is viewed from the front. To be provided. As shown in FIG. 8 and FIG. 9, the two visualized image acquisition units 102 on the downstream side are provided horizontally apart from each other with the predetermined position P interposed therebetween, and visualized images obtained by viewing the detection target person from the back are acquired. To be provided.

  Further, the terahertz image acquisition unit 101 and the visualized image acquisition unit 102 are provided above the detection target person passing through the predetermined position P, and are included in the terahertz image and the visualization image by persons existing before and after the detection target person. The target person is installed at an angle where it is difficult to hide. This makes it possible to acquire a terahertz image and a visualized image that include an image of the whole body of the person to be detected.

  Next, an example of a flow of classification processing of an object possessed by a person to be detected in the monitoring system according to the present embodiment will be described with reference to FIG. FIG. 10 is a flowchart illustrating an example of a flow of classification processing of an object possessed by a detection target person in the monitoring system according to the third embodiment.

  In this embodiment, since the monitoring system has a plurality of terahertz image acquisition units 101, the processes shown in step S202 and step S204 are executed in parallel for each terahertz image acquired by each terahertz image acquisition unit 101. Alternatively, the monitoring system integrates the terahertz image obtained by integrating the features of the terahertz images acquired by the two upstream terahertz image acquisition units 101 or the features of the terahertz image acquired by the two downstream terahertz image acquisition units 101. The processing shown in step S202 and step S204 may be executed for the terahertz image that has been performed.

  In addition, since the monitoring system includes a plurality of visualized image acquisition units 102, the processes shown in step S203 and steps S601 to 602 are executed in parallel on each visualized image acquired by each visualized image acquisition unit 102. . Alternatively, the monitoring system integrates the visualized image obtained by integrating the features of the visualized images acquired by the two upstream visualized image acquiring units 102 or the features of the visualized image acquired by the two downstream visualized image acquiring units 102. For the visualized image, the processes shown in step S203 and steps S601 to 602 may be executed.

  In the present embodiment, when an object possessed by the detection target person is detected (step S205: Yes), the clothing classification unit 701 uses the clothing model stored in the clothing database 702, from the same detection target person. Cloth is detected (step S1001). Then, clothing classification unit 701 estimates the position of the pocket associated with the clothing model of the detected clothing in clothing database 702 as the position of the pocket of the detected clothing. Next, the display unit 108 displays an alarm level based on the detected position of the object and the estimated position of the pocket (step S1002).

  Next, an example of a flow of processing for displaying an alarm level in the monitoring system according to the present embodiment will be described with reference to FIG. FIG. 11 is a flowchart illustrating an example of a flow of processing for displaying an alarm level in the monitoring system according to the third embodiment.

  The display unit 108 acquires the reliability of classification of the detected object (step S301). In the present embodiment, the object classifying unit 106 classifies the detected object classification into not only a predetermined object classification but also an object classification other than a predetermined object such as a smartphone or a wallet. Next, the display unit 108 determines whether or not the acquired reliability is greater than the threshold value R (step S302). If the acquired reliability is larger than the threshold value R (step S302: Yes), the display unit 108 determines whether the detected object is a predetermined object (step S1101).

  When the detected object is a predetermined object (step S1101: Yes), the display unit 108 sets the alarm level of the detected object to alarm level: “3” and displays the alarm level: “3” (step S303). ). On the other hand, when the detected object is an object other than the predetermined object (step S1101: No), the display unit 108 does not display the alarm level (step S1102).

  If the acquired reliability is equal to or less than the threshold value R (step S302: No), the display unit 108 determines whether the position of the detected object in the person position is within a suspicious position other than the estimated pocket position. Judgment is made (step S1103). When the position of the detected object is outside the suspicious position (step S1103: No), the display unit 108 sets the alarm level of the detected object to alarm level: “1” and displays the alarm level: “1”. (Step S1104). On the other hand, when the position of the detected object is within the suspicious position (step S1103: Yes), the display unit 108 sets the alarm level of the detected object to alarm level: “2”, and sets the alarm level: “2”. It is displayed (step S1105). As a result, the alarm level can be increased when there is a high possibility that the detection target person intentionally hides the object, so that a more appropriate alarm level can be displayed according to the object possessed by the detection target person. it can. Further, since it is not necessary for the manager who manages the entry / exit of the person to be detected to determine the alarm level, it is possible to save the security resources.

  Next, display examples of alarm levels by the monitoring system according to the present embodiment will be described with reference to FIGS. 12 and 13 are diagrams illustrating examples of alarm level display by the monitoring system according to the third embodiment.

  As illustrated in FIG. 12, the display unit 108 includes a visualization image 401 acquired by the upstream visualization image acquisition unit 102, a pocket region image 402, a terahertz image 403 acquired by the upstream terahertz image acquisition unit 101, and a terahertz A screen including a superimposed image 1201 in which a pocket region image 402 is superimposed on an image 403, an object image 404, a determination result 405, a classification result 406, a reliability 407, and an alarm level 408 is displayed. Further, the display unit 108 applies the visualization image 401, the pocket region image 402 acquired by the downstream visualization image acquisition unit 102, the terahertz image 403 acquired by the downstream terahertz image acquisition unit 101, and the terahertz image 403. A screen including a superimposed image 1201 on which the pocket area image 402 is superimposed, an object image 404, a determination result 405, a classification result 406, a reliability 407, and an alarm level 408 is displayed.

