Authentication system

Abstract

To enable an authentication system that opens gates to individuals who have been authenticated from images captured by cameras to accurately determine whether the individuals detected by the cameras are the same individuals entering the gates.SOLUTION: An authentication system according to an embodiment of the present disclosure includes a camera 10 installed toward a lane 4 in front of a gate 2 and a processing circuit 20 connected to the camera 10 for communication. The processing circuit 20 detects persons in images captured by the camera 10, measures the size of the detected persons' points of interest in the images, converts the measured sizes of the points of interest into distances from the camera 10, sets authentication ranges corresponding to the width of the lane 4 in the images according to the distances from the camera 10 to the points of interest, and authenticates the persons based on whether the points of interest fall within the set authentication ranges.SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] The present disclosure relates to an authentication system that takes a picture of a person entering a gate with a camera and authenticates the person from the captured image. [Background technology]

[0002] Patent Document 1 discloses a system configured to start a matching process for a person when it is determined that the eye distance of a person attempting to pass through an adjacent gate exceeds a condition value set for the area in which the person is located. However, there is a demand for more accurate determination of which gate a person captured on camera is heading toward. In addition to Patent Document 1, Patent Document 2 can also be cited as an example of a document that shows the state of the art in the technical field related to the present disclosure. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 2019-194875 [Patent Document 2] Japanese Patent Application Publication No. 2019-053675 Summary of the Invention [Problem to be solved by the invention]

[0004] The present disclosure has been made in view of the above-mentioned problems, and one objective of the present disclosure is to enable an authentication system that opens a gate to a person who can be authenticated from an image captured by a camera to accurately determine whether a person detected by the camera is the person entering the gate. [Means for solving the problem]

[0005] According to one embodiment of the present disclosure, an authentication system includes a camera positioned facing the lane ahead of the gate and a processing circuit communicatively connected to the camera. The processing circuit is configured to detect a person in an image captured by the camera, measure the size of a region of interest of the detected person in the image, convert the measured size of the region of interest into a distance from the camera to the region of interest, set an authentication range corresponding to the width of the lane in the image according to the distance from the camera to the region of interest, and authenticate the detected person when the region of interest is within the set authentication range. [Effects of the Invention]

[0006] With the authentication system configured as described above, when a person in front of the gate is detected by the camera, the authentication range corresponding to the width of the lane is set according to the distance from the camera to the person, so it is possible to determine whether the person is in the lane based on the positional relationship between the authentication range and the target part in the image. This makes it possible to accurately determine whether the person detected by the camera is a person entering the gate. [Brief explanation of the drawings]

[0007] [Figure 1] FIG. 1 is a diagram illustrating a problem with a security gate system. [Figure 2] FIG. 1 is a diagram illustrating a configuration of an authentication system according to an embodiment of the present disclosure. [Figure 3] FIG. 10 is a diagram illustrating a method for setting an authentication range by an authentication system according to an embodiment of the present disclosure. [Figure 4] 1A and 1B are diagrams illustrating specific examples of installation positions and angles of view of cameras in an authentication system according to an embodiment of the present disclosure. [Figure 5] FIG. 2 is a block diagram showing a configuration of a processing circuit of an authentication system according to an embodiment of the present disclosure. [Figure 6] FIG. 10 is a flowchart showing a flow of processing executed by a processing circuit of an authentication system according to an embodiment of the present disclosure. DETAILED DESCRIPTION OF THE INVENTION

[0008] 1. Issues with security gate systems The authentication system is part of a security gate system, and authenticates people from images captured by a camera, opening the gate to those who are successfully authenticated. Ensuring security within the gated area requires accurate determination of whether the person detected by the camera is the person entering the gate. The challenges in ensuring this accurate determination are explained using Figure 1.

[0009] Figure 1 shows a typical configuration of a security gate system. Gate 2A is installed in lane 4A, which is separated by partitions 6A and 6B. Gate 2B is installed in lane 4B, which is separated by partitions 6B and 6C. Lane 4A is a standard-width lane for general use, while lane 4B is wider than lane 4A and is suitable for wheelchairs, bicycles, strollers, etc. The edges of the partitions 6A, 6B, and 6C that separate lanes 4A and 4B are aligned. Camera 10A is installed in gate 2A. Camera 10B is also installed in gate 2B. Generally, the specifications of cameras 10A and 10B installed at gates 2A and 2B are the same, including the angle of view.

