JP2012048676A - Notification apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a notification apparatus capable of determining whether a person is suspicious, without increasing a processing load, even in the case where information on the person is lacked in image information when the person is present at a position proximate to an important object to be monitored.SOLUTION: The present invention relates to a notification apparatus for detecting and notifying a suspicious person in a monitoring space where a storage cabinet is installed, including: a person detection section for detecting the presence of a person at a position proximate to the storage cabinet; an image capturing section for imaging the monitoring space periodically and capturing a monitor image; a storage section for storing the monitor image captured by the image capturing section; a processing image determination section for determining a monitor image which is captured before the lapse of a first predetermined time from the person detection by the person detection section or after, as a processing image; and a suspicious person detection section for detecting a suspicious person of which the face is not sufficiently photographed in the processing image.

Description

  The present invention relates to a reporting device that detects and reports the presence of a suspicious person in a space where important monitoring objects are installed, and in particular, lacks human information in image information when a person is present in the proximity of an important monitoring object. It is an object of the present invention to provide a reporting device that can determine whether or not the person is a suspicious person.

2. Description of the Related Art Conventionally, a reporting device is known that automatically extracts feature information of a person region from an image input from an imaging device and automatically outputs a report to the outside when the feature information matches a feature of a suspicious person or the like.
In the notification device using such an image, the captured image is used as verification information when an abnormal situation occurs, and therefore, a range wider than the range to be monitored is captured by the imaging device. There is. Moreover, while the process of detecting a suspicious person from an image has become complex and sophisticated, the monitoring device itself that is to be monitored is required to be small and have low power consumption. There is also a risk that processing will not be completed within the period interval and processing may be delayed.

  For this reason, in order to prevent erroneous detection of a non-malicious third party that appears in the image as a suspicious person, and to reduce the processing load, the captured image is always processed to detect suspicious persons, etc. In general, the detection processing of a suspicious person is performed on a captured image only when the presence of a person is detected in a range to be monitored instead of extracting.

  For example, in Patent Document 1, in order to detect a suspicious person who may withdraw cash from an automatic transaction apparatus illegally, the camera detects when a human sensor detects that a person stands in front of the automatic transaction apparatus. There is described a monitoring device that captures an image and determines whether or not a suspicious person whose face is concealed with sunglasses or a mask is captured in the image.

JP 2010-79750 A

  In the monitoring apparatus of Patent Document 1, an image is displayed when it is detected that a person stands in front of an automatic transaction apparatus that becomes a monitoring range by using a camera in which a face of an operator of the automatic transaction apparatus can be photographed. And detecting the suspicious person based on the information of the person's face, thereby preventing erroneous detection and an increase in processing load.

  However, the image when the presence of a person is detected in the monitoring range is not necessarily suitable as an image for detecting a suspicious person. In such a case, the monitoring apparatus of Patent Document 1 detects a suspicious person. It may not be possible.

  For example, if a person who wants to open a vault such as a safe is to be monitored, it is difficult to incorporate an imaging device because the safe requires high fireproof performance and robustness. The person to be opened is imaged. The safe is generally fixedly installed on the floor, and the front door is provided with unlocking means. For this reason, a person who wants to open the safe door usually bends and operates the unlocking means. Therefore, in such a case, when an image of a person who wants to open the safe door is picked up by an image pickup device installed outside the safe, the person is behind the safe and picks up an image including the characteristic information of the person. It becomes difficult.

That is, unlike the automatic transaction apparatus that is operated while standing as exemplified in Patent Document 1, when a person who wants to open a safe is targeted for monitoring, the presence of a person in front of the safe is detected. There is a possibility that information for determining whether or not the person is a suspicious person is not always included in the image at that time, and the detection accuracy of the suspicious person may be reduced.
On the other hand, if the presence of a person is detected at a position away from the safe, a third party who does not attempt to open the safe may be erroneously detected as a suspicious person. Decreases.

  The present invention has been made in view of the above problems, and relates to a notification device that performs detection processing of a suspicious person only when the presence of a person is detected at a close position of a storage such as a safe, and the detection of a suspicious person The purpose is to provide a reporting device that can improve accuracy.

  In order to achieve the above object, the notification device according to the present invention is a notification device that detects and reports a suspicious person in a monitoring space where a storage is installed, and detects the presence of a person in the proximity of the storage. A person detection unit that periodically captures the monitoring space to acquire a monitoring image, a storage unit that stores a monitoring image acquired by the imaging unit, and a time when the person is detected by the person detection unit A processing image determination unit that determines, as a processing image, a monitoring image acquired after the first predetermined time; and a suspicious person detection unit that detects a suspicious person whose face is not sufficiently photographed with respect to the processing image. It was characterized by that.

