JP2006177086A - Entry and exit controller for room - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a walker type entry and exit controller capable of identifying any user entering or leaving a room without coming to a stop thereby reducing the occurrence of unfair ingress of any persons having no right to enter the room. <P>SOLUTION: This entry and exit controller comprises a picture acquiring means 101 of each moving person, an identifying means 102 for each person whose picture was already taken, a predicting means 103 for the possibility of entering the room for the person whose picture was taken, and a security controlling means 104 capable of controlling the state of locking of a gate in response to the result of identification by the identifying means 102 and the result of the prediction by the predicting means 103. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to an entrance / exit management device.

  Conventionally, a technique disclosed in Patent Document 1 is known as an entry / exit management device that authenticates a person using a camera image and manages entry.

  The technology disclosed in Patent Document 1 compares a face image captured by a camera with a face image registered in advance, and when it is determined that the images are sufficiently similar, an entrance / exit management device that permits entry It is. In the entrance / exit management device described in Patent Document 1, since it is necessary to photograph the face image to be authenticated in the same direction as the registered face image, a camera is installed at a height at which the face can be photographed from the front in front of the gate. A person stops before authentication.

  Further, as a person authentication device that can authenticate a person without depending on the photographing direction and size of the face, a technique disclosed in Patent Document 2 is known.

  The technique disclosed in Patent Document 2 uses a characteristic of the three-dimensional shape of a human face during face authentication to transform a registered face image in accordance with the orientation and size of the captured face image. Thus, a person is identified even in an image taken by a camera installed near the ceiling.

  By combining such conventional technologies, a camera is installed near the ceiling, a person walking is photographed, the person is authenticated using the photographed image, and the gate is unlocked according to the authentication result, It is possible to configure a walk-through type entrance / exit management device capable of managing entrance / exit without stopping.

  A conventional walk-through type entrance / exit management device will be described with reference to FIG.

  In FIG. 26, an image acquisition unit 2601 acquires image data from a camera (not shown) that captures a person. The camera is installed at a position where a wide area such as a ceiling can be photographed. The authentication means 2602 registers face images of persons who are permitted to enter the room in advance. The authentication unit 2602 extracts a person's face from the image acquired by the image acquisition unit 2601 using color information and a shading pattern, and the extracted face image and a face image of a person who is permitted to enter the room in advance. Are compared to determine whether or not they are the same person. Non-patent document 1 discloses a method for extracting a face portion from an image and a method for determining the identity of a face image.

  The security control unit 2603 controls the locked state of a gate (not shown) according to the determination result of the authentication unit 2602. When the authentication means 2602 successfully matches the person, the security control means 2603 unlocks the gate so that the person can enter the room.

In this way, by authenticating a person with an image of a person walking with a camera that can shoot a wide area and unlocking the gate, the walk can be completed without stopping the person. A through-type entry / exit management device can be configured.
JP-A-5-143651 JP 2003-263639 A Shigeru Akamatsu, “The Transactions of the Institute of Electrical, Information and Communication Engineers” Vol. J80-A No. 8, Institute of Electrical, Information and Communication Engineers, p. 1215-1230

  However, such a conventional entrance / exit management device has the following problems.

  In FIG. 27, an entrance / exit management device 2701 is a walk-through type entrance / exit management device configured according to the prior art.

  The gate 2802 is unlocked by the entrance / exit management device 2701. The person 2703 is an unauthorized intruder who attempts to pass through the gate 2702 illegally. A person 2704 is a person who has the authority to pass through the gate 2702.

  Since the entrance / exit management device 2701 can authenticate from a camera image capturing a wide range, when a person 2704 having authority to pass through the gate 2702 enters the imaging range of the camera, the authentication succeeds, and the gate 2702 It is unlocked.

  Further, when the person 2704 simply passes in front of the gate 2702 and moves to another place, the gate 2702 is unlocked only by entering the shooting range of the camera.

  Therefore, a person 2703 who is an unauthorized intruder can wait in front of the gate 2702 and pass the gate 2702 in anticipation of the unlocking of the gate 2702 when the authorized person 2704 passes. It is.

  In this way, in the conventional walk-through type entrance / exit management device 2701, since the authentication is completed without stopping, the entrance / exit can be managed without taking time and effort for the user, but a person with the entry authority Since the gate 2702 is unlocked just by entering the area where 2704 can be authenticated, there is a problem that the chance of unauthorized intrusion by the person 2703 who is an unauthorized intruder increases.