[0010] Camera 10A for gate 2A is installed with its imaging range 12A facing lane 4A. Specifically, camera 10A is installed on partition 6A, and its imaging direction is adjusted so that the left edge of its angle of view is parallel to lane 4A. As a result, the left edge of camera 10A's angle of view is approximately aligned with the left side of lane 4A, and the area inside partition 6A is within the imaging range of camera 10A. In addition, the right edge of camera 10A's angle of view passes through the tip of partition 6B. As a result, the entrance to lane 4A falls within imaging range 12A.

[0011] Camera 10B for gate 2B is installed with its imaging range 12B facing lane 4B. Specifically, camera 10B is installed on partition 6B, and its imaging direction is adjusted so that the left edge of its angle of view is parallel to lane 4B. As a result, the left edge of camera 10B's angle of view roughly coincides with the left side of lane 4B, and the area inside partition 6B is within the imaging range of camera 10B. On the other hand, the right edge of camera 10B's angle of view passes ahead of the tip of partition 6C. As a result, the entrance to lane 4B does not fall within imaging range 12B, and a gap 8 exists between the tip of partition 6C and imaging range 12B, which is a blind spot for camera 10B.

[0012] A typical method for determining whether a person is entering a gate is to determine whether the person is captured in an image captured by a camera installed at the gate. When this typical method is applied to the example shown in Figure 1, a person captured in an image captured by camera 10A for gate 2A, i.e., a person within capture range 12A, is determined to be a person entering gate 2A. Similarly, a person captured in an image captured by camera 10B for gate 2B, i.e., a person within capture range 12B, is determined to be a person entering gate 2B.

[0013] Now, let's assume that there are three people, HA, HB, and HC, heading towards a gate. Person HA is walking straight down the center of lane 4A as seen from gate 2A, towards gate 2A. Because person HA is within the shooting range 12A of camera 10A, it can be determined from the image from camera 10A that person HA is entering gate 2A. In this case, authentication to open gate 2A is performed based on the image of person HA captured by camera 10A.

[0014] Person HB is walking from the left end of lane 4B as seen from gate 2B towards gate 2B. Person HB is within the shooting range 12B of camera 10B, but is also within the shooting range 12A of camera 10A. For this reason, authentication to open gate 2B is performed based on the image of person HB captured by camera 10B, and at the same time, authentication to open gate 2A is also performed based on the image of person HB captured by camera 10A. In other words, person HB entering the adjacent gate 2B is mistakenly authenticated as a person entering gate 2A.

[0015] Person HC is walking towards gate 2B from the outside right of lane 4B as seen from gate 2B. Person HC is not within the shooting range 12B of camera 10B, and if he passes through gap 8, which is his blind spot, he can enter lane 4B without entering shooting range 12B. However, because person HC is not visible to camera 10B, he cannot be authenticated to open gate 2B. In this case, a situation occurs in which gate 2B does not open, even though person HC is qualified to pass through gate 2B.

[0016] As described above, typical conventional security gate systems have the problem of mistakenly identifying a person walking in an adjacent lane as someone entering the gate of their own lane. Another problem is that they are unable to identify a person entering the lane from outside the lane through the camera's blind spot. Hereinafter, the former problem will be referred to as the first problem, and the latter problem as the second problem.

[0017] 2. Overview of the authentication system An authentication system according to an embodiment of the present disclosure has been devised as a solution to the first and second problems described above. First, the overall configuration of the authentication system according to this embodiment will be described with reference to Fig. 2. For ease of explanation, the authentication system according to this embodiment is based on the security gate system shown in Fig. 1, and elements common to the security gate system shown in Fig. 1 are designated by common reference numerals.

[0018] The authentication system 100 according to this embodiment includes processing circuits 20A and 20B in addition to cameras 10A and 10B. The processing circuit 20A is communicatively connected to the camera 10A. The processing circuit 20A processes the image captured by the camera 10A and is programmed to set an authentication range 14A within the image capture range for authenticating a person. The authentication range 14A is a range within the image capture range of the camera 10A that corresponds to the width of lane 4A. Similarly, the processing circuit 20B is communicatively connected to the camera 10B. The processing circuit 20B is programmed to process the image captured by the camera 10B and set an authentication range 14B within the image capture range for authenticating a person. The authentication range 14B is a range within the image capture range of the camera 10B that corresponds to the width of lane 4B.