  Alternatively, as illustrated in FIG. 13, the display unit 108 displays a suspicious position image 1301 representing a suspicious position instead of the pocket area image 402, and a superimposed image in which the pocket area image 402 is superimposed on the terahertz image 403. Instead of 1201, a superimposed image 1302 in which a suspicious position image 1301 is superimposed on the terahertz image 403 may be displayed.

  In the present embodiment, the display unit 108 estimates the position of the pocket of the detected clothing using the clothing database 702, but is not limited thereto, and is based on at least one of the terahertz image and the visualized image. Thus, the position of the pocket of the detected clothing may be estimated.

  Thus, according to the monitoring system according to the third embodiment, the alarm level can be increased when the detection target person has a high possibility of intentionally hiding the object, so the detection target person possesses it. A more appropriate alarm level can be displayed according to the object.

  Further, the monitoring system according to the present embodiment is installed at a plurality of points (predetermined positions), and a detection target person whose alarm level is a predetermined alarm level (for example, alarm level: “2”) or higher is stored in the storage unit. You may save it. Accordingly, it is possible to track a person requiring attention by comparing the detection target person detected by the monitoring system at each point with the detection target person stored in the storage unit. In addition, when the detection target person detected by the monitoring system at each point is successfully matched with the detection target person stored in the storage unit, processing to increase the threshold R and the alarm level is performed, and the person who needs attention is possessed. The object classification process may be different from the object classification process carried by a general person to be detected.

  As described above, according to the first to third embodiments, an object held by a detection target person can be detected from a terahertz image without stopping the detection target person at a predetermined position.

  The program executed by the monitoring system of this embodiment is provided by being incorporated in advance in a ROM (Read Only Memory) or the like. A program executed in the monitoring system of the present embodiment is an installable or executable file and is read by a computer such as a CD-ROM, a flexible disk (FD), a CD-R, or a DVD (Digital Versatile Disk). You may comprise so that it may record on a possible recording medium and provide.

  Furthermore, the program executed by the monitoring system of the present embodiment may be provided by being stored on a computer connected to a network such as the Internet and downloaded via the network. Further, the program executed in the monitoring system of the present embodiment may be provided or distributed via a network such as the Internet.

  The program executed by the monitoring system of the present embodiment includes the above-described units (person detection unit 103, image processing unit 104, object detection unit 105, object classification unit 106, person tracking unit 501, and clothing classification unit 701). The module has a module configuration, and as the actual hardware, a CPU (Central Processing Unit) reads the program from the ROM and executes the program, so that each unit is loaded on the main storage device, and the person detection unit 103, the image processing unit 104, the object detection unit 105, the object classification unit 106, the person tracking unit 501, and the clothing classification unit 701 are generated on the main storage device.

  Although several embodiments of the present invention have been described, these embodiments are presented by way of example and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the scope of the invention. These embodiments and modifications thereof are included in the scope and gist of the invention, and are included in the invention described in the claims and the equivalents thereof.

DESCRIPTION OF SYMBOLS 101 Terahertz image acquisition part 102 Visualization image acquisition part 103 Person detection part 104 Image processing part 105 Object detection part 106 Object classification part 107 Detected object database 108 Display part 501 Person tracking part 701 Clothing classification part 702 Clothing database

Claims (5)

A first image acquisition unit that receives a terahertz wave emitted from a person passing through a predetermined position and acquires a first image obtained by imaging the predetermined position based on a reception result of the terahertz wave;
A second image acquisition unit that receives the light reflected at the predetermined position and acquires a second image obtained by imaging the predetermined position based on a light reception result of the light;
A person detecting unit for detecting the person from the second image;
An object detection unit for detecting an object possessed by the person from a first position corresponding to the detected person in the first image;
A classification unit that executes a classification process for determining whether the detected object is a predetermined object;
A display unit for displaying an execution result of the classification process and an alarm level of the detected object based on the execution result;
A monitoring system comprising:   The monitoring system according to claim 1, wherein the display unit further changes the alarm level based on a second position of the object at the first position. A person tracking unit that tracks the person between frames constituting the second image of the moving image;
The monitoring system according to claim 1, wherein the classification unit performs the classification process on the object detected from the first position of the same person being tracked. The object detection unit further tracks the same object between frames constituting the first image of the moving image,
The monitoring system according to claim 3, wherein the classification unit executes the classification process for each of the objects tracked as the same object being tracked. A clothing detector for detecting clothing worn by the person based on the second image and estimating a third position of the pocket of the detected clothing;
The monitoring system according to claim 2, wherein the display unit increases the alarm level when the second position is a position other than the position corresponding to the third position in the first position.

Images