A door access control method, device and electronic equipment

Abstract

The application provides a door access control method and device and electronic equipment. The application determines whether the terminal is in an idle state in a current time period according to the length of the current time period and the number of identification operations performed by the terminal in the current time period, and distributes target image data to be identified to the terminal in the idle state, so that the terminal in the idle state compares the target image data with an image matching template. That is, the application uses the terminal in the idle state to compare image data, avoids the problem of waste of computing resources caused by the terminal in the related art not performing other tasks after completing image data extraction, and improves the utilization rate of computing resources of the terminal in the idle state.

Description

Technical Field

[0001] This application relates to the field of security control technology, and in particular to an access control method, device and electronic equipment. Background Technology

[0002] With the development of the digital economy, the management of industrial parks, factories, residential communities, and schools is moving towards intelligent and digitalization. To facilitate the management of these facilities, access control systems are typically installed.

[0003] Access control systems typically consist of a host computer and multiple recognition terminals. The recognition terminals acquire facial images and extract image data, then report the extracted image data to the host computer. The host computer then compares the image data with the image matching template, and sends the comparison results to the recognition terminals. The recognition terminals then determine whether to unlock the access control system based on the comparison results.

[0004] In the above scheme, the image data comparison is performed by the host computer, while the recognition terminal only extracts the image data and does not perform image data comparison. After the recognition terminal completes the image data extraction, it is idle, resulting in a waste of the recognition terminal's computing resources. Summary of the Invention

[0005] To overcome the problems existing in related technologies, this application provides an access control method, device, and electronic device to improve the utilization rate of the computing resources of the identification terminal.

[0006] According to a first aspect of the embodiments of this application, an access control method is provided, the method being applied to an access control system, the access control system including a host and at least one identification terminal, the method comprising:

[0007] The first identification terminal collects the target image data to be identified and sends it to the host computer.

[0008] The host receives the target image data, performs a recognition operation on the target image data based on the image matching template assigned to the host, and selects at least one target recognition terminal that is idle in the current time period from all recognition terminals included in the access control system to send a comparison task to the target recognition terminal. The comparison task is used to instruct the recognition operation to be performed on the target image data. The current time period refers to a time period that includes the current time point and has a set duration. Whether any recognition terminal is idle in the current time period is determined based on the duration of the current time period and the number of recognition operations performed by the recognition terminal in the current time period.

[0009] When the host determines that one of the image matching templates assigned to the host matches the target image data, it controls the first recognition terminal to release the access control and sends a comparison stop command to the target recognition terminal;

[0010] The target recognition terminal receives a comparison task sent by the host, and performs a recognition operation on the target image data based on the comparison task and the image matching templates assigned to the terminal. If a comparison stop instruction is received during the recognition operation, the recognition operation on the target image data is stopped; otherwise, the recognition operation on the target image data continues to be performed to obtain a comparison result. If the comparison result indicates that one of the image matching templates assigned to the terminal matches the target image data, the comparison result is returned to the host.

[0011] After receiving the comparison result from the target identification terminal, the host controls the first identification terminal to unlock the access control.

[0012] According to a second aspect of the embodiments of this application, an access control device is provided, the device being applied to an access control system, the access control system including a host and at least one identification terminal, the device comprising:

[0013] The target image data transmission module is used by the first recognition terminal to collect target image data to be recognized and send it to the host.

[0014] The host identification module is used to receive the target image data, perform identification operations on the target image data based on the image matching template assigned to the host, and select at least one target identification terminal that is idle in the current time period from all identification terminals included in the access control system to send a comparison task to the target identification terminal. The comparison task is used to instruct the target image data to be identified. The current time period refers to a time period that includes the current time point and has a set duration. Whether any identification terminal is idle in the current time period is determined based on the duration of the current time period and the number of identification operations performed by the identification terminal in the current time period.

[0015] The first access control deactivation module is used by the host to control the first recognition terminal to deactivate the access control and send a comparison stop command to the target recognition terminal when the host determines that one of the image matching templates assigned to the host matches the target image data.

[0016] The target recognition terminal recognition module is used to receive a comparison task sent by the host, perform a recognition operation on the target image data based on the comparison task and the image matching templates assigned to the terminal, and if a comparison stop instruction is received during the recognition operation, the recognition operation on the target image data is stopped; otherwise, the recognition operation on the target image data continues to be performed to obtain a comparison result. If the comparison result indicates that one of the image matching templates assigned to the terminal matches the target image data, the comparison result is returned to the host.

[0017] The second access control deactivation module is used by the host to control the first identification terminal to deactivate the access control after receiving the comparison result from the target identification terminal.

[0018] According to a third aspect of the embodiments of this application, an electronic device is provided, the electronic device comprising: a processor and a memory; wherein the memory is configured to store machine-executable instructions; and the processor is configured to read and execute the machine-executable instructions stored in the memory to implement the method as described in the first aspect.