[0019] The authentication ranges 14A and 14B are set to extract only those persons who should be authenticated from among those within the image capture ranges of the cameras 10A and 10B. In the example shown in FIG. 2, person HB is within the image capture range of the camera 10A but is not within the authentication range 14A. Therefore, person HB is not determined to be a person entering gate 2A, and authentication based on the image captured by the camera 10A is not performed. On the other hand, person HB is within the authentication range 14B of the camera 10B. Therefore, person HB is determined to be a person entering gate 2B, and authentication based on the image captured by the camera 10B is performed. In other words, by setting the authentication ranges 14A and 14B within the image capture ranges of the cameras 10A and 10B, it is possible to prevent a person walking in an adjacent lane from being mistakenly identified as a person entering the gate of the lane in question. In other words, the first problem can be solved.

[0020] The authentication system 100 includes monitors 30A and 30B. Monitor 30A is installed near camera 10A. Monitor 30A is communicatively connected to processing circuit 20A and displays the detection and authentication results from camera 10A toward the person in lane 4A. Similarly, monitor 30B is installed near camera 10B. Monitor 30B is communicatively connected to processing circuit 20B and displays the detection and authentication results from camera 10B toward the person in lane 4B. From the displays on these monitors, the person in the lane can know whether they have been detected by the camera. In the example shown in FIG. 2, the detection and authentication results for person HA are displayed on monitor 30A, and the detection and authentication results for person HB are displayed on monitor 30B. Specifically, if a person is detected, a "Hello" logo screen is displayed as the detection result, and if the person is authenticated, a "Ready" logo screen is displayed as the authentication result. However, if person HC enters lane 4B through gap 8, which is in the blind spot of camera 10B, person HC will not be detected by camera 10B, and the detection result for person HC will not be displayed on monitor 30B. In this case, a logo screen prompting the person to show their face in front of the camera is displayed on monitor 30B. Therefore, person HC will know from the display on monitor 30B that they have not been detected by camera 10B, and will try to move to a position where they can be seen by camera 10B. This solves the second problem of not being able to authenticate a person entering the lane from outside the lane through the blind spot of the camera.

[0021] Furthermore, the authentication system 100 includes an attention warning indicator 32. The attention warning indicator 32 is provided at a location where lane 4B is entered from an area outside the field of view of camera 10B, i.e., in gap 8, which is a blind spot of camera 10B. The attention warning indicator 32 is provided, for example, on the floor and displays a warning that the area is outside the detection range of camera 10B. By providing such an attention warning indicator 32, the person HC will hesitate to enter lane 4B through gap 8 and will instead bypass the attention warning indicator 32 and head toward gate 2B through the authentication range 14B of camera 10B. In other words, the attention warning indicator 32 also contributes to solving the second problem. The display by the attention warning indicator 32 may be always displayed, or may be displayed only while gate 2B is functioning in response to a signal from processing circuit 20B.

[0022] 3. How to set the authentication scope using the authentication system As described above, the authentication system 100 solves the first problem by setting authentication ranges 14A and 14B for each gate 2A and 2B. A method for setting authentication ranges 14A and 14B by the processing circuit 20 of the authentication system 100 will be described below with reference to FIG. 3. In the following description, gates 2A and 2B will be collectively referred to as gate 2, lanes 4A and 4B will be collectively referred to as lane 4, cameras 10A and 10B will be collectively referred to as camera 10, shooting ranges 12A and 12B will be collectively referred to as shooting range 12, authentication ranges 14A and 14B will be collectively referred to as authentication range 14, and processing circuits 20A and 20B will be collectively referred to as processing circuit 20. Also in FIG. 3, each element will be referred to by the same reference numerals.