  In such a configuration, the reporting device periodically acquires monitoring images and stores them in the storage unit, and detects the presence of a person at a close position of the storage, after the first predetermined time from the time of the person detection. The acquired monitoring image is determined as a processed image, and acts to detect a suspicious person whose face is not sufficiently captured for the processed image.

  According to such a configuration, when a person is detected at a close position of the storage, such as a position where the storage can be opened, it is possible to detect a suspicious person retroactively immediately before the person approaches the storage. It becomes. As a result, even if the person's feature information cannot be captured at the close position of the storage, the feature information is extracted from the image until the person approaches the storage, and it is determined whether or not the person is a suspicious person. It is possible to make a determination, and it is possible to improve detection accuracy while preventing an increase in processing load.

  Further, in the notification device of the present invention, the processed image determination unit is acquired after the first predetermined time from the time of the person detection and before the second predetermined time shorter than the first predetermined time. A monitoring image may be determined as a processed image.

  In such a configuration, the notification device starts from the time when the presence of a person is detected at the close position of the storage until the second predetermined time before the first predetermined time (first predetermined time> second predetermined time). ) Acts to detect a suspicious person whose face is not sufficiently captured in the monitoring image acquired during

That is, characteristic information of the person is captured in images from the time when the presence of a person is detected at a close position of the storage, such as a position where the storage can be opened, to the second predetermined time before the near past. There is a possibility that the image processing load is unnecessarily increased. If such an image is set as a processing target for suspicious person detection, the image processing load increases unnecessarily, and other processing may be delayed.
Therefore, in such a configuration, when detecting a suspicious person for a past image from the time when the presence of a person is detected at a close position of the storage, the processing load is reduced by excluding the near past image from the processing target. It is possible to improve the detection accuracy while preventing the increase.

  Further, in the notification device of the present invention, the suspicious person detection unit is updated when the process for detecting the suspicious person is completed when the process for detecting the suspicious person is completed for the processed image determined by the processed image determination unit. Processing for detecting the suspicious person may be performed on the monitoring image.

  In such a configuration, when the reporting device detects the presence of a person at a close position of the storage, the reporting device performs a suspicious person detection process on a past image acquired before that time, and detects a suspicious person on the past image. When the process is completed, a suspicious person detection process is performed on the latest image acquired at present, and thereafter, a suspicious person detection process is performed on the latest image when the process is completed.

The processing for detecting a suspicious person from an image differs depending on factors such as the number of images to be processed (amount), the number of persons included in the image, and the degree of complexity (complexity). There is also a possibility that the process for is not completed within the image acquisition interval. For this reason, if all the images acquired periodically are processed, unprocessed images may be accumulated additionally, and the processing may be delayed.
Therefore, when the process of detecting a suspicious person for a past image that is started when the presence of a person is detected at the close position of the storage is completed, it is acquired when the previous process is completed thereafter. By repeating the process of detecting a suspicious person only for the latest image, it is possible to improve detection accuracy while preventing an increase in processing load.

  Further, in the notification device of the present invention, the suspicious person detection unit performs a process of detecting the suspicious person until a third predetermined time elapses after the person detection unit no longer detects the presence of the person. It may be.

  According to such a configuration, even if the presence of the suspicious person cannot be determined from the image until the person approaches the storage, the third person can no longer be detected in the proximity of the storage. It is possible to detect whether or not there is a suspicious person for an image acquired until a predetermined time elapses, that is, while the person is away from the storage, and the detection accuracy is prevented while increasing the processing load. Can be improved.

  According to the present invention, when a person is detected at a close position of the storage, it is possible to perform a suspicious person detection process retroactively before the person approaches the storage. As a result, even if the feature information of the person cannot be extracted at the close position of the storage, the feature information is extracted from the image until the person approaches the storage, and it is determined whether or not the person is a suspicious person. It is possible to make a determination, and it is possible to improve detection accuracy while preventing an increase in processing load.

It is the schematic which shows the whole structure of the notification system by the notification apparatus of this invention. It is a block diagram which shows the structure of the notification apparatus of this invention. It is a timing chart which shows the operation | movement outline | summary of the information apparatus of this invention. It is a flowchart which shows the suspicious person monitoring operation | movement process of the notification apparatus of this invention.

Embodiments of the present invention will be specifically described below with reference to the drawings.
In this embodiment, the case where an office area such as a financial institution or a store is monitored as a monitoring area (monitoring space) is exemplified, but the scope of the present invention is not limited to this.