[0019] The technical solutions provided by the embodiments of this application may include the following beneficial effects:

[0020] This application determines whether the recognition terminal is idle during the current time period based on the duration of the current time period and the number of recognition operations performed by the recognition terminal within the current time period. The target image data is then sent to the idle recognition terminal, enabling it to compare the target image data with an image matching template. In other words, in this application, the recognition terminal can also perform image data comparison with an image matching template. Compared to related technologies where only the host compares image data with the image matching template, the solution provided in this application effectively alleviates the pressure on the host to perform large-scale image data comparisons with image matching templates, improving the efficiency of image data comparison. Furthermore, by using an idle recognition terminal for image data comparison, this application avoids the problem in related technologies where the recognition terminal, after extracting image data, stops processing other tasks, leading to wasted computing resources, thereby improving the utilization rate of computing resources in the idle recognition terminal.

[0021] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0022] The accompanying drawings, which are incorporated in and form part of this application, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0023] Figure 1This is a flowchart illustrating an access control method according to an embodiment of this application.

[0024] Figure 2 This is an overall flowchart of an access control method shown in an embodiment of this application.

[0025] Figure 3 This is a schematic diagram illustrating the execution time of a task according to an embodiment of this application.

[0026] Figure 4 This is a block diagram of an access control device shown in an embodiment of this application.

[0027] Figure 5 This is a hardware structure diagram of an electronic device shown in an embodiment of this application. Detailed Implementation

[0028] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0029] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used in this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0030] It should be understood that although the terms first, second, third, etc., may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0031] The embodiments of this application will now be described in detail.

[0032] Before describing the methods provided in the embodiments of this application, the access control system involved in the embodiments of this application will be described first:

[0033] In this embodiment, the access control system can be deployed in a distributed system. The access control system includes a host and at least one identification terminal. The host is deployed at the center of the distributed system and has high computing power. It can manage image matching templates, access control, etc. The identification terminals are deployed in various nodes of the distributed system. Each identification terminal can be deployed in one node of the distributed system. It has lower computing power and can collect face images, receive various tasks issued by the host, and display data processing results (e.g., door lock opening results).

[0034] Based on the access control system described above, the method provided in the embodiments of this application will be described as follows:

[0035] See Figure 1 , Figure 1 This is a flowchart illustrating an access control method according to an embodiment of this application. Figure 1 As shown, the access control method provided in this application may include the following steps:

[0036] S102: The first identification terminal collects the target image data to be identified and sends it to the host.

[0037] For example, in this embodiment, the first identification terminal is one of the identification terminals in the access control system. The first identification terminal is the identification terminal that currently collects a face image and needs to determine whether to open the door lock. When someone enters the monitoring range of this identification terminal, for example, when an employee needs to enter the park, at the access control point, the first identification terminal collects the employee's face image, extracts features from the face image to obtain target image data, and then sends the target image data to the host so that the host can send the target image data to each identification terminal.

[0038] It should be noted that the target image data can be either the data obtained by the first recognition terminal extracting features from the face image, or the data obtained by the first recognition terminal capturing the face image, transmitting the face image to the host, and the host extracting features from the face image.

[0039] As an example, in scenarios where the recognition terminal is unable to extract facial features or the recognition terminal has poor computing performance and takes a long time to extract features from facial images, the first recognition terminal can send the collected facial images to the host, and the host can extract features from the facial images to obtain target image data.

[0040] As another embodiment, when the recognition terminal is able to extract target image data of a face image, or when the recognition terminal has high computing performance and can complete the feature extraction of a face image in a short time, after acquiring the face image, the target recognition terminal can perform feature extraction on the face image to obtain target image data, and send the target image data to the host so that the host can receive the target image data.

[0041] S104: The host receives the target image data, performs recognition operation on the target image data based on the image matching template assigned to the host, and selects at least one target recognition terminal that is idle in the current time period from all the recognition terminals included in the access control system to send a comparison task to the target recognition terminal.

[0042] For example, in this embodiment, after the host receives the target image data, it performs an identification operation on the target image data based on the image matching templates that have been assigned to the host. That is, it compares the target image data with the image matching templates that have been assigned to the host one by one to determine whether there is a template in the image matching templates assigned to the host that matches the target image data.

[0043] In this embodiment, after the host receives the target image data, it also sends a comparison task to the target recognition terminal in an idle state, instructing it to perform a recognition operation on the target image data. Here, the comparison task carries the target image data, so both the host and the target recognition terminal perform recognition operations on the target image data.

[0044] In this embodiment, the idle state of any identification terminal is determined by the duration of the current time period and the number of identification operations performed by the terminal within that time period. Here, the current time period refers to a time period including the current time point and lasting for a set duration (e.g., one hour). Specifically, the duration of the current time period and the number of identification operations performed by the terminal within that time period are input into the working state decision model. This model then outputs the idle probability of each identification terminal in the current time period. The host then determines the target identification terminal in an idle state from among multiple identification terminals based on the idle probability of the identification terminal and a preset probability threshold. The working state decision model is pre-built, and its specific construction will be described below, and will not be elaborated upon here.