[0023] As described above, authentication range 14 is a range within the imaging range of camera 10 that corresponds to the width of lane 4. As can be seen from the relationship between imaging range 12 and authentication range 14 of camera 10, the ratio of authentication range 14 to imaging range 12 changes depending on the distance from camera 10. For example, imaging range 12 and authentication range 14 coincide at distance D1, but the ratio of authentication range 14 to imaging range 12 becomes smaller at distance D2 that is farther than distance D1, and the ratio of authentication range 14 to imaging range 12 becomes even smaller at distance D3 that is farther than distance D2. The relationship between the ratio of authentication range 14 to imaging range 12 and the distance from camera 10 is calculated in advance and registered in the memory of processing circuit 20.

[0024] FIG. 3 shows images IMG1, IMG2, and IMG3 of camera 10 corresponding to the relationship between the shooting range 12 and the authentication range 14 shown in the plan view. Image IMG1 is an image corresponding to the shooting range 12 of camera 10 at distance D1. In image IMG1, the area corresponding to the authentication range 14 of camera 10 at distance D1 is authentication range R1. Since the shooting range 12 and the authentication range 14 coincide at distance D1, the authentication range R1 and image IMG1 also coincide. Image IMG2 is an image corresponding to the shooting range 12 of camera 10 at distance D2. In image IMG2, the area corresponding to the authentication range 14 of camera 10 at distance D2 is authentication range R2. The ratio of authentication range R2 to image IMG2 coincides with the ratio of authentication range 14 to shooting range 12 at distance D2. Image IMG3 is an image corresponding to the shooting range 12 of camera 10 at distance D3. In image IMG3, the area corresponding to the authentication range 14 of camera 10 at distance D3 is authentication range R3. The ratio of the authentication range R3 to the image IMG3 matches the ratio of the authentication range 14 to the shooting range 12 at the distance D3.

[0025] The above authentication ranges R1, R2, and R3 are set based on the distance from the camera 10 to the person in the images IMG1, IMG2, and IMG3. If a part of the human body that is relatively consistent in size across ages and genders is defined as the site of interest, the size of that site of interest in the image is proportional to the person's distance from the camera 10. Based on this fact, the processing circuit 20 measures the distance from the camera 10 to the person in the image captured by the camera 10. First, the processing circuit 20 detects the person in the image captured by the camera 10. Next, the processing circuit 20 measures the size (number of pixels) of the person's site of interest in the image and converts the size of the site of interest into the distance from the camera 10 to the site of interest. An example of the site of interest is the head, and examples of the size of the site of interest are the length of the face and the distance between the eyes. The authentication range is then set according to the distance of the site of interest from the camera 10.

[0026] For example, when image IMG1 is acquired, the processing circuit 20 measures the face length P1 of person H1 in image IMG1 and measures the distance D1 of person H1 from the camera 10 based on a pre-registered proportional relationship. Then, an authentication range R1 is set based on the relationship between the ratio of the authentication range to the shooting range and the distance from the camera 10, which is pre-registered. Furthermore, when image IMG2 is acquired, the processing circuit 20 measures the face length P2 of person H2 in image IMG2 and measures the distance D2 of person H2 from the camera 10 based on the pre-registered proportional relationship. Then, an authentication range R2 is set based on the relationship between the ratio of the authentication range to the shooting range and the distance from the camera 10, which is pre-registered. Person H4 has the same face length as person H2, and the same authentication range R2 as person H2 is set for such person. Furthermore, when image IMG3 is acquired, the processing circuit 20 measures the face length P3 of person H3 in image IMG3 and measures the distance D3 of person H3 from the camera 10 based on the pre-registered proportional relationship. Then, an authentication range R3 is set based on the relationship between the ratio of the authentication range to the pre-registered shooting range and the distance from the camera 10. Person H5 has the same face length as person H3, and the same authentication range R3 as person H3 is set for such a person.

[0027] The processing circuit 20 determines whether a person is a target for authentication based on whether the person's interest area is within the set authentication ranges R1, R2, and R3. In the case of image IMG1, the face of person H1 is within authentication range R1 and is therefore a target for authentication. In the case of image IMG2, the face of person H2 is within authentication range R2 and is therefore a target for authentication, but the face of person H4 is not within authentication range R2 and is therefore not a target for authentication. In the case of image IMG3, the face of person H3 is within authentication range R3 and is therefore a target for authentication, but the face of person H5 is not within authentication range R3 and is therefore not a target for authentication.