FIG. 1 is a configuration diagram showing a notification system 1 using the notification device of the present invention.
FIG. 1 schematically shows a monitoring area in which a storage 2 having a storage space inside a safe or the like and a notification device 3 are installed on a side view. As shown in FIG. 1, the notification system 1 of the present embodiment includes a communication line between a notification device 3 placed on the upper part of a storage 2 fixedly installed on the floor of a monitoring area and a remote monitoring center 5. They are connected via the network 6.

In the case of the reporting device 3, a monitoring camera 31 as an imaging unit and a human sensor 32 as a person detection unit are arranged in the same direction as the storage door 21. The surveillance camera 31 is arranged so that a person approaching the storage 2 with the front of the storage door 21 as an imaging field of view can be imaged, and the human sensor 32 is a proximity position where the storage door 21 can be opened (for example, The human body 7 existing within 80 cm in front of the storage door is disposed so as to be detectable.
Note that the monitoring camera 31 and the human sensor 32 are not limited to being incorporated in the housing of the notification device 3, and can capture images of at least a person approaching the storage 2, and are close to the storage door 21. A monitoring camera 31 and a human sensor 32 may be separately installed in the upper space or rear wall surface of the storage 2 so that the human body 7 can be detected and connected to the notification device 3. In addition, the reporting device 3 is not limited to the example of being placed on the upper part of the storage 2 and may be set in the upper space of the storage 2 or the rear wall surface.

The monitoring camera 31 captures a monitoring area at every predetermined shooting interval (for example, 0.2 seconds), generates image data as a monitoring image, and outputs it. Further, the human sensor 32 outputs whether or not a human body is detected within the detection range so as to be identifiable.
Based on the information output from the monitoring camera 31 and the human sensor 32, the notification device 3 refers to the image data when the human sensor 32 is in a detection state, and a suspicious person is located around the storage 2. Determine if it exists.

  In this embodiment, a suspicious person refers to a person who may impair the maintenance of a monitored area such as a burglar, and specifically, a person whose appearance matches a predetermined feature pattern. In general, a suspicious person who performs a robbery or the like tends to hide his / her face in order to prevent recognition of the identity. Therefore, in the present embodiment, as the person's feature information in the image data, for example, the human face is hidden by partially covering all or part of the face by wearing sunglasses, a mask, a full face helmet, etc. A person, that is, a person whose face is not sufficiently photographed is determined as a suspicious person.

Here, an action mode when a suspicious person such as a robber enters the surveillance area and tries to take away the contents stored in the storage 2 will be described. As an action mode in which a suspicious person takes away the contents stored in the storage 2, it can be roughly classified into the following two types.
First, the suspicious person has the unlocking method of the storage 2. If the suspicious person already has the unlocking method of the vault 2 such as a key or a password, the suspicious person restrains the user in the surveillance area and cannot move, and then approaches the vault 2 Unlock and take away the stored items.
Second, the suspicious person does not have the unlocking method of the storage 2. When the suspicious person does not have the unlocking method of the storage 2, the suspicious person threatens the user in the monitoring area with a weapon such as a blade and causes the user to unlock the storage 2. In this case, while the user unlocks the storage 2, the suspicious person monitors the user behind it.

In general, since a person who intends to unlock the storage 2 approaches the storage 2 and bends and unlocks the storage door 21, an image acquired when the person approaches the storage 2 There is a possibility that the person is behind the storage 2 and does not include the feature information of the person.
For this reason, in the first type described above, there is a possibility that a suspicious person cannot be detected from the image data acquired when a person approaches the storage 2 and an emergency situation cannot be detected. In the image data acquired when a person approaches the storage 2, a part of the user's body close to the image data is greatly reflected, and a suspicious person existing in the vicinity cannot be detected. Like the type of, there is a possibility that an emergency cannot be detected.

  Therefore, as a characteristic matter in the present embodiment, when the human sensor 32 detects the human body 7 at a position close to the storage door 21, the notification device 3 uses image data acquired in the past from the present time. It is used to determine whether or not there is a suspicious person around the storage 2. In any of the first and second types described above, the image data before the person reaches the close position of the storage door 21 has a high possibility that the human body when approaching the person is imaged. By using such image data, the detection accuracy of the suspicious person is improved. When it is determined that there is a suspicious person, the reporting device 3 transmits an emergency signal indicating that an emergency has occurred in the monitoring area to the remote monitoring center 5 and detects a suspicious person. The used image data is transmitted to the monitoring center 5.