[0045] It should be noted that the aforementioned target recognition terminal may include a first recognition terminal, meaning that a recognition terminal that has completed image acquisition can also perform image data comparison. Furthermore, in this embodiment, the host can issue image matching templates to the host and all recognition terminals connected to the host. The image matching templates possessed by the host and recognition terminals may not overlap or may partially overlap, and the number of corresponding image matching templates may be different; that is, the host and each recognition terminal have image matching templates different from other recognition terminals or the host. As an example, the host can issue image matching templates to the host and each recognition terminal based on their computational performance parameters. Related details will be explained in detail below and will not be repeated here.

[0046] S106: When the host determines that one of the image matching templates assigned to the host matches the target image data, it controls the first recognition terminal to release the access control and sends a comparison stop command to the target recognition terminal.

[0047] For example, in this embodiment, if the host determines that one of the image matching templates assigned to the host matches the target image data during the recognition operation of the target image data, it indicates that the object corresponding to the target image data (e.g., an employee) has the right to enter the park. The host controls the first recognition terminal to unlock the door and open the door lock; at the same time, it sends a comparison stop command to each target recognition terminal so that each target recognition terminal stops performing recognition operations on the target image data.

[0048] S108: The target recognition terminal receives the comparison task sent by the host, and performs recognition operation on the target image data based on the comparison task and the image matching templates assigned to the terminal. If a comparison stop command is received during the recognition operation, the recognition operation on the target image data is stopped; otherwise, the recognition operation on the target image data continues to be performed to obtain the comparison result. If the comparison result indicates that one of the image matching templates assigned to the terminal matches the target image data, the comparison result is returned to the host.

[0049] For example, in this embodiment, after receiving the comparison task sent by the host, the target recognition terminal uses the image matching template assigned to itself to perform a recognition operation on the target image data carried in the comparison task, that is, to compare the target image data with the image matching templates assigned to the terminal one by one, so as to determine whether there is a template in the image matching templates assigned to the terminal that matches the target image data.

[0050] In this embodiment, if the target recognition terminal receives a comparison stop command from the host during the recognition operation, it stops performing the recognition operation on the target image data; otherwise, it continues to perform the recognition operation on the target image data to obtain the comparison result. If the comparison result indicates that one of the image matching templates assigned to this terminal matches the target image data, the comparison result is returned to the host.

[0051] S110: After receiving the comparison result from the target identification terminal, the host controls the first identification terminal to release the access control.

[0052] For example, in this embodiment, after receiving the comparison result from the target recognition terminal, the host controls the first recognition terminal to unlock the door and open the door lock. At the same time, it stops the currently ongoing recognition operation on the target image data and sends a comparison stop command to each of the other target recognition terminals. When other target recognition terminals receive the comparison stop command while performing recognition operations on the target image data based on the image matching template assigned to them, they stop performing recognition operations on the target image data.

[0053] Furthermore, if the host determines that none of the image matching templates assigned to it match the target image data, and no comparison result is received from the target recognition terminal within a set time period, it indicates that based on the current comparison result, the object corresponding to the target image data (e.g., an employee) does not have the right to enter the park, and the host controls the first recognition terminal to maintain access control.

[0054] Alternatively, if the host determines that none of the image matching templates assigned to it match the target image data, and no comparison result is received from the target recognition terminal within a set time period, it indicates that based on the current comparison result, the object corresponding to the target image data (e.g., an employee) does not have permission to enter the park. The host performs a recognition operation on the target image data based on the image matching templates assigned to other recognition terminals besides the target recognition terminal that it has cached. If it is found that none of the image matching templates assigned to other recognition terminals that it has cached match the target image data, it indicates that the object corresponding to the target image data (e.g., an employee) does not have permission to enter the park, and the host controls the first recognition terminal to maintain access control. If it is found that one of the image matching templates assigned to other recognition terminals that it has cached matches the target image data, it indicates that the object corresponding to the target image data (e.g., an employee) has permission to enter the park, and the host controls the first recognition terminal to unlock the access control and open the door lock.

[0055] It should be noted that the above-mentioned setting time can be set by the user based on experience, or it can be determined by the host based on the maximum time consumed by the target recognition terminal with the worst computing performance to match the target image data with the image matching template, or it can be determined in other ways. This embodiment does not make specific limitations on this.

[0056] As another embodiment of this application, if no template matching the target image data is found in any target recognition terminal, the target recognition terminal may report the message of no template matching to the host, which will then record it.

[0057] This concludes the process. Figure 1 The process is shown below.

[0058] pass Figure 1 The illustrated process enables the recognition terminal to compare image data with image matching templates. Compared to related technologies where only the host compares image data with image matching templates, the solution provided in this application effectively alleviates the pressure on the host to compare large amounts of image data with image matching templates, thus improving the efficiency of image data comparison. Furthermore, this application uses the recognition terminal in an idle state to perform image data comparison, avoiding the problem in related technologies where the recognition terminal stops processing other tasks after extracting image data, leading to wasted computing resources and thereby improving the utilization rate of computing resources of the recognition terminal in an idle state.

[0059] The following explains how to construct a work status decision model.

