Face recognition gate pass determination method with secondary verification

Abstract

The application discloses a kind of secondary verification face recognition gate pass by false pretenses determination method, comprising the following steps: S1: the face appearing in camera video is detected, and the face track information is recorded;S2: determine the passenger ID for identity verification as the first ID, and record the first ID face information, empty all historical track information;S3: re-detect the face appearing in camera video, and record the face track information;S4: when passenger face leaves camera video edge, determine that the passenger is passing through gate door, while recording the face information of current ID;S5: compare whether the face of pass passenger and the face of verification are the first ID face, if not, it is determined as pass by false pretenses.The application uses passenger face imaging to leave camera edge, and does secondary face verification, directly compares and verifies the face of pass passenger, and if not the same person, it is alarmed, so as to ensure that the face of verification and the face of pass passenger are the same person.

Description

Technical Field

[0001] The present invention relates to the technical field of face recognition gate machines, and in particular to a method for determining impersonation through a face recognition gate machine and a face recognition gate machine. Background Art

[0002] With the development of facial recognition technology, facial recognition turnstiles are becoming increasingly popular. Facial recognition uses cameras to capture pedestrians' facial information as they move, quickly determining their access rights. This unique advantage makes facial recognition turnstiles seamless. They are increasingly popular in applications requiring high efficiency and a high user experience, such as airport security, boarding, and campus / hospital security.

[0003] In high-security situations such as airport security checks and airport access, it is necessary to verify that the face of the person passing through is the same person as the passenger passing through, otherwise it is easy to cause security accidents. Existing facial recognition gates cannot handle the following situations when performing facial recognition: (1) Passenger A completes the recognition at the front of the gate, and the gate wing opens. At this time, passenger B intentionally or unintentionally passes through the gate from the back, and passenger A ultimately fails to complete the passage; (2) Passenger A and passenger B stand in front of each other. Because passenger B's facial features are more obvious, the gate captures passenger B. After the gate wing opens, passenger A passes through, resulting in the verified face and the passenger passing through being different people, posing a security risk; (3) For dual-channel gates, passenger A stands in front of the gate channel, the gate captures the face of the person outside the next channel, and after the gate wing opens, passenger A passes through, resulting in the verified face and the passenger passing through being different people. In all three of the above situations, the verified face and the passenger passing through are different people, which is easy to cause security accidents and needs to be improved. Summary of the Invention

[0004] The main purpose of the present invention is to overcome the shortcomings of the existing technology, provide a method for determining the impersonation of users through facial recognition gates with secondary verification, and improve the security problems caused by the impersonation of users in existing facial recognition gates.

[0005] In order to solve the above-mentioned purpose, the solution of the present invention is:

[0006] A method for determining impersonation through a facial recognition gate with secondary verification, comprising the following steps:

[0007] S1: The face recognition terminal detects faces appearing in the camera video and records face trajectory information for different IDs;

[0008] S2: When a user performs facial verification, the ID of the passenger currently undergoing identity verification is determined to be the first ID, and the facial information of the first ID that has passed the verification is recorded, and all historical trajectory information is cleared at the same time;

[0009] S3: The face recognition terminal re-detects the face appearing in the camera video and records the face track information of different IDs;

[0010] S4: When the passenger face leaves the edge of the camera video close to the side of the gate channel, it is determined that the passenger is passing through the gate door, and the face information of the current ID is recorded;

[0011] S5: The best snapshot face of the passing passenger ID within a preset time is selected from the historical track, and if the face quality meets the requirements, the face is verified again;

[0012] S6: Whether the passing passenger face and the verification face are both the first ID face is compared, and if not, it is determined that the passing is false.

[0013] Further, the face recognition terminal is installed with a depth camera, and in S4, when the depth of the passenger face leaving the edge of the camera video is less than a preset distance, it is determined that the passenger is passing through the gate door.

[0014] Further, when the passenger leaves the edge of the face recognition terminal, if the depth information is missing, the last depth information of the passenger ID face track is taken, and if the last face depth is less than a preset distance, it is determined that the passenger is passing through the gate door.