  The present invention provides an entry / exit management device that reduces the chances of unauthorized entry by a person without entry authority.

  The entrance / exit management device according to the present invention includes an image acquisition unit that acquires an image of a moving object, an authentication unit that authenticates the object in the acquired image, and a prediction unit that predicts the presence / absence of an entry operation of the object in the acquired image. And security control means for controlling the locked state of the gate according to the authentication result by the authentication means and the prediction result by the prediction means.

  With this configuration, the gate can be unlocked only for a person who has the authority to enter the room and is willing to enter the room.

  ADVANTAGE OF THE INVENTION According to this invention, the entrance / exit management apparatus with few opportunities of unauthorized intrusion can be provided by unlocking a gate only with respect to the person who has room entry authority and is willing to enter.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

(Embodiment 1)
FIG. 1 is a block diagram showing a configuration of a walk-through type entrance / exit management device according to Embodiment 1 of the present invention.

  In FIG. 1, an image acquisition unit 101 is installed at a position where a wide area such as a ceiling can be imaged, and acquires image data from a camera (not shown) that captures a person.

  The authentication unit 102 compares the image of the head face of a person who is permitted to enter the room in advance (hereinafter referred to as a face image) with the image acquired by the image acquisition unit 101, and the same person It is determined whether or not there is. The prediction unit 103 determines whether or not a person has an intention to enter the room based on a temporal change in the image acquired by the image acquisition unit 101. The security control means 104 controls the lock state of a gate (not shown) according to the determination result of the authentication means 102.

  FIG. 2 is a flowchart showing a processing procedure of the authentication means according to the first embodiment of the present invention.

  In FIG. 2, when the processing is started (step S200), the authentication unit 102 notifies the prediction unit 103 of an initialization command (step S201), and resets the collation flag in the internal memory to the OFF state (step S202). ). Next, an image is acquired from the image acquisition unit 101 (step S203), and it is detected whether a face image exists in the image (step S204). In the face detection process, whether the color information and shading pattern of each region in the image matches the face pattern of the person model created by the statistical method is compared, and it is determined that there is a face in the region having a high degree of matching. A specific method of face detection is disclosed in Non-Patent Document 1.

  If face detection fails (NO in step S204), the process is executed again from step S201. If the face detection is successful (YES in step S204), it is confirmed whether the collation flag in the internal memory is ON (step S205). When the collation flag is ON (YES in step S205), it is confirmed whether the processing end is notified from the predicting means 103 (step S206), and when the end is not notified (NO in step S206). The detected face area (hereinafter referred to as face size) is output to the prediction means 103 (step S210). When the prediction unit 103 notifies the end of processing (YES in step S206), the collation flag in the internal memory is changed to OFF (step S207), and face collation processing is executed (step S208).

  Further, when the collation flag is OFF when the collation flag is referred to in step S205 (NO in step S205), face collation processing is executed (step S208). The face collation process compares the detected face area image (hereinafter referred to as a collation image) with a face image of a person who is permitted to enter the room in advance (hereinafter referred to as a registered image). This is a process for determining whether or not they are the same person.

  The face matching process detects the position of the eyes from the matching image, enlarges or reduces the size of the matching image so that the eye position of the registered image matches the eye position of the matching image, and then a specific part such as a light and dark pattern or a corner Judgment is made by comparing whether or not the image features match. A specific face matching method is disclosed in Non-Patent Document 1.

  If the face collation fails (NO in step S208), the process is executed again from step S201.

  If the face collation is successful (YES in step S208), the collation flag in the internal memory is changed to ON (step S209), and the face size is output to the prediction means 103 (step S210).

  The predicting unit 103 can determine that the person is approaching the authentication device when the change amount of the face size detected by the authentication unit 102 is larger than a predetermined change amount per time. It is judged that there is, and the security control means 104 is instructed to unlock the gate.

  FIG. 3 is a flowchart showing a processing procedure of the prediction means according to the first embodiment of the present invention.

  In FIG. 3, when the prediction unit 103 starts processing (step S300), it resets the detection counter in the internal memory to 0 (step S302), and changes the face size memory in the internal memory to 0 (step S303). .

  Next, the face size detected by the face detection process (hereinafter referred to as the detected face size) is acquired from the authentication unit 102 (step S304).