[0060] In this embodiment, the host can obtain a working state decision model through neural network model training, and use the working state decision model to detect and identify the idle probability of the terminal in the current time period.

[0061] Specifically:

[0062] Step A: The host counts the number of recognition operations of each recognition terminal in the access control system within each divided time period, and calculates the idle probability of each recognition terminal in each time period based on the number of recognition operations of each recognition terminal and the maximum number of recognition operations that each terminal can perform in each time period.

[0063] Step B: The host inputs the number of recognition operations and idle probability of each recognition terminal in each time period into the neural network for training to obtain the working state decision model.

[0064] For example, the host can divide a day into r time periods, each of which may have the same or different durations. For each recognition terminal, the host counts the number of recognition operations performed on image data by each terminal in each time period of the day. Based on the number of recognition operations performed on image data in each time period and the maximum number of recognition operations that each terminal can perform in each time period, the host calculates the idle probability of each recognition terminal in any given time period. The idle probability of each recognition terminal in any given time period can be expressed by the following formula:

[0065]

[0066] In the above formula, X is the idle probability of the recognition terminal in a time period, c is the number of recognition operations performed by the recognition terminal on the image data in that time period, and L is the maximum number of recognition operations that the recognition terminal can perform in that time period.

[0067] After determining the idle probability of each identification terminal in each time period, the host determines whether the identification terminal is in an idle state in each time period based on the idle probability of the identification terminal in each time period and the probability threshold corresponding to each time period, thus obtaining the initial working state decision model shown in Table 1. For example, if the idle probability of the identification terminal in a time period is less than the probability threshold corresponding to that time period, then the identification terminal is determined to be in a non-idle state in that time period; if the idle probability of the identification terminal in a time period is greater than or equal to the probability threshold corresponding to that time period, then the identification terminal is determined to be in an idle state in that time period.

[0068] Table 1

[0069]

[0070] Table 1 illustrates the number of operations performed by a single terminal within a given time period each day. Here, T1…Tr represents the time period, c1…cr represents the number of recognition operations performed by the terminal on image data within each time period, and X1…Xr represents the idle probability of the terminal in each time period. 0 indicates the terminal is not idle, and 1 indicates it is idle. Initially, the terminal is assumed to be in an idle state.

[0071] Then, the host can divide the number of recognition operations and idle probability of each recognition terminal in each time period into a training set and a test set. Based on the self-learning nature of the neural network algorithm, an initial working state decision model is trained using the training set, and the algorithm performance of this initial working state decision model is measured using the test set. Within an acceptable error range for the user, the working state decision model is output. After obtaining the working state decision model, the host can use it to determine the probability that each recognition terminal is in an idle state in each time period.

[0072] This completes the construction of the work status decision model.

[0073] After constructing the working state decision model, in practical applications, the host can use this model to determine the idle probability of any identification terminal in the current time period. Specifically:

[0074] First, the host inputs the duration of the current time period and the number of recognition operations performed by the recognition terminal within the current time period into the trained working state decision model to obtain the idle probability of the recognition terminal in the current time period. Here, the current time period refers to the time period selected from the divided time periods. For example, if a day is divided into 4 time periods, the host determines the time period in which the current time is located based on the time range corresponding to each time period.

[0075] Secondly, the host determines whether the identification terminal is in an idle state during the current time period based on the idle probability and a preset probability threshold.

[0076] For example, for each time period, each identification terminal will set a probability threshold (i.e., the aforementioned preset probability threshold). When the idle probability of the identification terminal in the current time period is obtained, it is determined whether the identification terminal is in an idle state in the current time period based on the idle probability and the preset probability threshold of the identification terminal in the current time period. For example, if the idle probability is greater than or equal to the preset probability threshold, it is considered that the identification terminal is in an idle state in the current time period; if the idle probability is less than the preset probability threshold, it is considered that the identification terminal is in a non-idle state in the current time period.

[0077] It should be noted that in practical applications, instead of constructing a working state decision model, the maximum number of recognition operations that each recognition terminal can perform within each time period can be stored in the host's storage area. When determining whether a recognition terminal is idle, the host can read the maximum number of recognition operations that each recognition terminal can perform within the current time period from the aforementioned storage area. Based on the number of recognition operations performed by each recognition terminal on image data within the current time period, and the maximum number of recognition operations that each recognition terminal can perform within the current time period, the host determines the probability that each recognition terminal is idle within the current time period. Then, by combining the probability of each recognition terminal being idle within the current time period with a preset probability threshold for each recognition terminal, the host can determine whether each recognition terminal is idle within the current time period.

[0078] By determining whether each recognition terminal is idle within the current time period, the idle target recognition terminal can be identified from multiple recognition terminals. Then, the idle target recognition terminal performs recognition operations on the target image data. This method can effectively improve the utilization rate of the recognition terminal's computing resources.

[0079] After identifying an idle target recognition terminal, the host sends a comparison task carrying the target image data to the idle target recognition terminal, enabling the idle target recognition terminal to compare the target image data with the image matching template assigned to it. As explained in the explanation of step 104, the number of image matching templates assigned to each recognition terminal and the host is determined by the computational performance parameters of the recognition terminal and the host. Therefore, before assigning image matching templates to the recognition terminal and the host, the computational performance parameters of each recognition terminal and the host must be determined.