[0028] As described above, according to the authentication system 100, when a person ahead of gate 2 is detected by camera 10, the authentication range corresponding to the width of lane 4 is set according to the distance from camera 10 to the person, so it is possible to determine whether the person is in lane 4 from the positional relationship between the authentication range and the target part in the image. This makes it possible to accurately determine whether the person detected by camera 10 is a person entering gate 2.

[0029] 4. Specific examples of camera installation positions and field of view The installation position and angle of view of the camera 10 are important for accurately detecting a person entering the gate 2. FIG. 4 is a diagram illustrating a specific example of the installation position and angle of view of the camera 10 in the authentication system 100.

[0030] In the example shown in Fig. 4, camera 10 is installed at a height of 1250 mm from the floor. Camera 10 is integrated with monitor 30, and camera 10 is provided above monitor 30. The vertical angle of view of camera 10 is set to 63 degrees. Also, the authentication distance using the image from camera 10 is assumed to be 2000 mm.

[0031] In FIG. 4, F1 indicates the facial height of a wheelchair-bound Japanese first-grade elementary school student, F2 indicates the average facial height of a Japanese first-grade elementary school girl, and F3 indicates the average facial height of a Japanese adult woman. Furthermore, F4 indicates the average facial height of a Japanese adult male, and F5 indicates the maximum anticipated facial height of a Japanese adult male. By setting the installation position and field angle of camera 10 as described above, an authentication range of 1000 mm or more can be provided for all anticipated facial heights. This eliminates the need for any person to stand still in front of camera 10, enabling so-called walk-through authentication.

[0032] 5. Details of the authentication system processing circuit Next, details of the processing circuit 20 of the authentication system 100 will be described with reference to Fig. 5. Fig. 5 is a block diagram showing the configuration of the processing circuit 20. The processing circuit 20 is composed of a face detection circuit 22, a position determination circuit 24, a face authentication circuit 26, and a gate control circuit 28. Each of the circuits 22, 24, 26, and 28 may be composed of separate hardware and separate software, or may be composed of common hardware and separate software running on it.

[0033] The face detection circuit 22 is a circuit that captures an image captured by the camera 10 and detects the faces of people appearing in the image. The output of the face detection circuit 22 is input to a position determination circuit 24. The output of the face detection circuit 22 is also input to a monitor 30 connected to the processing circuit 20.

[0034] The position determination circuit 24 is a circuit that determines whether the image of the face detected by the face detection circuit 22 is located within the authentication range. The output of the position determination circuit 24 is input to a face authentication circuit 26.

[0035] The face authentication circuit 26 is a circuit that performs authentication processing on a face image that is determined by the position determination circuit 24 to be located within the authentication range. The authentication processing refers to face information registered in a face information storage database 42. The face information storage database 42 is provided on the cloud and is shared among multiple authentication systems 100. The face authentication circuit 26 is connected to the face information storage database 42 via a network 40. The output of the face authentication circuit 26 is input to a gate control circuit 28. The output of the face authentication circuit 26 is also input to a monitor 30 connected to the processing circuit 20.

[0036] The gate control circuit 28 is a circuit that controls the opening and closing of the gate 2 based on the authentication result of the face authentication circuit 26 when it detects a person approaching the gate 2. If the face to be authenticated is authenticated by the face authentication circuit 26, the gate control circuit 28 opens the gate 2 to the person corresponding to that face. On the other hand, if the face to be authenticated is not authenticated by the face authentication circuit 26, the gate control circuit 28 keeps the gate 2 closed to the person corresponding to that face.

[0037] The detailed functions of the processing circuit 20 configured as above will be described with reference to Fig. 6. Fig. 6 is a flow chart showing the flow of processing executed by the processing circuit 20.

[0038] First, the face detection circuit 22 starts acquiring an image from the camera 10. In step S11, the face detection circuit 22 determines whether a face has been detected in the image acquired from the camera 10. The face detection circuit 22 repeats the determination in step S11 until a face is detected. If a face is detected, in step S12, the face detection circuit 22 cuts out the detected face image from the image acquired from the camera 10. The face detection circuit 22 inputs the cut-out face image to the position determination circuit 24 and displays it on the monitor 30. Alternatively, instead of displaying the face image, the face detection circuit 22 displays a logo screen on the monitor 30 indicating that a face has been detected.