  Next, the detected face size is compared with the value of the face size memory (step S305). If the detected face size is larger (YES in step S305), the detected face size is recorded in the face size memory (step S305). In step S306, the internal memory detection counter is incremented (step S307). Next, it is compared whether the value of the detection counter is larger than a preset threshold value (step S308), and if it is larger (YES in step S308), a release command is output to the security control means 104 (step S309). The authentication unit 102 is notified that the prediction process has been completed (step S310).

  If the detected face size is smaller than the face size memory (NO in step S305), it is determined that there is no intention to enter the room, and the authentication unit 102 is notified that the prediction process has been completed (step S310).

  If the detection counter does not reach the threshold value (NO in step S308), it is checked whether an initialization command is notified from the authentication unit 102 (step S301). If the initialization command is notified (step S301). If YES), the process of step S302 is executed to initialize the prediction process, and if the initialization command is not notified (NO of step S301), the process of step S304 is executed to continue the prediction process.

  After notifying the end of the prediction process in step S310, the initialization process in step S302 is executed to determine the next person's intention to enter the room.

  When the predicting unit 103 determines that the person has an intention to enter the room, the security control unit 104 unlocks the gate and permits the person to enter the room.

  Hereinafter, the operation of this apparatus will be described.

  First, an operation when a person passes through the gate will be described.

  FIG. 4 is a diagram illustrating an operation example when a person permitted to enter the room enters the room.

  In FIG. 4, a person 401 is a person permitted to enter the room. The entrance / exit management device 402 is the entrance / exit management device of the first embodiment. The gate 403 is a security gate managed by the entrance / exit management device 402.

As shown in FIG. 4, the person 401 continues to approach the gate 403 in the order of positions P1, P2, and P3.

  FIGS. 5A, 5B, and 5C show the image data taken by the image acquisition unit 101 at the respective positions P1, P2, and P3 in FIG. 4, and the broken line area in each image data is detected. The size of the face image is shown.

  At position P1 in FIG. 4, since the person 401 is far from the image acquisition unit 101, the face size detected by the authentication unit 102 is small as shown in FIG.

  When the authentication unit 102 succeeds in authenticating the person 401 at the position P1 in FIG. 4, the area of the broken line area in FIG. 5A is output to the prediction unit 103 as the face size. Since the face size memory in the prediction unit 103 is 0 in the initial state, the initial prediction process is always successful, and the prediction unit 103 records the face size in the face size memory and increments the detection counter to 1. To do.

  When the person 401 moves to the position P2 in FIG. 4, as shown in FIG. 5B, the person is photographed larger than in the case of FIG. 5A, so the face detection area indicated by the broken line is also large. Become. The area of the face detection area is detected by the authentication unit 102 and output to the prediction unit 103. The predicting means 103 compares the face size memory with the newly detected face size, so it is determined that there is an intention to enter the room, the detected face size is recorded in the face size memory, and the detection counter Increases to 2.

  Further, when the person 401 comes to the position P3 in FIG. 4, the face size further increases as shown in FIG. 5C, and therefore the detection counter of the prediction unit 103 becomes 3 by the same processing as described above.

  When the threshold value in step S308 of the prediction unit 103 is set to 3, it is determined that the person has an intention to enter the room at this point, and the security control unit 104 unlocks the gate 403 and the person 401 enters the room. It becomes possible.

  Next, an operation when a person who does not intend to enter the room passes around the gate will be described.

  FIG. 6 is a diagram for explaining an operation example in which a person permitted to enter the room passes in front of the gate.

  In FIG. 6, a person 604 is a person who stands by near the gate 403 and tries to intrude the gate 403 illegally.

  As shown in FIG. 6, the person 401 moves from the position P <b> 1 to the position P <b> 2 and approaches the gate 403, but then changes direction at the position P <b> 2, moves to the position P <b> 3, and moves away from the gate 403.

  FIGS. 7A, 7B, and 7C show image data captured by the image acquisition unit 101 when the person 401 is at each of the positions P1, P2, and P3 in FIG.

  The operations at the positions P1 and P2 in FIG. 6 are the same as the operations at the positions P1 and P2 shown in FIG.

When the person 401 moves from the position P2 to the position P3, the face size shown in FIG. 7C detected by the authentication unit 102 at the position P3 in FIG. 6 is detected in the position P2 in FIG. ) Is smaller than the face size shown. Therefore, the comparison process in step S305 of the prediction unit 103 fails, the prediction process ends, and the detection counter and the face size memory are reset. As a result, the gate 403 is not unlocked, and the person 604 who is an unauthorized intruder is not given an opportunity to pass through the gate 403.

