A gate access picture issuing method and system, an intermediate server and a storage medium

Abstract

The application relates to the technical field of access control, and in particular to an access control picture issuing method and system, an intermediate server and a storage medium, which comprise the following steps: an intermediate server acquires an access request of an access control device and responds to the access request to enable the access control device to access; a face picture issued by a platform and a target device are acquired, at least three feature points and feature data in the face picture are acquired based on random selection logic, the at least three feature data are packed with corresponding selection logic to obtain a feature set; corresponding device information is matched according to the target device to send the feature set to the corresponding access control device; a feedback signal sent by the access control device is acquired, and the feedback signal is forwarded to the platform; and the feedback signal represents a feedback sent after the access control device stores and processes the feature set after receiving the feature set. The application has the effect of improving the processing capacity and efficiency of face picture issuing.

Description

Technical Field

[0001] This application relates to the technical field of access control, and in particular to a method, system, intermediate server and storage medium for sending access control images. Background Technology

[0002] Access control systems are commonly used security products, widely applied in residential, corporate, factory, and shopping mall environments. By installing access control systems at main entrances and important locations and using methods such as facial recognition, password input, and fingerprint recognition, access can be opened and closed, effectively managing and controlling entry and exit security.

[0003] One of the main tasks of facial recognition access control is the distribution of facial images. Users or administrators need to transmit facial images of people who can enter and exit through the access control system to the access control device via the network. P2P technology is used to solve the link problem. On the platform, P2P hole punching technology can directly connect to the device through NAT. However, with the addition of various complex situations at the user site, the distribution speed is greatly reduced, whether it is P2P or forwarding. Sometimes only a few hundred facial images can be distributed per day. However, the platform is often connected to a large number of access control devices, and each access control device also needs to process hundreds or thousands of facial images. This makes the required computing and processing volume extremely large, and the processing difficulty extremely high. Summary of the Invention

[0004] To improve the processing capability and efficiency of face image distribution, this application provides a method, system, intermediate server, and storage medium for distribution of access control images.

[0005] Firstly, this application provides a method for distributing access control images, which employs the following technical solution:

[0006] A method for distributing access control images includes the following steps:

[0007] Obtain an access request from an access control device and respond to the access request to enable the access control device to access the device, wherein the access request includes device information of the access control device;

[0008] Obtain the face image and target device sent by the platform, and obtain at least three feature points from the face image based on random selection logic;

[0009] Obtain feature data of at least three feature points, and package the at least three feature data and their corresponding selection logic to obtain a feature set;

[0010] Based on the target device, match the corresponding device information to send the feature set to the corresponding access control device;

[0011] The system acquires the feedback signal emitted by the access control device and forwards the feedback signal to the platform. The feedback signal represents the feedback emitted by the access control device after receiving, storing, and processing the feature set.

[0012] Preferably, at least three feature points are obtained from the face image based on a random selection rule, specifically including the following steps:

[0013] The face is divided into three feature areas along the vertical direction: upper, middle and lower. The lower area extends from the lowest point of the chin to the highest point of the upper lip. The middle area extends from the highest point of the upper lip to the highest point of the bridge of the nose. The upper area extends from the highest point of the bridge of the nose to the highest point of the forehead.

[0014] At least one feature point is selected from each of the three feature regions to obtain at least three feature points.

[0015] Preferably, selecting at least one feature point in each of the three feature regions further includes the following steps:

[0016] Determine whether there are identical symmetrical feature points at symmetrical positions along the vertical line from the center of the face;

[0017] If such a point exists, it is defined as a backup feature point, and its feature data is obtained and packaged together into the feature set.

[0018] Preferably, determining whether there are identical symmetrical feature points at symmetrical positions along the vertical line from the center of the face further includes the following steps:

[0019] If it does not exist, a circular area is obtained with the feature point as the center and a preset distance as the radius, and the area outside the circular area in the feature area where the feature point is located is defined as the spare feature point selection area;

[0020] Select a backup feature point within the backup feature point selection area, obtain the feature data corresponding to the backup feature point, and package it into the feature set.