[0039] Upon receiving a facial image input from the face detection circuit 22, the position determination circuit 24 begins to determine the position of the facial image. In step S21, the position determination circuit 24 measures the size of the input facial image. Next, in step S22, the position determination circuit 24 calculates the distance from the camera 10 to the detected face based on the size of the facial image. Furthermore, in step S23, the position determination circuit 24 sets an authentication range for the facial image based on the distance from the camera 10 to the detected face. Then, in step S24, the position determination circuit 24 determines whether the input facial image is located within the authentication range. If the facial image input from the face detection circuit 22 is located within the authentication range, the position determination circuit 24 inputs the facial image to the face authentication circuit 26 as an object to be authenticated.

[0040] Upon receiving the facial image input from the position determination circuit 24, the facial authentication circuit 26 begins facial authentication processing. In step S31, the facial authentication circuit 26 performs authentication on the facial image by comparing the facial image of the authentication target with the facial information storage database 42. In step S32, the facial authentication circuit 26 determines whether the facial image of the authentication target can be authenticated. If the facial image of the authentication target can be authenticated, in step S33, the facial authentication circuit 26 displays an OK message on the monitor 30, indicating that authentication was successful, and inputs the facial authentication result to the gate control circuit 28. On the other hand, if the facial image of the authentication target cannot be authenticated, in step S34, the facial authentication circuit 26 displays an NG message on the monitor 30, indicating that authentication was not successful.

[0041] Upon receiving the face authentication result from the face authentication circuit 26, the gate control circuit 28 begins controlling the gate 2. In step S41, the gate control circuit 28 determines whether a passing object has been detected based on a signal from the gate sensor 2b. The gate sensor 2b is a sensor provided in front of the gate 2 and outputs a signal when it detects a passing object. The gate sensor 2b is connected to the gate control circuit 28. If the gate sensor 2b detects a passing object, in step S42, the gate control circuit 28 determines whether a positive face authentication result has been input. If a positive face authentication result has been input, in step S43, the gate control circuit 28 instructs the gate opening / closing device 2a to open the gate 2. The gate opening / closing device 2a is an actuator that opens and closes the gate 2. On the other hand, if a positive face authentication result has not been input, the gate control circuit 28 displays on the monitor 30 a message prompting authentication using a camera. If a passing object is not detected by the gate sensor 2b despite the input of a face authentication result, in step S44, the gate control circuit 28 waits for a predetermined time. If no passing object is detected after a predetermined time has elapsed, the gate opening and closing device 2a resets the face authentication result input from the face authentication circuit 26 in step S45. [Explanation of symbols]

[0042] 2, 2A, 2B...gate, 4, 4A, 4B...lane, 6A, 6B, 6C...partition, 8...gap, 10, 10A, 10B...camera, 12, 12A, 12B...shooting range, 14, 14A, 14B...authentication range, 20, 20A, 20B...processing circuit, 30, 30A, 30B...monitor, 32...attention indicator, 100...authentication system

Claims

1. A camera is positioned facing the lane in front of the gate, a processing circuit communicatively connected to the camera; The processing circuitry Detecting a person in an image captured by the camera; measuring the size of a region of interest of the person in the image; converting the size of the region of interest into a distance of the region of interest from the camera; setting an authentication range corresponding to a width of the lane in the image according to the distance; and performing authentication of the person when the target part is found to be within the authentication range. An authentication system comprising:

2. 2. The authentication system according to claim 1, The camera is installed at the right or left end of the gate, and the photographing direction is adjusted so that either the right or left end of the camera's angle of view is parallel to the lane. An authentication system comprising:

3. 2. The authentication system according to claim 1, The camera is installed so that the right edge of the image coincides with the right side of the lane or the left edge of the image coincides with the left side of the lane. An authentication system comprising:

4. 2. The authentication system according to claim 1, A warning indicator is provided at a location where the vehicle enters the lane from an area outside the camera's angle of view, and alerts the driver that the vehicle is outside the camera's detection range. An authentication system comprising:

5. 2. The authentication system according to claim 1, a monitor provided near the camera for displaying the detection results of the camera to a person in the lane; An authentication system comprising:

Images