  As described above, the entrance / exit management device 402 according to the first embodiment uses the prediction means 103 to predict the entrance prediction of the person 401 from the temporal change in the face size detected by the authentication means 102, thereby It is possible to realize a walk-through type entrance / exit management device with few opportunities for the person 604 to enter the room.

(Embodiment 2)
In the first embodiment, after the entrance prediction is successful (YES in step S308 in FIG. 3), the gate is unlocked (step S309), and the end process (step S310) and the initialization process (step S302) of the prediction unit 103 are performed. In this example, step S303) is performed. In the second embodiment, the face size comparison is continued without performing the prediction unit termination process and initialization process, and the face size becomes smaller. This is an example in which the unlocked gate is locked again.

  FIG. 8 is a flowchart showing the processing procedure of the predicting means of the entrance / exit management device according to Embodiment 2 of the present invention.

  The operation until the gate is unlocked (step S309) is the same as in FIG. 3, and after the gate is unlocked, the process of the prediction unit is continued without ending the process of the prediction unit, The difference is that when the face size comparison fails, the unlocked gate is locked.

  Specifically, when the face size is input from the authentication unit 102, the detected size and the face size memory are compared (step S305), and when the detected face size is smaller than the face size memory (NO in step S305). Confirms whether or not the gate has been unlocked (step S811). If the gate has been unlocked, the gate is locked (step S812), and notification of the end of prediction is made (step S310).

  FIG. 9 shows an operation example when a person permitted to enter the room enters the room.

  After the person 401 is authenticated as shown in FIG. 4, when the person 401 changes the moving direction at the position P3 in FIG. 9 and moves away from the gate 403 and moves to the position P4, the person is detected at the position P4 by the authentication means 102. Since the face size is smaller than the face size detected at the position P3 in FIG. 9, the gate 403 unlocked at the position P3 in FIG. 9 is locked.

  As described above, according to the second embodiment, it is possible to always determine the entrance prediction of the person 401, and when the intention to enter the room is changed midway, the gate 403 can be immediately locked, and an unauthorized person It is possible to realize a walk-through type entry / exit management device with few opportunities for entry.

(Embodiment 3)
In the first and second embodiments, the entrance prediction is determined according to the change in the size of the face of the person's head according to the passage of time. This is an example in which it is predicted that there is an entry operation due to a change in the orientation of the face of the person's head.

  FIG. 10 is a flowchart showing the processing procedure of the authentication means of the entry / exit management device according to Embodiment 3 of the present invention.

In FIG. 10, the steps from step S200 to step S209 are the same as the authentication unit 102 shown in FIG. 2, and the authentication unit 102 in the third embodiment has succeeded in face matching (YES in step S208). After the authentication flag is turned on (step S209), the face orientation is calculated (step S1010).

  FIG. 11 is a diagram for explaining a face orientation determination method.

  In FIG. 11, the authentication unit 102 further detects the positions of both eyes from the detected face. For the detection of the eye position, eye image templates created by statistical methods from many human eye images are prepared in advance, the templates are compared with the detected face area, and a portion with a high degree of similarity is determined as the eye position. The distance between the center line 1103 of the positions of both eyes and the center line 1102 of the detected face area 1101 is calculated to obtain a face orientation size 1104. In FIG. 11, reference numerals 1106 and 1107 denote the eyes.

  Next, in order to normalize this size so that it can be compared without depending on the size of the detected image, the face orientation is obtained by equation (1).

Face orientation = face orientation size 1104 / (width of face area 1105/2) × 90 (1)
In the equation (1), the size of the face orientation is calculated so that the face orientation is 90 degrees when the face is facing sideways. However, the face orientation is set to 90 when only one eye can be detected by the eye detection process, and is set to 180 when both eyes cannot be detected.

  The authentication unit 102 calculates the face orientation (step S1010), and then outputs the calculated face orientation size to the prediction unit 103 (step S1011).

  FIG. 12 is a flowchart showing the processing procedure of the predicting means of the entrance / exit management device according to Embodiment 3 of the present invention.

  In FIG. 12, step S300 to step S310 are the same as those in FIG. 3, and in the third embodiment of the present invention, a face orientation is obtained instead of the face size in that processing relating to the face size memory is unnecessary. (Step S1204) Instead of comparing the point and the face size, the detected face direction is compared with a preset face direction limit value (Step S1205).