[0021] Preferably, after obtaining three feature points, the following steps are also included:

[0022] Determine whether one of the aforementioned feature points is a specific feature point, wherein the specific feature point is characterized by a feature with a specific identification function, including at least moles, spots, and scars;

[0023] If so, the specific feature point is defined as a special identification point. The special identification point represents the situation where, when the access control equipment performs face recognition, all other feature points in the feature set except for the specific feature point cannot be recognized, but the specific feature point can be recognized, the face recognition passes.

[0024] If not, the feature point located in the upper region is defined as a special case identification point.

[0025] Preferably, after obtaining the facial image and target device from the platform, the following steps are also included:

[0026] At least three verification points are obtained from the face image based on a fixed selection logic;

[0027] Obtain feature data corresponding to at least three of the verification points and package them to obtain a verification feature set;

[0028] Determine whether a set exists in the duplicate image verification library that is identical to the set of verification features;

[0029] If it exists, then define the face image as a duplicate image;

[0030] If the face image does not exist, the face image is defined as a new image, and the set of verification features is stored in the duplicate image verification library. At the same time, at least three feature points in the face image are obtained based on random selection logic.

[0031] Preferably, the method further includes the following steps: if present, the face image is defined as a duplicate image, and the method further includes the following steps:

[0032] Determine whether the time difference between the entry time of the duplicate verification feature set in the duplicate image verification library and the current time is greater than a preset time length;

[0033] If so, at least three feature points and corresponding feature data in the face image are re-obtained based on random selection logic and sent to the access control device to update the feature points.

[0034] If not, feature points in the face image are obtained based on random selection logic, and duplicate upload error information is returned to the platform.

[0035] Secondly, this application provides an access control image distribution system, which adopts the following technical solution:

[0036] An access control image distribution system includes an intermediate server, a platform, and access control devices, wherein...

[0037] The platform sends facial images and target devices to the intermediate server based on user operations;

[0038] The intermediate server is used to obtain access requests from access control devices and respond to the access requests to enable access control devices to access the system. The access requests include device information of the access control devices. It also obtains a face image and target device from the platform, and acquires at least three feature points from the face image based on a random selection logic. Furthermore, it acquires feature data for the at least three feature points, packages the feature data with their corresponding selection logic to obtain a feature set, and matches the target device with the corresponding device information to send the feature set to the corresponding access control device.

[0039] The access control device is used to acquire the feature set, store and process the feature set, and then generate a feedback signal;

[0040] The intermediate server receives the feedback signal and forwards it to the platform.

[0041] Thirdly, this application provides an intermediate server, which adopts the following technical solution:

[0042] An intermediate server, based on the Libevent framework, includes a number of kernel-level TCP protocol threads and a number of kernel-level HTTPS protocol threads. The TCP protocol threads are used to connect the access control device and the platform, and the HTTPS protocol threads are used to initiate HTTPS requests to the platform and forward the HTTPS requests sent by the platform to the access control device.

[0043] Fourthly, this application provides a computer storage medium, which adopts the following technical solution:

[0044] A computer storage medium storing a computer program, which, when executed by a processor, implements the access control image distribution method.

[0045] In summary, this application includes at least one of the following beneficial technical effects:

[0046] 1. By extracting features from facial images uploaded to the platform, a feature set is obtained. When distributing data to access control devices, the data set is distributed instead of the entire image. This transforms the existing image transmission and processing into the transmission and processing of the feature set. Since the feature set can contain at least three feature points, the workload during transmission is far less than that of transmitting the entire image. This enables the distribution of tens of thousands of images per day, improving the processing capacity and efficiency of facial image distribution.