[0080] The following explains how to determine the computing performance parameters of the identification terminal:

[0081] Step a: The host sends a performance test certificate to all identification terminals in the access control system. The performance test certificate includes at least: test image data and N test image templates for the test image data; N is greater than or equal to 1.

[0082] Step b: Any identification terminal receives the performance test certificate, compares the test image data in the performance test certificate with each test image template in the performance test certificate, and sends the time consumed to the host after the comparison is completed.

[0083] Step c: The host receives the time consumed and determines the computing performance parameters of the identification terminal based on the time consumed and N.

[0084] Step d: The host assigns image matching templates to each identification terminal and itself based on the computing performance parameters of each identification terminal and the computing performance parameters of the host. The computing performance parameters are positively correlated with the number of image matching templates assigned. The set of image matching templates assigned to each identification terminal and the set of image matching templates assigned to the host constitutes the set of image matching templates required by the intelligent access control system.

[0085] For example, in this embodiment, after the recognition terminal connects to the host, the host sends test image data and N test image templates to the recognition terminal. In subsequent use, when the recognition terminal cannot identify a template matching the test image data among the N test image templates, it indicates that the recognition terminal has completed all comparisons of the N test image templates. At this time, the host calculates the time T consumed by the recognition terminal in completing all comparisons of the N test image templates, and determines the computational performance parameter k of the recognition terminal based on this time consumption. n The calculation formula is as follows:

[0086] k n =T / N

[0087] In the above formula, k n This represents the computational performance parameter of the nth recognition terminal, where N represents the number of test image templates allocated to the recognition terminal, and T represents the time consumed by the recognition terminal to complete the comparison between the test image data and each test image template.

[0088] At this point, the computational performance parameters of the identification terminal have been determined.

[0089] It should be noted that whenever an identification terminal connects to the host, the host calculates the computing performance parameters of that identification terminal. Furthermore, the host's computing performance parameters can also be determined using the methods described above.

[0090] Furthermore, after determining the computational performance parameters of each identification terminal and the host, the host can assign image matching templates to the identification terminals and the host itself. That is, the host can assign image matching templates to each identification terminal and the host based on the computational performance parameters of both the identification terminal and the host itself. The set of image matching templates assigned to each identification terminal and the host constitutes the set of image matching templates required by the access control system. This set of image matching templates required by the access control system is cached in the host. It should be noted that even after the host assigns image matching templates to each identification terminal and the host itself, the host still caches the set of image matching templates required by the access control system.

[0091] As an example, the host can assign image matching templates to each recognition terminal and the host itself according to the following formula:

[0092] P x =Mk x / (k0+k1+…+k n )

[0093] Among them, P x This represents the number of image matching templates assigned to the x-th identification terminal or the host, M is the total number of image matching templates, k0 is the computing performance parameter of the host, and k1~k n These are the computational performance parameters for the 1st to nth identification terminals, respectively, k x This represents the computing performance parameter of the xth identification terminal or the host.

[0094] As can be seen from the above formula, the number of image matching templates allocated to each recognition terminal is positively correlated with the computing performance parameters of that recognition terminal. That is, the better the computing performance of the recognition terminal, the more image matching templates the host sends to it.

[0095] In this embodiment, after a new recognition terminal is detected to be connected to the host, the host calculates the computational performance parameters for the newly connected recognition terminal, and determines the number of image matching templates to be allocated to the newly connected recognition terminal and the number of image matching templates to be reallocated to other recognition terminals based on the computational performance parameters of the newly connected recognition terminal. Then, based on the number of image matching templates reallocated to other recognition terminals, an adjustment increment for the number of image matching templates for other recognition terminals is determined, and the image matching templates allocated to the newly connected recognition terminal are obtained from the image matching templates of the corresponding recognition terminals based on the adjustment increment.

[0096] For example, there are 10 recognition terminals already connected to the host, and each recognition terminal has 10 image matching templates. When a new recognition terminal is connected to the host, the host calculates and determines that 10 image matching templates should be allocated to the new recognition terminal. The host can then determine to obtain one (i.e., adjust the increment) image matching template from each of the recognition terminals already connected to the host and allocate the obtained image matching template to the newly connected recognition terminal.

[0097] For example, if 10 recognition terminals are already connected to a host, and each terminal has 10 image matching templates, and a new terminal connects to the host, the host calculates that 9 image matching templates should be allocated to the new terminal. The host can then determine the number of image matching templates to allocate from each of the existing 10 terminals based on their computational performance. For instance, it might allocate 2 image matching templates from terminals with lower computational performance to the new terminal, while allocating 1 or none from terminals with higher computational performance.