[0015] Further, the face track information includes ID number, time point, face position and face depth; and the first ID face information includes ID number, time point, face position, face depth, face picture and face feature.

[0016] Further, in S4, when the passenger face leaves the edge of the camera video close to the side of the gate channel, the face appearing in the current camera is snapped, and the largest face snapped is selected, and when the largest face is less than a preset size, it is determined that the passenger with the largest face is passing through the gate door.

[0017] Further, in step 5, if the face with the quality meeting the requirements cannot be snapped, the loose strategy is not to verify the face again, and the strict strategy is to compare whether the passing passenger ID and the verification face ID are the same, and if not, it is determined that the passing is false.

[0018] After adopting the above structure, the face recognition terminal directly uses the passenger face imaging to leave the edge of the camera close to the side of the gate channel to verify the face again, directly compares the verification face and the passing passenger face information, and if they are not the same person, an alarm is given, so as to ensure that the verification face and the passing passenger are the same person and avoid false passing. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1 The present application is a passing process imaging schematic diagram.

[0020] Figure 2 The flowchart of the preferred embodiment of the present application.

[0021] Figure 3 The structure diagram of the face recognition terminal with a depth camera according to the present application. DETAILED DESCRIPTION

[0022] In order to further explain the technical solutions of the present application, the present application will be described in detail below through specific embodiments.

[0023] As shown in Figure 1 and Figure 2 The present application discloses a face recognition gate passing determination method with secondary verification, which comprises the following steps:

[0024] S1: The face recognition terminal 10 detects the face in the camera video through a face detection algorithm, and records the face track information of different IDs;

[0025] S2: In the case that there is a passenger for face verification, the passenger ID currently for identity verification is determined as the first ID, and the face information of the first ID that passes the verification is recorded, while all the historical track information of the first ID is emptied. The historical track information refers to the face (position), ID number, time and depth information recorded in a tracking record by the camera for each video frame. The tracking record of the same passenger in different frames is called historical track information, and the ID number of the same ID in the passing process does not change;

[0026] S3: The face recognition terminal 10 re-detects the face appearing in the camera video, and records the face track information of different IDs;

[0027] S4: When the passenger face leaves the edge of the camera video near the side of the gate passage, it is determined that the passenger is passing through the gate, and the face information of the current passenger ID is recorded;

[0028] S5: The best snapshot face of the passing passenger ID within a preset time is selected from the historical track information, and if the face quality meets the requirements, the face is subjected to secondary verification;

[0029] S6: Whether the face of the passing passenger and the face for verification are both the face of the first ID is compared, and if not, it is determined as passing by impersonation.

[0030] In step 5, if the face with quality meeting the requirements cannot be captured, the loose strategy is: no secondary face verification is performed, and it is determined as normal passing; the strict strategy is: whether the ID of the passing passenger and the ID of the face for verification are the same is compared, and if not, it is determined as passing by impersonation.

[0031] As shown in Figure 3As shown, as the preferred embodiment of the present application, the face recognition terminal 10 is installed with a depth camera 20, and the depth information of the face in the passage is calculated by the depth camera 20.

[0032] In the above steps, the face trajectory information and the historical trajectory information each include an ID number, a time point, a face position, and a face depth, and the face information includes an ID number, a time point, a face position, a face depth, a face picture, and a face feature.

[0033] In the above S4, when the depth camera detects that the passenger face is less than a preset distance (such as 0.7 meters) from the edge of the camera video on the side close to the gate passage, it is determined that the passenger is passing through the gate, and the face information of the passenger in the passage is recorded. The passenger face leaving the edge of the camera imaging is a necessary condition for the passenger to pass through, but the passenger face outside the gate passage can also be on the left and right edges of the camera imaging, and the RGB imaging cannot distinguish the two scenarios, and the depth information can accurately distinguish the two scenarios. The field of view angle of the depth camera is smaller than that of the ordinary RGB camera, and when the passenger leaves the left and right edges of the recognition terminal at the time point, if the depth information is lacking, the last depth information of the face trajectory of the same ID can be used to determine whether the passenger passes through the gate.