[0047] 2. Using TCP transmission, an intermediate server can handle 30,000 connections simultaneously. The number of threads is set according to the Linux server kernel. All request and response are asynchronous, greatly enhancing processing capabilities. Attached Figure Description

[0048] Figure 1 This is a flowchart illustrating the access control image distribution method in an embodiment of this application;

[0049] Figure 2 This is an example diagram illustrating feature points, backup feature points, and verification points in an embodiment of this application;

[0050] Figure 3 This is a schematic diagram of the module connection between the access control image distribution system and the intermediate server in an embodiment of this application. Detailed Implementation

[0051] The following is in conjunction with the appendix Figure 1-3 This application will be described in further detail.

[0052] This application discloses a method for sending access control images.

[0053] like Figure 1 and Figure 2 As shown, a method for sending access control images includes the following steps:

[0054] S100: Obtain the access request from the access control device and respond to the access request to enable the access control device to access. The access request includes the device information of the access control device.

[0055] Access control devices must first connect to an intermediate server to establish connectivity. Only then can all access control devices connected to the intermediate server perform subsequent image downloading tasks.

[0056] The device information includes the device number, device location, device permissions, and device processing capacity of the access control device.

[0057] S200: Obtain the face image and target device sent by the platform, and obtain at least three feature points from the face image based on random selection logic.

[0058] Users send facial images and target devices through the platform. The platform can be a mobile terminal platform, such as a mobile phone, laptop, or tablet, or a fixed terminal platform, such as a desktop computer.

[0059] Three feature points are the minimum number of intermediate servers required to minimize processing power while ensuring security. Feature analysis of each face image is achieved by acquiring at least three feature points. The target device is the information needed to subsequently send at least three feature points to the corresponding access control device.

[0060] Specifically, the following steps are included:

[0061] S210 divides the face into three feature regions along the numerical direction of the face: upper region, middle region, and lower region. The lower region extends from the horizontal line where the lowest point of the chin is located to the horizontal line where the highest point of the upper lip is located. The middle region extends from the horizontal line where the highest point of the upper lip is located to the horizontal line where the highest point of the bridge of the nose is located. The upper region extends from the horizontal line where the highest point of the bridge of the nose is located to the horizontal line where the highest point of the forehead is located.

[0062] S220, select at least one feature point in each of the three feature regions to obtain at least three feature points.

[0063] The face is divided into three regions, each centered around a uniquely characterized organ. For example, the lower region is the mouth and chin, the middle region is the nose, and the upper region is the eyes and eyebrows.

[0064] In each feature region, at least one feature point is selected. This can prevent the positions of several feature points from being too concentrated when randomly selecting feature points, thus improving security.

[0065] Meanwhile, the random selection logic means that at least one feature point is selected for each feature region, and the selected feature point within each feature region can be random. For example, the upper region can select a corner of one eyebrow or a point on the inner side of the other eye; the lower region can select a corner of the upper lip or a point on the chin. The selection logic is randomized differently for each face image, further enhancing security.

[0066] Selecting at least one feature point in each of the three feature regions also includes the following steps:

[0067] S221, determine whether there are identical symmetrical feature points at symmetrical positions along the vertical line from the center of the face.

[0068] S222, if it exists, then define the same symmetrical feature point as a spare feature point, and subsequently obtain the feature data of the spare feature point and package them together into the feature set.

[0069] like Figure 1 and Figure 2 As shown, in some cases, users of access control devices may obscure certain areas of their faces due to certain circumstances, or certain areas may be unintentionally invalidated. In order to ensure security and recognition rate when the number of feature points is reduced, a certain number of backup feature points can be set. When normal feature points cannot be recognized, in order to avoid the situation where recognition is not possible due to obstruction, the backup feature points can be identified for judgment.

[0070] The first step is to find if a feature point has a symmetrical counterpart. The symmetry benchmark here is the vertical line from the center of the face. This vertical line is a near-vertical line passing through the center of the eyebrows, eyes, nose, and mouth, dividing the face vertically into two symmetrical sides. If a feature point is located on one side, and a symmetrical feature point with the same characteristics exists on the other side, then that point is a symmetrical feature point. Feature symmetry here means that the original feature point is located on a unique organ, such as the eyes, nose, mouth, or ears, rather than on the cheeks or forehead. Furthermore, after symmetry, the symmetrical point is also located on the other side of the same organ. For example, a point on the left eye becomes a point on the right eye after symmetry; these two points are symmetrical and have the same characteristics, thus qualifying as symmetrical feature points.