[0098] It should be noted that different recognition terminals have different computational performance parameters. Among them, the image matching templates available to the recognition terminal represent the computational load of the terminal performing recognition operations on image data. During the recognition operation, if the recognition terminal has a small number of image matching templates, it indicates insufficient computational load, which may cause the recognition terminal to enter an idle state after completing the recognition operation, waiting for the host to complete its recognition operation before the host can allocate tasks to it again, resulting in wasted computational resources. Conversely, if the recognition terminal has a large number of image matching templates, it indicates excessive computational load, which may cause the host to wait for the recognition terminal, resulting in host idleness and wasted computational resources.

[0099] Furthermore, recognition terminals can vary in computing performance, and allocating the same number of image matching templates may result in wasted computing resources for terminals with higher computing performance. Therefore, allocating a reasonable number of image matching templates to each recognition terminal based on its computing performance parameters can effectively avoid wasting computing resources on either the recognition terminal or the host computer.

[0100] As an example, typically, the image matching templates assigned to each recognition terminal are different. For instance, the image matching template assigned to recognition terminal 1 includes image matching template A, while the image matching template assigned to recognition terminal 2 does not include image matching template A. It is easy to note that because different image matching templates are assigned to different recognition terminals, the image matching templates compared by each recognition terminal do not overlap during the recognition operation. That is, each recognition terminal does not perform repetitive work, and the recognition operations are performed in parallel, enabling fast and efficient acquisition of comparison results.

[0101] It should be noted that in practical applications, during the process of assigning image matching templates to the recognition terminal, the host also considers the frequency of use of each image matching template (i.e., the frequency of use for feature comparison), and assigns image matching templates to the recognition terminal based on the frequency of use of each image matching template. The frequency of use of an image matching template represents the frequency with which the object corresponding to that image matching template (e.g., an employee) is recognized by the recognition terminal, that is, the frequency with which the object corresponding to the image matching template passes through the access control system.

[0102] Specifically, after the access control system has been running for a period of time (e.g., 1 month), the host can count high-frequency objects for at least one time period. High-frequency objects refer to objects included in the image matching templates that are used more frequently than a preset frequency within at least one time period within a specified time interval (i.e., the aforementioned time period). The host will then distribute the image matching templates that match the high-frequency objects to the host and each recognition terminal.

[0103] As an example, during the recognition operation, if the recognition terminal identifies a target image matching template that matches the target image data, the recognition terminal will report the comparison result to the host. Simultaneously, it will also report the template identifier of the target image matching template to the host, or the object information (e.g., employee name, employee number, etc.) of the objects contained in the target image matching template to the host, so that the host can calculate the usage frequency of the target image matching template within a preset time period. If the usage frequency of the target image matching template is greater than the preset frequency, then the objects contained in the target image matching template are determined to be high-frequency objects.

[0104] Furthermore, after identifying the high-frequency objects, the host will distribute the image matching templates that match the high-frequency objects to the host, each recognition terminal, and the host.

[0105] As an example, the host distributes image matching templates that match high-frequency objects to all recognition terminals and the host, and loads the image matching templates that match high-frequency objects in front of the template libraries corresponding to the host and all recognition terminals, so as to prioritize the comparison of image matching templates that match high-frequency objects when performing recognition operations.

[0106] As another embodiment, the host can also assign image matching templates that match high-frequency objects to the recognition terminal and the host according to the computing performance parameters of the recognition terminal and the host. For example, the host assigns a larger number of image matching templates corresponding to high-frequency objects to the recognition terminal with higher computing performance; while the host assigns a smaller number of image matching templates corresponding to high-frequency objects to the recognition terminal with lower computing performance. For example, the host assigns the top 10 most frequently used image matching templates to the recognition terminal with lower computing performance; while assigning the top 30 most frequently used image matching templates to the recognition terminal with higher computing performance.

[0107] This completes the allocation of image matching templates.

[0108] By allocating a reasonable number of image matching templates to the recognition terminal and host based on computing performance, the waste of computing resources of the recognition terminal or host can be effectively avoided, thereby improving the computing utilization rate of the recognition terminal.

[0109] The following describes the processing flow after the host detects the comparison results.

[0110] After the target image data is sent to the idle recognition terminal, both the host and the idle recognition terminal perform recognition operations on the target image data. The following explains the processing flow after the host detects the comparison result in two scenarios: the first recognition terminal extracts image data and the host extracts image data.

[0111] In the scenario where the first recognition terminal extracts image data, the first recognition terminal acquires the image to be recognized and extracts features from the image to obtain target image data. The target image data is then transmitted to the host so that the host can obtain the target image data and distribute the target image data to each recognition terminal in an idle state.

[0112] As an example, while the recognition terminal performs a recognition operation on the target image data, the host also performs a recognition operation on the target image data. If the host obtains a comparison result from the recognition operation on the target image data, the host controls each target recognition terminal to stop performing the recognition operation on the target image data.

[0113] For example, recognition terminal 1 acquires an image to be recognized and extracts the target image data. After reporting the target image data to the host, the host sends the target image data to all idle recognition terminals except recognition terminal 1 to perform recognition operations. If the host detects a template matching the target image data in its assigned image matching template or in the image matching template corresponding to a non-idle recognition terminal, the host sends a comparison stop command to all idle target recognition terminals to control each target recognition terminal to stop performing recognition operations on the target image data. Simultaneously, the host also controls recognition terminal 1 to unlock the access control.