[0034] If the depth camera is not installed on the face recognition terminal 10, the face depth information cannot be obtained, and in this case, the face trajectory information and the historical trajectory information only contain an ID number, a time point, and a face position. In order to determine whether the passenger passes through the gate, the face size can be used instead of the distance, and when the maximum face is smaller than a preset size, the maximum face is taken as the passenger in the passage, and when the maximum face leaves the edge of the camera video close to the side of the gate passage, it is determined that the passenger is passing through the gate. It is compared whether the face of the passenger in the passage is the same as that of the verification passenger, and if not, it is determined that the passenger is passing through by impersonation, and an alarm is triggered.

[0035] The above embodiments and drawings are not limited to the product form and style of the present application, and any appropriate changes or modifications made by those skilled in the art shall be considered as not departing from the scope of the present application.

[0036] It should be noted that the information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data for analysis, stored data, displayed data, etc.), and signals (including but not limited to signals transmitted between user terminals and other devices, etc.) involved in the present application are all authorized by the user or fully authorized by all parties, and the collection, use, and processing of related data need to comply with relevant laws, regulations, and standards in relevant countries and regions. For example, the "face" and other related information or data involved in the present application are obtained under full authorization.

Claims

1. A secondary verification face recognition gate impostor pass determination method, characterized in that, The method comprises the following steps: S1: the face recognition terminal detects the face appearing in the camera video and records the face track information of different IDs; S2: in the case that a user performs face verification, the face recognition terminal determines the passenger ID currently performing identity verification as a first ID, the face of the passenger corresponding to the first ID as a verification face, records the first ID face information of the verification, and clears all historical track information; S3: the face recognition terminal re-detects the face appearing in the camera video and records the face track information of different IDs; S4: when the passenger face leaves the edge of the camera video close to the side of the gate channel, it is determined that the passenger is passing through the gate, and the passenger passing through the gate is a passing passenger, and the face information of the current ID is recorded; S5: the best captured face of the passing passenger ID within a preset time is selected from the face track information, and if the face quality meets the requirements, a second face verification is performed; S6: whether the face of the passing passenger and the verification face both correspond to the first ID face information is compared, and if not, it is determined that the passenger is passing by impersonation; Further, if the face recognition terminal is installed with a depth camera, the depth information of the face passing through the gate is calculated by using the depth camera; in S4, when the face of the passenger leaves the edge of the camera video close to the side of the gate channel, the depth information of the face is less than a preset distance, it is determined that the passenger is passing through the gate, and the face information of the current ID recorded includes the depth information of the face; If the face recognition terminal is not installed with a depth camera, in S4, when the face of the passenger leaves the edge of the camera video close to the side of the gate channel, the face appearing in the current camera is captured, and the largest face captured is selected; when the largest face is less than a preset size, it is determined that the passenger corresponding to the largest face is passing through the gate.

2. The secondary verification face recognition gate impost pass determination method of claim 1, wherein: For the case that the face recognition terminal is installed with a depth camera, when the face of the passenger leaves the edge of the camera video close to the side of the gate channel, if the depth information of the face is missing, the depth information of the last face of the face track information of the passenger ID is taken, and if the depth information of the last face is less than a preset distance, it is determined that the passenger is passing through the gate.

3. The secondary verification face recognition gate impostor pass determination method of claim 1, wherein: For the case that the face recognition terminal is installed with a depth camera, the face track information in S1 and S3 both includes the ID number, time point, face position and depth information of the face; the first ID face information in S2 includes the ID number, time point, face position, depth information of the face, face picture and face feature.

4. The secondary verification face recognition gate impost pass determination method of claim 1, wherein: In S5, if the face with quality meeting the requirements cannot be captured, a loose strategy is adopted: no second face verification is performed, and it is determined as normal passing; or a strict strategy is adopted: whether the ID of the passing passenger and the ID of the verification face are the same is compared, and if not, it is determined as passing by impersonation.

Images