  As shown in FIG. 4, when the person 401 faces the front and approaches the gate 403, the face direction calculated by the authentication unit 102 is always a value close to 0, and thus the face direction limit is determined in the comparison process of the prediction unit 103. It is smaller than the value (YES in step S1205). For this reason, the detection counter gradually increases, and the gate is unlocked beyond the threshold (step S309).

  However, when the person 401 moves away from the gate 403 during the movement as shown in FIG. 6, the face orientation changes as shown in FIG. 7C and is larger than the face orientation limit value in the comparison process of the prediction means 103. (NO in step S1205), the prediction process ends (step S310), and the gate is not unlocked, so that unauthorized person 604 is not allowed to enter.

  As described above, the third embodiment predicts a person's entrance from the change of the face direction calculated by the authentication unit 102 with the passage of time. An exit management device can be realized.

  Further, since the eye detection process in the authentication unit 102 can use the one used in the face matching process (step S208), the apparatus can be configured at low cost without increasing the scale of the apparatus and software constituting the authentication unit 102. can do.

(Embodiment 4)
The fourth embodiment of the present invention is an example in which the unlocked gate is re-locked when the person 401 changes direction during the movement.

  FIG. 13 is a flowchart showing the processing procedure of the prediction means of the entry / exit management device according to Embodiment 4 of the present invention.

  In FIG. 13, in the fourth embodiment, as in FIG. 8, the detection counter matches the threshold value (step S308), and the intention confirmation process (step S1206) is continued even after the gate is unlocked (step S309). If the face orientation has changed (NO in step S1206) and the gate has been unlocked (YES in step S811), the gate is locked (step S812).

  As described above, in the fourth embodiment, it is possible to always determine the predicted entry of the person 401. If the intention to enter the room is changed in the middle, the gate is immediately locked to It is possible to realize a walk-through type entrance / exit management device with few entrance opportunities.

(Embodiment 5)
Embodiment 5 of the present invention is an example in which the security control means 104 is provided with authentication status presentation means 1501 for presenting an authentication status to a person.

  FIG. 14 is a block diagram showing an example of the authentication status presenting means of the entry / exit management device according to Embodiment 5 of the present invention.

  In FIG. 14, the authentication status presenting means 1401 is provided with an authentication OK lamp 1402 and an authentication NG lamp 1403. Depending on the authentication status, one of the lamps lights up to authenticate the person walking. Present the situation.

  If it is determined that there is an intention to enter the room due to restrictions such as the configuration of the passage even though there is no intention to enter the room, it is confirmed that the authentication OK lamp 1402 of the authentication status presenting means 1401 is turned on, By performing an operation such as greatly changing the value, the prediction unit 103 can lock the gate in step S1206.

  As described above, the fifth embodiment is provided with the authentication status presenting means 1401, so that even when the intention to enter the room is erroneously determined and the gate is unlocked, Can be locked again, and a walk-through type entrance / exit management device with few opportunities for unauthorized persons to enter can be realized.

  It should be noted that the same effect can be obtained without providing an authentication status presenting means if an operation sound is generated when the gate is unlocked, or a gate that opens the door at the same time as unlocking is used, such as an automatic gate.

(Embodiment 6)
The sixth embodiment of the present invention is an example in which it is predicted that the operation is to enter the room when there is little change in the face orientation with time.

  FIG. 15 is a flowchart showing the processing procedure of the prediction means of the entry / exit management device according to Embodiment 6 of the present invention.

  In FIG. 15, steps S300 to S310 and step S1204 are the same as the prediction unit 103 of FIG. 12, but the sixth embodiment relates to a face orientation memory that records the face orientation acquired from the authentication unit 102. The difference is that the processing is added and the face orientations are compared to determine the intention to enter the room.

  The prediction unit 103 resets the detection counter to 0 at the time of initialization (step S302) and clears the face orientation memory to an unset state in which no numerical value is set (step S1502).

  Next, the face orientation is acquired (step S1204), the difference between the face orientation value and the face orientation value recorded in the face orientation memory is calculated (step S1505), and the difference is set in advance. If it is smaller than the limit value (YES in step S1505), it is determined that there is an intention to enter the room, and the detection counter is incremented (step S307). At that time, the acquired face orientation value is recorded in the face orientation memory (step S1506).

  If the face direction change amount is larger than the face direction change limit value (NO in step S1505), it is determined that there is no intention to enter the room, and the end of the prediction process is notified (step S310). However, when the face direction change amount is calculated (step S1505), if the face direction memory is not set, the comparison is not performed and the face direction acquired in the face direction memory is determined as successful (YES in step S1505). Is recorded (step S1506).