  The monitoring center 5 is a facility provided with a center device 51 operated by a security company or the like. The center device 51 is composed of one or a plurality of computers, and realizes a monitoring center function related to the present invention. In the monitoring center 5, various devices are controlled by the center device 51, the emergency signal received from the reporting device 3 is recorded, and the image data transmitted from the reporting device 3 is displayed on the display 52, so that the monitor can monitor. Monitors multiple target surveillance areas.

<Notification device>
Next, the configuration of the notification device 3 will be described with reference to FIG. FIG. 2 is a block diagram illustrating a configuration of the notification device 3.
The reporting device 3 is connected to the communication network 6 so as to be communicable with the monitoring center 5, and is fixedly placed on the upper part of the storage 2.

  The reporting device 3 includes a monitoring camera 31 that captures the front of the storage 2, a human sensor 32 that detects a human body located in the vicinity of the storage door 21, a communication unit 33 connected to the communication network 6, The storage unit 34 configured by an HDD, a memory, and the like, and a control unit 35 configured by an MPU, a microcomputer, and the like for controlling each unit are schematically configured.

  The monitoring camera 31 generates a monitoring image of a monitoring area including the front of the storage 2. The monitoring camera 31 converts an image signal input from the image sensor into a digital signal at a predetermined shooting interval (for example, a cycle of 0.2 seconds, that is, 5 fps (frame / second)), performs compression encoding processing, and performs a predetermined standard. Image data compliant with (for example, JPEG standard) is generated. The generated image data is output to the control unit 35.

The human sensor 32 detects a human body existing in the proximity position of the storage 2 and, in the present embodiment, a detection range of a proximity position (for example, within 80 cm in front of the storage door) where the storage door 21 can be opened. As described above, it is detected whether or not a human body exists in the detection range. The proximity position at which the storage door 21 can be opened may be set as a distance from a general person's arm to the head in front of the storage door 21. The detection range set in front of the storage 21 may be arbitrarily set according to the environment so that a predetermined depression angle is set toward the human sensor 32 horizontally or toward the floor. In the present embodiment, the human sensor 32 includes a laser light projecting / receiving distance measuring sensor (PSD), and determines that a human body exists in the detection range when the projected search signal is reflected within the detection range. . The human sensor 32 outputs the detection state of the human body within the detection range as contact information, and outputs a detection signal when the human body is detected and an undetected signal when the human body is not detected. To do.
The human sensor is not limited to this, and may be realized by a known mat sensor, a heat ray sensor that detects a heat ray of a human body, or other proximity sensors.

  The communication unit 33 is connected to the center device 51 via the communication line network 6 and communicates with the monitoring center 5. When the control unit 35 determines that there is a suspicious person in the monitoring area, the communication unit 33 transmits the emergency signal including its own address information and the image data stored in the storage unit 34 to the monitoring center 5.

  The storage unit 34 is configured by a ROM, a RAM, or an HDD, stores address information for identifying itself, various programs, and the like, and further stores various information for operating the reporting device 3. Specifically, the storage unit 34 determines whether or not a suspicious person has been detected from the accumulated image information for storing the image data input from the monitoring camera 31, the processed image information determined by the control unit 35, and the image data. And a person flag indicating whether or not a person is staying near the storage 2 is stored.

  The stored image information functions as a ring buffer for temporarily recording image data input from the monitoring camera 31. The stored image information holds image data input from the monitoring camera 31 for a predetermined number of frames (for example, 100 frames) together with information on the imaging time. In the stored image information, the oldest of the stored image data is overwritten with newly received image data, and the latest image data is always stored for a predetermined number of frames. Since the number of frames held in the accumulated image information is the number of images immediately before the human sensor 32 detects the human body, the number of frames is traced back to the past when the control unit 35 performs suspicious person detection processing. Set accordingly.

The processed image information stores the image data determined as processing target image data for suspicious person detection by the control unit 35 among the image data stored in the accumulated image information. As described above, since the accumulated image information is information that can be overwritten sequentially updated with newly acquired image data, the processed image information is temporarily stored so that the image data determined as the processing target image data is not overwritten. Is holding.
In the present embodiment, the accumulated image information and the processed image information are described as separate storage information. However, the processed image information is flag information that prohibits overwriting of image data determined as a processing target image in the accumulated image information. May be stored.