[0071] Through the above steps, if a user takes a photo of their face at the access control device to extract a facial feature point located in the left eye, and the left eye happens to be closed due to an anomaly while the right eye is open, the user can be identified by comparing and matching the symmetrical feature point on the right eye; or if the feature point is located in the left ear, and the user happens to be making a phone call or wearing a Bluetooth headset through the left ear when taking the photo, the user can be identified through the right ear. Therefore, selecting symmetrical feature points can ensure security while maintaining an effective recognition rate even when feature points are missing.

[0072] Furthermore, it also includes:

[0073] S223, if it does not exist, then take the feature point as the center and the preset distance as the radius to obtain a circular area, and define the area outside the circular area in the feature area where the feature point is located as the spare feature point selection area.

[0074] S224, Select a backup feature point in the backup feature point selection area, and subsequently obtain the feature data corresponding to the backup feature point and package it into the feature set.

[0075] When a feature point is located at or near the center, or in a location where there are no symmetrical feature points, such as the tip of the nose, chin, or cheek, alternative feature points can be obtained through other methods.

[0076] First, draw a circle with the feature point as the center and a preset distance as the radius. This circle represents a short distance range. If the backup feature point is selected within this circular area, the newly selected backup feature point will be close to the original feature point. On the one hand, this may cause confusion during subsequent recognition. On the other hand, if the original feature point is occluded or cannot be recognized, the backup feature point that is very close may also be occluded or cannot be recognized.

[0077] Therefore, after obtaining the circular region, select a feature point in the remaining area after subtracting the circular region from the feature area where the feature point is located. This ensures that the newly selected backup feature point is at a certain distance from the original feature point.

[0078] In other embodiments, after obtaining the three feature points, the following steps are also included:

[0079] S230, determine whether a certain feature point is a specific feature point among several feature points.

[0080] Specific feature points are characteristics with specific identifying functions, including at least moles, spots, and scars. These specific feature points can serve as unique markers for a person to a certain extent, because people with moles, spots, and scars on their faces are generally located in different places, and it is rare for people to have the same mole or scar in the same place. Therefore, a person's identity can be quickly identified through these specific markers.

[0081] S231, if so, then the specific feature point is defined as a special case identification point.

[0082] The special case verification point is characterized as follows: when the subsequent access control equipment performs face recognition, if all feature points in the feature set except for the specific feature point cannot be recognized, but the specific feature point can be recognized, the face recognition passes.

[0083] In other words, if a person has a specific feature among their many feature points, then that specific feature is a special identification point. If all feature points except that specific feature are obscured during subsequent access control facial recognition, but this specific feature can still be identified, that is, a specific mole, spot, or scar is identified, then passage can be granted directly because these features are relatively specific and there are very few cases of recognition errors.

[0084] S232, if not, then the feature points located in the upper region are defined as special case identification points.

[0085] If there are no specific feature points, feature points located in the upper region can be selected as feature recognition points.

[0086] Since many people wear masks daily, they sometimes don't want to or forget to take them off when using access control. This means that the area below the user's nose is basically covered by the mask. In this case, facial recognition can be performed on the feature recognition points in the upper area. As long as the feature points in the upper area are recognized, passage can be achieved.

[0087] It should be noted that the use of special case identification points is only possible when all feature points except for that point are occluded.

[0088] like Figure 1 and Figure 2As shown, in some other embodiments, after obtaining the face image and target device issued by the platform, the following steps are also included:

[0089] S240, obtain at least three verification points in the face image based on fixed selection logic.

[0090] S241, Obtain feature data corresponding to at least three verification points, and package them to obtain a set of verification features.

[0091] S242, Determine whether there exists a set in the duplicate image verification library that is the same as the verification feature set.

[0092] S243, if it exists, then define the face image as a duplicate image.