  FIG. 16 shows an operation example when there is an intention to enter the room.

  In FIG. 16, the person 401 approaches the gate 403 from an oblique direction in the order of positions P1, P2, and P3 in FIG. 17 as time elapses.

  FIGS. 17A, 17B, and 17C respectively show image data obtained by photographing the person 401 at the positions P1, P2, and P3 in FIG.

  As shown in FIG. 17, since the face direction of the person 401 does not change with time, it is determined in step S1505 of the prediction unit 103 that there is always no change in the face direction (YES in step S1505). The detection counter is increased (step S307), and the gate 403 is unlocked (step S309).

  FIG. 18 shows an operation example when there is no intention to enter the room.

  In FIG. 18, the person 401 approaches the gate in the order of the positions P1 and P2 in FIG. 16 as time elapses. However, since there is no intention to enter the room, the person 401 changes direction at the position P2 and moves to the position P3.

  FIGS. 19A, 19B, and 19C respectively show image data obtained by photographing the person 401 at the positions P1, P2, and P3 in FIG.

  As shown in FIGS. 19A and 19B, the face direction of the person 401 is constant from the position P1 to the position P2, and the face direction change amount of the prediction unit 103 is determined (step S1605). ) Succeeds and the detection counter is increased (step S307), but one eye is not visible at the position P3, and the face direction changes greatly, so the amount of change in the face direction becomes larger than the face direction limit value (NO in step S1605). ) Notifying the end of the prediction process without unlocking the gate (step S310).

As described above, the sixth embodiment is a walk-through type in which an unauthorized person has few opportunities to enter the room by determining the entrance prediction of the person 401 from the change of the face direction calculated by the authentication unit 102 over time. An entrance / exit management device can be realized.

  In the above, the face direction is detected from the position of both eyes, but it is detected by detecting a part of the other face such as the nose and mouth from the detected face area or by using the contrast of the face. The determination may be made by folding and comparing the face area at the center.

(Embodiment 7)
The seventh embodiment of the present invention is an example in which it is predicted that there is an entry operation due to a change in the position of a person over time.

  FIG. 20 is a flowchart showing the processing procedure of the authentication means of the entry / exit management device according to Embodiment 7 of the present invention.

  20, the operations from step S200 to step S209 are the same as those in FIG. 2, and the seventh embodiment is different in that the position of a person is output to the prediction means 103 instead of the face size.

  In FIG. 20, the authentication unit 102 calculates the position of the person from the detected face area when the face collation succeeds (YES in step S208) and turns on the collation flag (step S209), and then to the prediction unit 103. Output (step S2010).

  FIG. 21 is a diagram for explaining a method for calculating a person position.

  In FIG. 21, the face authentication unit 102 obtains a center line 2102 of the detected face area 2101, calculates a displacement 2104 from the center line 2103 of the screen, and outputs it to the prediction unit 103 as a person position.

  FIG. 22 shows a flowchart of the processing procedure of the prediction means of the entry / exit management device according to Embodiment 7 of the present invention.

  In FIG. 22, steps S300 to S310 are the same as those in FIG. 3, and the seventh embodiment of the present invention processes the person position instead of the face size, and the person position is stored in the person position memory in the internal memory. It is different in that it is determined that there is an intention to enter the room when it is smaller than the value.

  The prediction unit 103 includes a person position memory in the internal memory, and clears the person position memory to an unset state at the time of initialization (step S2203).

  Next, the person position is acquired from the authentication means 102 (step S2204), and the acquired person position is compared with the value of the person position memory (step S2205), and the person position is small, that is, close to the center of the image. It is determined that there is an intention to enter the room (YES in step S2205), the value of the person position is recorded in the value of the person position memory (step S2206), and the detection counter is incremented (step S307).

  In addition, when a person approaches the camera from the front, that is, when the position of the person in the acquired image does not change from a predetermined range over time (YES in step S2205), there is an intention to enter the room. And process in the same manner as above.

  However, if the value of the person position memory is not set, it is determined that there is an intention to enter the room unconditionally, and the process of YES in step S2205 is performed.

  Next, the operation of the present apparatus will be described using the example of FIG.

  FIG. 23 is a diagram illustrating the positions of persons in the image data at the positions P1, P2, and P3 in FIG.