  The control unit 35 includes a microcomputer including a CPU, a ROM, a RAM, and the like and peripheral circuits thereof, and controls each unit described above. For this purpose, the control unit 35 is a processing module that performs suspicious person detection from image data input from the monitoring camera 31 and held as a functional module realized by the microcomputer and a computer program executed on the microcomputer. A processing image determination unit 351 that determines image data, a suspicious person detection unit 352 that detects a suspicious person from image data, and a stay determination unit 353 that determines whether or not a person is staying in the proximity of the storage 2 A notification control unit 354 that controls notification from the communication unit 33 to the monitoring center 5.

  The processed image determination unit 351 determines image data to be subjected to suspicious person detection processing from among the image data held in the accumulated image information when the human sensor 32 is in a detection state and a detection signal is input. Specifically, when a detection signal is input, the processed image determination unit 351 performs suspicious person detection processing on image data acquired from a predetermined time T1 to a predetermined time T2 before the current time (T1> T2). It determines as the process target image data (process image) to perform, and memorize | stores these process target image data in the process image information of the memory | storage part 34. FIG. For example, when the detection signal is input, the processed image determination unit 351 receives the image data acquired from 1.8 seconds before to 1 second before the current time (9 frames if captured every 0.2 seconds). 5 frames from the previous 5 frames) are determined as processing target image data.

The processing of the processed image determination unit 351 means that image data until the person 7 detected by the human sensor 32 approaches the storage 2 is targeted for suspicious person detection processing. As described above, in the image data acquired when a person approaches the storage 2, there is a possibility that the person is behind the storage 2 and does not include characteristic information of the person. On the other hand, as illustrated in FIG. 1, the image data of a predetermined time before the person 7 is detected at the close position of the storage 2 walks in front of the storage as a past image of the person 7. There is a high possibility that the incoming person 7a has been imaged. In the image data when the person 7a approaches by walking in front of the storage, there is a high possibility that the feature information including the face portion of the person 7a is captured. For this reason, the processing image determination unit 351 excludes image data immediately before the person approaches the position where the storage door 21 can be unlocked (from the current time to a time before the predetermined time T2), The image data when approaching the storage is determined as a target for suspicious person detection processing.
Not limited to this, the processed image determination unit 351 may determine all the image data acquired from the predetermined time T1 to the present time as the processing target image data, and processes the image data immediately before the person approaches. Whether or not to exclude the target may be arbitrarily set according to the processing load of the suspicious person detection process.

  The suspicious person detection unit 352 is activated when the human sensor 32 enters a detection state, sequentially reads out the processing target image data stored in the processing image information, and whether or not there is a person having the image characteristics of the suspicious person. Determine whether. When the suspicious person detecting process is completed for all the processing target image data stored in the processed image information, the suspicious person detecting unit 352 continues to store the latest image data (current image data currently stored in the accumulated image information). ) As a processing target, and it is determined whether or not there is a person having image characteristics of a suspicious person. Thereafter, the suspicious person detection unit 352 ends the suspicious person detection process for the image data to be processed until a predetermined delay time (for example, 5 seconds) elapses after the human sensor 32 no longer detects a person. Repeating the suspicious person detection process with the newest current image data at that time as the processing target. When a predetermined delay time (for example, 5 seconds) elapses after the human sensor 32 no longer detects a person, the suspicious person detection unit 352 ends the operation.

Performing suspicious person detection processing on the current image data until the predetermined delay time has elapsed after the human sensor 32 is in the undetected state means that the person who has taken out the stored items in front of the storage 2 This means that the suspicious person is also detected for the image data acquired when leaving. This makes it possible to determine whether or not there is a suspicious person from the current image data at the time of leaving, even if the suspicious person cannot be determined from the processing target image data that becomes an image until approaching the storage 2. .
In addition, by performing suspicious person detection processing only for the latest current image data at the time when the previous processing is completed, even if a large number of people are reflected in the image data, the processing time becomes long. The load is reduced by preventing unprocessed image data from accumulating due to image data sequentially acquired during one frame processing.

  Next, suspicious person detection processing by the suspicious person detection unit 352 will be described. In this embodiment, as image characteristics of a suspicious person, each part (eyes, nose, mouth) that is a facial feature part whose human phase can be identified is hidden in the image data (each part of the face cannot be detected) Is detected and the presence of a suspicious person is determined. If it is determined that there is a suspicious person, the suspicious person detection unit 352 turns on the suspicious person flag in the storage unit 34.