[0114] As another embodiment, if any target recognition terminal performs a recognition operation on the target image data and obtains a comparison result, the host controls the host and other target recognition terminals to stop performing the recognition operation on the target image data.

[0115] For example, if target recognition terminal 1 performs a recognition operation on target image data and obtains a comparison result, then the host control host and other target recognition terminals will stop performing recognition operations on the target image data.

[0116] In one optional embodiment, when the target recognition terminal receives a feature extraction instruction during the process of performing a recognition operation on the target image data, the target recognition terminal will prioritize the feature extraction task. That is, the priority of the target recognition terminal in performing the feature extraction task is higher than the priority of performing the recognition operation.

[0117] In this scenario, if the target recognition terminal that received the feature extraction instruction has not yet completed the recognition operation on the target image data, the target recognition terminal will send the recognition progress of the target image data to the host, so that the host can complete the recognition operation that the target recognition terminal has not yet completed according to the recognition progress. For example, after receiving the recognition progress sent by the target recognition terminal, the host determines the image matching template corresponding to the target recognition terminal, determines which image matching templates have been matched and which have not yet been matched, and performs matching on the image matching templates that have not yet been matched.

[0118] For example, when recognition terminal 1 is performing a recognition operation, if a user enters the recognition range of recognition terminal A, recognition terminal 1 will capture the user's face image and extract features from the face image. At the same time, the recognition progress of recognition terminal 1 will be fed back to the host, which will complete the recognition operation that recognition terminal 1 has not completed. In subsequent recognition operations, the task of recognition operation of recognition terminal 1 will be temporarily assigned to other recognition terminals that are in an idle state until recognition terminal 1 returns to an idle state.

[0119] In another optional embodiment, when multiple recognition terminals have acquired face images and image comparison is required, after a certain recognition terminal completes the recognition operation on the previous target image data assigned by the host, the host sends the next target image data to that recognition terminal without waiting for other recognition terminals to complete the recognition operation on the previous target image data.

[0120] For example, if identification terminals 1 and 2 successively report target image data a and target image data b to the host, the host first sends target image data a to identification terminals 3 and 4, which are in an idle state. If identification terminal 3 completes the identification operation for target image data a but does not obtain a comparison result, and identification terminal 4 and the host also do not complete the identification operation for target image data a, then after identification terminal 3 reports the lack of a comparison result to the host, the host sends target image data b to identification terminal 3, and identification terminal 3 performs the identification operation for target image data b without waiting for other identification terminals to complete the comparison.

[0121] As an example, if no comparison result is detected within the set time period, or if the host finds that the target image data does not match the image matching template assigned to the host, the host performs a recognition operation on the target image data based on the image matching templates assigned to each recognition terminal other than the target recognition terminal that it has cached, and controls each target recognition terminal to stop performing the recognition operation on the target image data.

[0122] For example, if the host does not detect a template that matches the target image data in the image matching templates it has assigned, or does not receive the comparison result from the target recognition terminal within the set time period, the host will continue to compare the target image data from the image matching templates corresponding to other recognition terminals that are not idle.

[0123] It should be noted that the host can cache not only the image matching templates assigned to itself, but also the image matching templates assigned to each recognition terminal (including recognition terminals in an idle state and recognition terminals in a non-idle state).

[0124] This concludes the description of the target image data recognition operation in the scenario where the target recognition terminal extracts image data.

[0125] The following describes the recognition operation for target image data in the scenario where the host extracts image data.

[0126] In a scenario where the host extracts image data, the first recognition terminal acquires the image to be recognized and then sends the image to the host so that the host can extract features from the image to obtain the target image data.

[0127] As an example, in a scenario where the host extracts image data, the host sends the target image data to the first recognition terminal and the target recognition terminal in an idle state, so that the first recognition terminal and the target recognition terminal in an idle state can perform recognition operations on the target image data.

[0128] In this scenario, the recognition terminal only needs to acquire the face image, then send the acquired face image to the host, which will extract the features of the face image and send the extracted target image data to the idle target recognition terminal so that the idle target recognition terminal can perform the recognition operation.

[0129] It should be noted that in this scenario, the recognition terminal that captures facial images also performs the recognition operation. It's easy to observe that feature extraction tasks are typically more computationally intensive than recognizing the target image data. In this scenario, having the more computationally demanding feature extraction task performed by a more powerful host computer further reduces the hardware performance requirements of the recognition terminal.

[0130] The following combination Figure 2 The flowchart shown illustrates the access control method provided in this application as a whole.

[0131] like Figure 2As shown, the host assigns image matching templates to the recognition terminals based on their computing performance. When a user enters the recognition range of a certain recognition terminal (i.e., the first recognition terminal), the first recognition terminal acquires the user's facial image, extracts target image data from the facial image, and sends the extracted target image data to the host. After receiving the target image data sent by the first recognition terminal, the host checks whether there is an idle target recognition terminal among the recognition terminals connected to the host. If there is no idle target recognition terminal among the recognition terminals connected to the host, the host compares the target image data against the image matching templates corresponding to the n recognition terminals to find an image matching template that matches the target image data. If the host finds an image matching template that matches the target image data, the host controls the first recognition terminal to deactivate the access control and unlock the door.