  In FIG. 23, the distance between the straight lines 2301 and 2302 indicated by the broken lines is the person position calculated by the authentication means 102, and when the person 401 approaches the authentication means 102 from an angle as shown in FIG. The position is getting smaller. That is, the person position moves away from the center of the image. In this case, the prediction means 103 always determines that there is an intention to enter the room in step S2205, the detection counter continues to increase (step S307), and the gate is unlocked (step S309).

  The operation when the gate is not unlocked will be described using the example of FIG.

  24A, 24B, and 24C are diagrams respectively showing the positions of persons in the image data at the positions P1, P2, and P3 in FIG.

  In FIG. 24, the distance between the straight lines 2401 and 2402 indicated by the broken lines is the person position calculated by the authentication means 102, and when the person 401 approaches the authentication device from the position P1 to the position P2 in FIG. The person position gradually decreases, and it is determined by the prediction means 103 that there is an intention to enter the room (YES in step S2205). However, when the person moves from the position P2 to the position P3, the person 401 moves away from the center position, and the value in the person position memory Since the value of the calculated person position becomes larger, it is determined that there is no intention to enter the room (NO in step S2305), and the notification of the end of the prediction process (step S310) is performed without unlocking the gate.

  As described above, the seventh embodiment is a walk-through type in which an unauthorized person has few opportunities to enter the room by determining his / her intention to enter the room from the change of the person position calculated by the authentication unit 102 over time. An entrance / exit management device can be realized.

(Embodiment 8)
Embodiment 8 of the present invention is an example in which the gate is locked again when the intention to enter the room changes by continuing the prediction process after the gate is unlocked.

  FIG. 25 is a flowchart showing the processing procedure of the prediction means of the entry / exit management device according to Embodiment 8 of the present invention.

  25, step S300 to step S310, step S811, and step S2202 to step S2206 are the same as those in FIGS. 8 and 22, and in the eighth embodiment, the gate is unlocked (step S309). It is determined whether or not there is an intention to enter the room (step S2205). If it is determined that the position of the person has moved away from the center of the image data and the intention to enter has been lost (NO in step S2205), the gate is locked (step S2205). S812).

  The operation of this apparatus will be described using the example of FIG.

In FIG. 9, when the person 401 moves from the position P1 to the position P3, since the person position approaches the center of the image data, the detection counter is increased and the gate is unlocked. However, if the direction is changed at the position P3 and moved to the position P4, the person position moves away from the center of the image data, so it is determined that there is no intention to enter the room (NO in step S2205), and the gate is locked (
Step S812).

  As described above, in the eighth embodiment, the intention of the person 401 is changed after the gate is unlocked by determining the intention of the person 401 to enter the room from the change of the person position calculated by the authentication unit 102 over time. In this case, by locking the gate again, it is possible to realize a walk-through type entrance / exit management device with few opportunities for unauthorized persons to enter the room.

  In the present invention, the object position is determined based on the position of the head or the eye included in the object in the acquired image as a predicting means for predicting the presence or absence of the room entry operation based on the temporal change in the object position in the acquired image. What is determined may be used.

  In the above embodiment, the position of the person is compared only in the horizontal direction. However, the present invention can also perform the comparison of the position of the object in the vertical direction.

  Further, the determination of the intention to enter the room can be similarly performed by combining the plurality of methods described above. For example, by taking into account temporal changes in the size and orientation of the head at the same time, it is possible to perform processing such as determining that there is an intention to enter the room only when looking at the camera and a person approaches. Can be improved.

  The entrance / exit management device of the present invention has the effect of reducing the chances of unauthorized intrusion by unlocking the gate only for persons who have entry authority and are willing to enter the room. This is useful as an entrance / exit management device with a low user load that can authenticate a person without stopping.