  The suspicious person detection unit 352 can use various known methods for detecting a facial feature in the face area. Specifically, as a suspicious person detection process, the suspicious person detection unit 352 compares reference data (image data when there is no person) captured in the past with image data to be processed, and a variation in luminance value exceeds a predetermined value. In addition to extracting the variation pixels that are, the group of substantially continuous variation pixels is grouped as one variation region, and each variation region is labeled with a unique label. Then, the edge component of the labeled fluctuation region is extracted, and an elliptical edge distribution approximate to the contour shape of the face is detected in the edge image data, and the color information in the elliptical region surrounded by the edge distribution is When the skin color (for example, a pixel having a hue (H) component from 0 to 30 in the HSV color system) is included, the elliptical area is extracted as a face area. Then, it is determined whether or not an edge (for example, a horizontal edge such as an eye or a mouth) that becomes a facial feature portion appears in the facial region, and a facial region where such a facial feature portion cannot be detected by a suspicious person. Judge that there is.

In addition, the suspicious person detection unit 352 detects a face area in which sunglasses and a mask are worn from the luminance distribution of the face area. In the face area where the sunglasses and the mask are worn, dark pixels concentrate on the upper side, so that the luminance center of gravity in the vertical direction is lower than the center. Further, the luminance value for each line viewed in the horizontal direction is small, whereas the variance of the luminance value for each line viewed in the vertical direction is large. Furthermore, there is a property that the luminance value is extremely low depending on the part of the sunglasses. Therefore, the Y coordinate of the luminance center of gravity of the face area is below the center of the face area, and the ratio between the standard deviation in the X direction and the standard deviation in the Y direction is larger than a predetermined threshold, and the pixel having a low luminance value Is equal to or greater than the threshold value, it is determined that the image in the face area is an image of a face wearing sunglasses and a mask, and the face area is determined to be a suspicious person.
The suspicious person detection process is not limited to this, and it is only necessary to determine that the appearance or behavior of a person matches a preset suspicious pattern.

Based on the output of the human sensor 32, the stay determination unit 353 determines whether or not a person is continuously present (stay) at a close position where the storage door 21 can be opened. Since it is difficult to complete the work of unlocking the storage door 21 and taking out the stored items in an instant, it usually requires work for a certain time in front of the storage 2. Therefore, when the stay determination unit 353 receives a detection signal from the human sensor 21, the stay determination unit 353 starts counting time and continuously detects a person's presence for a predetermined time (for example, 2 seconds), that is, from the input of the detection signal. If the undetected signal is not input even after a lapse of time, it is determined that a person is continuously present in the proximity position where the storage door 21 can be opened, and the person flag in the storage unit 34 is turned ON. Further, when the stay determination unit 353 receives a non-detection signal from the human sensor 21, it determines that there is no person and turns off the person flag in the storage unit 34.
Here, the person flag is flag information indicating whether or not a person is staying in the proximity position of the storage 2. When the reporting device 3 is activated, the flag is OFF as an initial state, and is turned on when the stay determination unit 353 determines that a person is staying in the vicinity of the storage door 21, and is turned OFF when it is determined that the person is not staying.

The notification control unit 354 controls notification processing to the remote monitoring center 5. The notification control unit 354 determines whether or not the suspicious person detection unit 352 determines the presence of the suspicious person between the time when the human sensor 21 enters the detection state and the time when the suspicious person detection unit 352 starts operation and ends. When the two conditions that the part 353 determines that a person is staying at the close position of the storage 2 are satisfied, an emergency signal including its own address information is generated, and the suspicious person detecting part 352 has the presence of the suspicious person. Is output to the communication unit 33, and transmitted from the communication unit 33 to the monitoring center 5.
Not limited to this, the notification control unit 354 may be configured to perform notification processing only on the condition that the suspicious person detection unit 352 determines the presence of the suspicious person. Whether or not to be set may be arbitrarily set according to the environment of the monitoring area.

<Description of operation>
About the notification system 1 comprised as mentioned above, the operation | movement is demonstrated with reference to drawings. Here, the operation related to the reporting device 3 will be mainly described. First, the outline | summary of operation | movement of the notification apparatus 3 is demonstrated using FIG. FIG. 3 is a timing chart schematically showing the operation of the reporting device 3.

  As shown in FIG. 3A, image data is acquired by the monitoring camera 31 at predetermined intervals (for example, 0.2 seconds) and stored in the accumulated image information in the storage unit 34. Here, as shown in FIG. 3B, when the human sensor 32 detects the human body 7 at a close position where the storage door 21 can be opened (t in the figure), a detection signal is input to the control unit 35, As shown in FIG. 3C, the processed image determination unit 351 suspiciously acquires image data acquired from a current time before a predetermined time T1 to a predetermined time T2 (for example, from 1.8 seconds to 1 second before). Is determined as processing target image data to be subjected to person detection processing, and stored in the processing image information of the storage unit 34.