[0093] Verification points are primarily used to determine whether duplicate face images have been sent to the platform. First, a fixed selection logic is established, i.e., a fixed selection number. Each time a face image arrives, a set of verification features is generated using these fixed verification points and compared with a set in the duplicate image verification database. Since the set of verification points corresponding to a new face image is sent to the duplicate image verification database for storage whenever a new face image is sent, if the comparison results are the same, it means the face image has been sent before; if the comparison results are different, it means the face image has not been sent. Then, feature point extraction and subsequent steps are performed on the new face image, and the verification feature set of this face image is added to the duplicate image verification database for storage.

[0094] Specifically, it also includes,

[0095] S2431, determine whether the difference between the entry time of the duplicate verification feature set in the duplicate image verification library and the current time is greater than the preset time length.

[0096] S2432, if so, then re-obtain at least three feature points and corresponding feature data from the face image based on random selection logic and send them to the access control device to update the feature points.

[0097] S2433, if not, then obtain the feature points in the face image based on random selection logic, and return the duplicate upload exception information to the platform.

[0098] S244, if it does not exist, then define the face image as a new image, store the verification feature set in the duplicate image verification library, and obtain at least three feature points from the face image based on random selection logic.

[0099] If an image is a duplicate image, then determine whether the time difference between the current time the image was sent and the time the image was first sent and entered into the database is greater than a preset time.

[0100] If the current distribution time is May 20th, the database entry time (first distribution) is May 10th, and the preset time is 1 month, then if the difference is not greater than the preset value, it means that the image was distributed repeatedly within a short period of time. In this case, the duplicated image will be returned, and corresponding duplicate upload error information will be generated. However, if the current distribution time is May 20th, and the database entry time is February 10th, then the time difference is greater than the preset time. In this case, the feature values ​​of the currently distributed face image will be re-obtained based on random selection logic to update the data and improve security.

[0101] S300: Obtain feature data of at least three feature points, and package the at least three feature data and their corresponding selection logic to obtain a feature set.

[0102] Feature data for several feature points is obtained. This data includes the coordinates of each feature point, the distances between them, and their identifiers. The location information of the feature points is obtained from a pre-uploaded facial feature map. This map labels the various feature locations on the face with several points, assigning a corresponding identifier to each point. The distances between feature points are calculated by converting the distance from each feature point to a center point, typically the tip of the nose. To determine the distance between point A and point B, simply calculate the distances from A to the center point and from B to the center point; this will allow you to calculate the distance between A and B.

[0103] By packaging the feature data and selection logic of each feature point, a set is obtained, which serves as the object subsequently transmitted to the access control device. Specifically, the selection rule here refers to which feature points with specific numbers are selected, and these numbers change randomly for each different face image.

[0104] S400 matches the corresponding device information based on the target device to send a feature set to the corresponding access control device.

[0105] S500 acquires feedback signals from access control devices and forwards them to the platform. The feedback signals represent the feedback sent by the access control devices after receiving, storing, and processing the feature set.

[0106] After receiving the feature set from the intermediate server, the access control device stores and processes it, and sends feedback information to the intermediate server. The intermediate server then forwards the feedback information to the platform for users to view.

[0107] like Figure 3 As shown in the illustration, this application also discloses an access control image distribution system, including an intermediate server, a platform, and access control devices, wherein...

[0108] The platform sends facial images and target device information to an intermediate server based on user actions.

[0109] The intermediate server is used to obtain access requests from access control devices and respond to access requests to enable access control devices to connect. The access request includes the device information of the access control device; obtain face images and target devices issued by the platform, and obtain at least three feature points in the face image based on random selection logic; obtain feature data of at least three feature points, package the at least three feature data and their corresponding selection logic to obtain a feature set; match the corresponding device information according to the target device to send the feature set to the corresponding access control device;

[0110] Access control equipment is used to acquire feature sets, store and process the feature sets, and then generate feedback signals;

[0111] The intermediate server receives the feedback signal and forwards it to the platform.