The block diagram which shows the structure of the entrance / exit management apparatus in Embodiment 1 of this invention. The flowchart which shows the process sequence of the authentication means of the entrance / exit management apparatus in Embodiment 1 of this invention. The flowchart which shows the process sequence of the prediction means of the entrance / exit management apparatus in Embodiment 1 of this invention. The figure explaining the example of operation when the person who was permitted to enter the room enters the room The figure which shows an example of the image data when a person approaches continuously The figure explaining the operation example where the person who is allowed to enter the room passes in front of the gate The figure which shows an example of the image data of the person in each position of FIG. The flowchart which shows the process sequence of the prediction means of the entrance / exit management apparatus in Embodiment 2 of this invention. The figure explaining the example of operation when the person who was permitted to enter the room enters the room The flowchart which shows the process sequence of the authentication means of the entrance / exit management apparatus in Embodiment 3 of this invention. The figure explaining the method of face orientation judgment The flowchart which shows the process sequence of the prediction means of the entrance / exit management apparatus in Embodiment 3 of this invention. The flowchart which shows the process sequence of the prediction means of the entrance / exit management apparatus in Embodiment 4 of this invention. The block diagram of the authentication condition presentation means of the entrance / exit management apparatus in Embodiment 5 of this invention The flowchart which shows the process sequence of the prediction means of the entrance / exit management apparatus in Embodiment 6 of this invention. Diagram explaining example of operation when there is intention to enter the room The figure which shows an example of the image data of the person in each position of FIG. Diagram explaining an operation example when there is no intention to enter the room The figure which shows an example of the image data of the person in each position of FIG. The flowchart which shows the process sequence of the authentication means of the entrance / exit management apparatus in Embodiment 7 of this invention. The figure explaining the calculation method of a person position The flowchart which shows the process sequence of the prediction means of the entrance / exit management apparatus in Embodiment 7 of this invention. The figure which shows the position of the person in the image data image | photographed in each position of FIG. The figure which shows the position of the person in the image data image | photographed at each position of FIG. The flowchart which shows the process sequence of the prediction means of the entrance / exit management apparatus in Embodiment 8 of this invention. Block diagram showing the configuration of a conventional entrance / exit management device The figure explaining the subject of the conventional entrance / exit management apparatus

Explanation of symbols

DESCRIPTION OF SYMBOLS 101 Image acquisition means 102 Authentication means 103 Prediction means 104 Security control means 401 Person 402 Entrance / exit management device 403 Gate 1401 Authentication status presentation means

Claims (17)

Image acquisition means for acquiring an image of a moving object;
Authentication means for authenticating the object in the acquired image;
A predicting means for predicting presence / absence of an entry operation of an object in the acquired image;
Security control means for controlling the locked state of the gate according to the authentication result by the authentication means and the prediction result by the prediction means;
Entrance / exit management device including The entrance / exit management device according to claim 1, wherein the predicting unit predicts presence / absence of a room entry operation based on a temporal change in the size of the head of an object in an acquired image. The entrance / exit management device according to claim 2, wherein the predicting unit predicts that there is an entrance operation when the amount of change in the size of the head in the acquired image is greater than a predetermined amount of change per time. The entrance / exit management device according to claim 2, wherein the predicting unit predicts that there is no entrance operation when a change amount of the size of the head in the acquired image is smaller than a predetermined change amount per time. The entrance / exit management device according to claim 1, wherein the predicting unit predicts presence / absence of an entrance operation based on a temporal change in the orientation of the head of an object in the acquired image. The entrance / exit management device according to claim 5, wherein the predicting unit predicts that there is an entrance operation when the orientation of the head in the acquired image does not change from the front. The entrance / exit management device according to claim 5, wherein the predicting unit predicts that there is an entrance operation when the orientation of the head in the acquired image approaches the front as time passes. The entrance / exit management device according to claim 5, wherein the predicting unit predicts that there is no entrance operation when the orientation of the head in the acquired image moves away from the front as time passes. The entrance / exit management device according to claim 5, wherein the predicting unit predicts that there is no entrance operation when the orientation of the head in the acquired image changes more than a predetermined size with time. The entrance / exit management device according to claim 5, wherein the predicting unit determines the orientation of the head in the acquired image based on a change in the position of an eye included in the head in the acquired image. The entrance / exit management device according to claim 1, wherein the predicting unit predicts presence / absence of an entrance operation based on a temporal change in an object position in an acquired image. The entrance / exit management device according to claim 11, wherein the predicting unit predicts that there is an entry operation when the object position in the acquired image does not change from a predetermined range over time. The entrance / exit management device according to claim 11, wherein the predicting unit predicts that there is an entrance operation when the object position in the acquired image approaches a predetermined range as time elapses. The entrance / exit management device according to claim 11, wherein the predicting unit predicts that there is no entrance operation when the object position in the acquired image moves away from a predetermined range as time elapses. The entrance / exit management device according to claim 11, wherein the predicting unit predicts that there is no entrance operation when the object position in the acquired image changes more than a predetermined range over time. The entrance / exit management device according to claim 11, wherein the predicting unit determines the position of the object based on the position of the head included in the object in the acquired image. The entrance / exit management device according to claim 11, wherein the prediction unit determines the position of an object based on a position of an eye included in the object in the acquired image.

Images