  When the human sensor 32 detects a human body, the suspicious person detection unit 352 is activated as shown in FIG. 3D, and the processing target image stored in the processing image information as shown in FIG. Data is read out in order from the oldest, and the suspicious person detection process is performed on the sequentially read out image data as shown in FIG. When the suspicious person detection process is completed for all the processing target image data stored in the processed image information (tb in the figure), as shown in FIG. 3E, the latest suspicious person detection process is completed when the previous suspicious person detection process is completed. Image data (current image data) is read from the stored image information, and the suspicious person detection process is performed on the read image data. Thereafter, when the suspicious person detection process ends, the suspicious person detection unit 352 repeatedly reads out the latest image data from the accumulated image information and performs the suspicious person detection process. In any of the series of suspicious person detection processes, when the presence of a suspicious person is detected, the suspicious person detecting unit 352 turns on the suspicious person flag in the storage unit 34. In addition, when the human sensor 32 is in an undetected state of the human body (tc in the figure), an undetected signal is input to the control unit 35, and the suspicious person detecting unit 352 starts measuring a delay time (for example, 5 seconds). When the delay time elapses, the operation ends (td in the figure). At this time, the suspicious person flag in the storage unit 34 is initialized to OFF.

As shown in FIG. 3G, when the detection state of the human sensor 32 continues for a predetermined time (for example, 2 seconds), the person continues to a proximity position where the stay determination unit 353 can open the storage door 21. It is determined that it exists (ta in the figure), and the person flag is turned ON. When the human sensor 32 is not detected, the stay determination unit 353 determines that there is no person (tc in the figure), and turns off the person flag.
The notification control unit 354 starts the operation of the suspicious person detection unit 352 (t in the figure) and ends (td in the figure) until the suspicious person flag is displayed in the suspicious person detection process shown in FIG. If the condition that the flag is ON and the condition that the retention determination unit 353 shown in FIG. 3G satisfies the condition that the person flag is set to ON are satisfied, whichever of the later conditions is satisfied is monitored. An emergency signal is transmitted to the center 5.

Next, the suspicious person monitoring operation process by the processing image determination unit 351 and the suspicious person detection unit 352 shown in FIGS. 3C to 3F will be described with reference to FIG. FIG. 4 is a flowchart showing the operation of the monitoring program repeatedly executed by the control unit 35 of the reporting device 3.
The notification device 3 accepts input of image data captured by the monitoring camera 31 at a predetermined cycle (for example, a cycle of 0.2 seconds, that is, 5 fps (frame / second)) and stores it in the accumulated image information of the storage unit 34. When the human sensor 32 detects the human body 7 in proximity to the storage door 21 and a detection signal is input (step ST1-Yes), the processed image determination unit 351 is as shown in FIG. Then, image data acquired in the past from the present time among the image data stored in the accumulated image information is determined as processing target image data, and stored in the processed image information in the storage unit 34 (step ST2).

  If the undetected signal indicating that the human sensor 32 no longer detects the human body is not input to the suspicious person detection unit 352 (step ST3-No), the process of the storage unit 34 is performed as illustrated in FIG. The processing target image data is read from the image information (step 4), and a suspicious person detection process is executed for the image data (step ST5). In the suspicious person detection process, as the image characteristics of the suspicious person, each part (eyes, nose, mouth) that is a facial feature part that can distinguish the human phase is concealed in the read image data (each part of the face is hidden). A person who cannot be detected is detected as a suspicious person. When a suspicious person is detected from the image data to be processed in the suspicious person detection process, the suspicious person detection unit 352 turns on the suspicious person flag in the storage unit 34.

  When the suspicious person detection processing is completed for the read image data, it is determined whether the suspicious person detection processing is completed for all the processing target image data (step ST6), and if there is unprocessed processing target image data (step ST6-No). ), The process returns to step ST3. If no undetected signal is input, the process of reading the unprocessed processing target image data and executing the suspicious person detection process is repeated (steps ST3 to ST6).

  When an undetected signal indicating that the human sensor 32 no longer detects a human body is input in step ST3 (step ST3-Yes), the suspicious person detecting unit 352 has a predetermined delay time (for example, 5). A delay time timer is started (step ST7), and unprocessed processing target image data is read out for all processing target image data until the suspicious person detection processing is completed (step ST8). Processing is executed (step ST9).