[0112] This application also discloses an intermediate server based on the Libevent framework, which includes a number of kernel-level TCP protocol threads and a number of kernel-level HTTP(S) protocol threads. The TCP protocol threads are used to connect the access control device and the platform, and the HTTP(S) protocol threads are used to initiate HTTP(S) requests to the platform and forward the HTTP(S) requests issued by the platform to the access control device.

[0113] It also includes a TCP listening service, the size of which is determined by the number of deployed devices, and is generally recommended to be between 8,000 and 10,000.

[0114] TCP protocol thread:

[0115] Thread framework: It is recommended to start threads corresponding to the number of kernels and bind them to the corresponding kernels. After the listening service accepts a connection, perform simple load balancing to distribute the connections relatively evenly among the processing threads. This can make full use of the functions of each kernel.

[0116] Interaction message format: The protocol header between the access control device and the intermediate server is fixed at 16 bytes, the message is in JSON format, and the JSON part is decrypted using a private protocol.

[0117] HTTP(s) protocol thread:

[0118] Thread framework: It is recommended to start threads corresponding to the number of kernels and bind them to the corresponding kernels. After the listening service accepts a connection, perform simple load balancing to distribute the connections relatively evenly among the processing threads. This can make full use of the functions of each kernel.

[0119] Interaction message format: The protocol header between the access control device and the intermediate server is fixed at 16 bytes, the message is in JSON format, and the JSON part is decrypted using a private protocol.

[0120] The HTTP(s) protocol thread is set to a 30-second timeout. If no response is received from the device within 30 seconds, a timeout is returned to the platform.

[0121] If the corresponding UUID cannot be found in the linked list (or other storage method) of the TCP protocol thread, it means that the device is offline.

[0122] The specific communication process is as follows:

[0123] The access control device is mainly connected to the central server (after successful registration, the device will send a heartbeat to the central server every 20 seconds to keep it alive; if the central server does not receive a heartbeat for 20*n (n is recommended to be 2 or 3) seconds, it considers the device offline).

[0124] The platform initiates an HTTP(s) request;

[0125] The central server forwards HTTP(s) requests from the platform;

[0126] After processing the information, the access control equipment sends a response to the central server.

[0127] Central server response platform.

[0128] This application also discloses a computer storage medium, in which a computer program, when executed by a processor, implements the above-described method for sending access control images.

[0129] The implementation principle is as follows:

[0130] By extracting features from facial images uploaded to the platform, a feature set is obtained. When distributing data to access control devices, the data set is distributed to the feature set instead of the entire image. This transforms the existing image transmission and processing into the transmission and processing of the feature set. Since the feature set can contain at least three feature points, the workload during transmission is far less than that of transmitting the entire image. This allows for the distribution of tens of thousands of images per day, improving the processing capacity and efficiency of facial image distribution.

[0131] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A method for sending access control images, characterized in that, Includes the following steps: Obtain an access request from an access control device and respond to the access request to enable the access control device to access the device, wherein the access request includes device information of the access control device; Obtain the face image and target device sent by the platform, and obtain at least three feature points from the face image based on random selection logic; Obtain feature data of at least three feature points, and package the at least three feature data and their corresponding selection logic to obtain a feature set; Based on the target device, match the corresponding device information to send the feature set to the corresponding access control device; The system acquires the feedback signal emitted by the access control device and forwards the feedback signal to the platform. The feedback signal represents the feedback emitted by the access control device after receiving, storing, and processing the feature set. At least three feature points are obtained from the face image based on a random selection rule, specifically including the following steps: The face is divided into three feature areas along the vertical direction: upper, middle and lower. The lower area extends from the lowest point of the chin to the highest point of the upper lip. The middle area extends from the highest point of the upper lip to the highest point of the bridge of the nose. The upper area extends from the highest point of the bridge of the nose to the highest point of the forehead. At least one feature point is selected from each of the three feature regions to obtain at least three feature points.