  When the suspicious person detection process is completed for all the processing target image data in the processes of steps ST8 to ST10 (step ST10-Yes), the suspicious person detection unit 352 determines whether or not the delay time timer has timed the elapsed delay time. When the delay time has elapsed (step ST11-Yes), the reporting device 3 initializes the suspicious person flag in the storage unit 34 to OFF and ends the suspicious person monitoring operation process. If the delay time has not elapsed (step ST11-No), the process proceeds to step ST16.

  On the other hand, when no undetected signal is input in step ST3 (the human sensor is in a detection state) and the suspicious person detection processing is completed for all the processing target image data (step ST6-Yes), the suspicious person detecting unit 352 performs processing. Is advanced to step ST12. If no undetected signal is input (step ST12-No), the suspicious person detecting unit 352 reads the latest image data from the accumulated image information (step ST13), as indicated by tb in FIG. A person detection process is executed (step ST14), and the process returns to step ST12. As described above, when the suspicious person detection process for the processing target image data, which is the past image data determined by the processing image determination unit 351, is completed, the suspicious person detection unit 352 immediately before the undetected signal is input. Whether or not there is a suspicious person is monitored for the latest image data (current image data) at the time when the suspicious person detection process ends.

  Here, when an undetected signal is input from the human sensor 32 in step ST12 (step ST12-Yes), as shown by tc in FIG. 3, the suspicious person detecting unit 352 has a predetermined delay time timer (for example, 5 seconds) is started (step ST15), the latest image data (current image data) is read from the stored image information (step ST16), and a suspicious person detection process is executed (step ST17). If the delay time timer has not timed the elapsed delay time (step ST18-No), the process returns to step ST16, and the latest suspicious person detection process at the time when the previous suspicious person detection process is completed until the delay time elapses. It is monitored whether or not there is a suspicious person in the image data (current image data). When the delay time elapses (step ST18-Yes), that is, when the delay time elapses after the human body 7 close to the storage door 21 is not detected, the reporting device 3 turns off the suspicious person flag in the storage unit 34. And the suspicious person monitoring operation process is terminated.

  As described above, when the human sensor 7 detects the human body 7 in a proximity position where the human sensor 32 can open the storage door 21, the notification device 3 is in the past from the time of detection of the human body 7 and before the storage 2. It is determined whether or not there is a suspicious person for image data in which the approaching human body 7a may be imaged, and then it is determined whether or not there is a suspicious person for the latest image data. When a suspicious person whose face is not sufficiently photographed is detected for any image data, the suspicious person flag is turned ON. Further, as shown in FIG. 3G, when the human body 7 is continuously present for a predetermined time (for example, 2 seconds) at a proximity position where the storage door 21 can be opened, the stay determination unit sets the human flag to ON. . When both the suspicious person flag and the stay flag are turned ON, the notification control unit 354 transmits an emergency signal from the communication unit 33 to the remote monitoring center 5.

  The embodiments of the present invention have been described above by way of example, but the scope of the present invention is not limited to these embodiments, and can be changed or modified according to the purpose within the scope of the claims. is there.

DESCRIPTION OF SYMBOLS 1 Notification system 2 Storage 21 Storage door 3 Notification apparatus 31 Monitoring camera 32 Human sensor 33 Communication part 34 Memory | storage part 35 Control part 351 Process image determination part 352 Suspicious person detection part 353 Stay determination part 354 Notification control part 5 Monitoring center 51 Center device 52 Display 6 Communication network 7 Person close to storage 7a Person approaching close position of storage

Claims (4)

A reporting device that detects and reports a suspicious person in a monitoring space where a vault is installed,
A person detection unit for detecting the presence of a person in the proximity of the storage;
An imaging unit that periodically images the monitoring space to obtain a monitoring image;
A storage unit for storing a monitoring image acquired by the imaging unit;
A processing image determination unit that determines, as a processing image, a monitoring image acquired after a first predetermined time from the time of detection of the person by the person detection unit;
A suspicious person detection unit for detecting a suspicious person whose face is not sufficiently photographed with respect to the processed image;
A notification device characterized by comprising:
The processing image determination unit determines, as a processing image, a monitoring image acquired after the first predetermined time from the time of the person detection and before a second predetermined time shorter than the first predetermined time. The reporting device according to Item 1.
When the process for detecting a suspicious person is completed for the processed image determined by the processed image determination unit, the suspicious person detecting unit detects the suspicious person for the monitoring image that is the latest when the process for detecting the suspicious person is completed. The notification device according to claim 1 or 2, wherein a notification process is performed.
The notification device according to claim 3, wherein the suspicious person detection unit performs a process of detecting the suspicious person until a third predetermined time elapses after the person detection unit no longer detects the presence of the person.

Images