2. The access control image distribution method according to claim 1, characterized in that: Selecting at least one feature point in each of the three feature regions also includes the following steps: Determine whether there are identical symmetrical feature points at symmetrical positions along the vertical line from the center of the face; If such a point exists, it is defined as a backup feature point, and its feature data is obtained and packaged together into the feature set.

3. The access control image distribution method according to claim 2, characterized in that: Determining whether there are identical symmetrical feature points at symmetrical positions along the vertical line from the center of the face, further includes the following steps: If it does not exist, a circular area is obtained with the feature point as the center and a preset distance as the radius, and the area outside the circular area in the feature area where the feature point is located is defined as the spare feature point selection area; Select a backup feature point within the backup feature point selection area, obtain the feature data corresponding to the backup feature point, and package it into the feature set.

4. The access control image distribution method according to claim 1, characterized in that: After obtaining the three feature points, the following steps are also included: Determine whether one of the aforementioned feature points is a specific feature point, wherein the specific feature point is characterized by a feature with a specific identification function, including at least moles, spots, and scars; If so, the specific feature point is defined as a special identification point. The special identification point represents the situation where, when the access control equipment performs face recognition, all other feature points in the feature set except for the specific feature point cannot be recognized, but the specific feature point can be recognized, the face recognition passes. If not, the feature point located in the upper region is defined as a special case identification point.

5. The access control image distribution method according to claim 1, characterized in that: After obtaining the facial image and target device information from the platform, the following steps are also included: At least three verification points are obtained from the face image based on a fixed selection logic; Obtain feature data corresponding to at least three of the verification points and package them to obtain a verification feature set; Determine whether a set exists in the duplicate image verification library that is identical to the set of verification features; If it exists, then define the face image as a duplicate image; If the face image does not exist, the face image is defined as a new image, and the set of verification features is stored in the duplicate image verification library. At the same time, at least three feature points in the face image are obtained based on random selection logic.

6. The access control image distribution method according to claim 5, characterized in that: It also includes the following steps: If it exists, the face image is defined as a duplicate image, and the following steps are also included: Determine whether the time difference between the entry time of the duplicate verification feature set in the duplicate image verification library and the current time is greater than a preset time length; If so, at least three feature points and corresponding feature data in the face image are re-obtained based on random selection logic and sent to the access control device to update the feature points. If not, feature points in the face image are obtained based on random selection logic, and duplicate upload error information is returned to the platform.

7. An access control image distribution system, characterized in that: This includes intermediate servers, platforms, and access control equipment, among which, The platform sends facial images and target devices to the intermediate server based on user operations; The intermediate server is used to obtain access requests from access control devices and respond to the access requests to enable access control devices to access the system. The access requests include device information of the access control devices. It also obtains a face image and target device from the platform, and acquires at least three feature points from the face image based on a random selection logic. Furthermore, it acquires feature data for the at least three feature points, packages the feature data with their corresponding selection logic to obtain a feature set, and matches the target device with the corresponding device information to send the feature set to the corresponding access control device. The access control device is used to acquire the feature set, store and process the feature set, and then generate a feedback signal; The intermediate server receives the feedback signal and forwards it to the platform; The process of obtaining at least three feature points from the face image based on a random selection rule specifically includes the following steps: The face is divided into three feature areas along the vertical direction: upper, middle and lower. The lower area extends from the lowest point of the chin to the highest point of the upper lip. The middle area extends from the highest point of the upper lip to the highest point of the bridge of the nose. The upper area extends from the highest point of the bridge of the nose to the highest point of the forehead. At least one feature point is selected from each of the three feature regions to obtain at least three feature points.

8. An intermediate server, characterized in that: The access control image distribution system described in claim 7 is based on the Libevent framework and includes a number of kernel-level TCP protocol threads and a number of kernel-level HTTPS protocol threads. The TCP protocol threads are used to connect the access control device and the platform, and the HTTPS protocol threads are used to initiate HTTPS requests to the platform and forward the HTTPS requests sent by the platform to the access control device.

9. A computer storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the access control image distribution method according to any one of claims 1 to 6.

Images