[0132] If any of the target recognition terminals connected to the host is in an idle state, the host will send the comparison task carrying the target image data to the idle target recognition terminal. That is, the host will distribute the target to m (n greater than or equal to m) idle target recognition terminals. If the idle target recognition terminal does not need to recognize the image data, it will search for a template that matches the target image data in its corresponding image matching template. At the same time, the host will also search for a template that matches the target image data in its own allocated image matching template and in the image matching templates allocated to the recognition terminals that are not in an idle state. If the idle target recognition terminal has a feature extraction task to perform, it will stop performing the recognition operation on the target image data and perform the feature extraction task. At the same time, the host will search for a template that matches the target image data in its own allocated image matching template.

[0133] It should be noted that after identifying an idle target recognition terminal, the host sends the target image data to the idle terminal, thereby utilizing idle computing resources for face comparison and improving the overall utilization of the access control system's computing resources. Furthermore, when performing recognition operations on the target image data, the host's corresponding image matching template is dynamically adjusted based on the idle computing resources of the recognition terminal. For example, when the recognition terminal has sufficient idle computing resources, the host will allocate some of its image matching templates to that terminal, thus changing the host's image matching template as well.

[0134] In addition, in this embodiment, the host can receive image data from multiple recognition terminals, thereby enabling the host and recognition terminals to perform parallel processing tasks in a meaningful sense, further improving the speed of facial feature comparison.

[0135] Furthermore, it should be noted that when the host is under computational saturation, the total time for the host to perform the identification operation is L. If only the host is used for identification, and the number of cores in the host hardware is m1, then the shortest time for the host to perform the identification operation is L / m1. If the access control system has m2 idle target identification terminals, then the shortest time to perform the same identification operation is L / (m1+m2). For example, in... Figure 3 In the diagram illustrating the execution time of the identification operation, Figure 3 The comparison results show the execution time of performing the same identification operation using a dual-core host and four idle target identification terminals. Figure 3 It is known that using a dual-core host can effectively reduce the time it takes for the host to perform identification operations. Furthermore, by using the solution provided in this application, the idle computing resources of the identification terminal can be fully utilized, which can effectively reduce the time it takes for the host to perform identification operations with dual cores.

[0136] In addition, in actual calculations, uneven task allocation and the hardware performance gap between the host (single core) and the identification terminal (sub-core) may lead to inconsistent task completion times, resulting in an increased maximum execution time. However, compared to the scheme that only uses the host to perform the identification operation, the scheme provided in this application makes full use of the idle computing resources of the identification terminal and achieves fast calculation of large-scale computing tasks.

[0137] As described above, the solution provided in this application can determine the idle computing resources of the recognition terminal based on the current time period and the number of recognition operations performed by the recognition terminal within that time period. The host dynamically allocates target image data based on the idle computing resources of the recognition terminal, the computing performance of the recognition terminal, and personnel attendance, making full use of the computing resources in the entire access control system and significantly improving the utilization rate of the entire access control system's computing resources. In addition, the recognition operation for the target image data can be performed simultaneously on the host and each idle target recognition terminal, thereby achieving true parallel processing. Moreover, the more recognition terminals there are, the faster the feature comparison speed, providing users with a better user experience.

[0138] Corresponding to the embodiments of the aforementioned methods, this application also provides embodiments of access control devices and the electronic devices and storage media used therein.

[0139] like Figure 4 As shown, Figure 4 This is a block diagram illustrating an access control device according to an embodiment of this application. The access control device is applied to an access control system, which includes a host computer and at least one identification terminal, such as... Figure 4 As shown, the device includes:

[0140] The target image data transmission module is used by the first recognition terminal to collect target image data to be recognized and send it to the host.

[0141] The host identification module is used to receive target image data, perform identification operations on the target image data based on the image matching template assigned to the host, and select at least one target identification terminal that is idle in the current time period from all identification terminals included in the access control system to send a comparison task to the target identification terminal. The comparison task carries the target image data and is used to instruct the target image data to be identified. The current time period refers to a time period that includes the current time point and has a set duration. Whether any identification terminal is idle in the current time period is determined based on the duration of the current time period and the number of identification operations performed by the identification terminal in the current time period.

[0142] The first access control deactivation module is used by the host to control the first recognition terminal to deactivate the access control and send a comparison stop command to the target recognition terminal when the host determines that one of the image matching templates assigned to the host matches the target image data.

[0143] The target recognition terminal recognition module is used to receive the comparison task sent by the host, and perform recognition operation on the target image data based on the comparison task and the image matching templates assigned to the terminal. If a comparison stop command is received during the recognition operation, the recognition operation on the target image data is stopped; otherwise, the recognition operation on the target image data continues to be performed to obtain the comparison result. If the comparison result indicates that one of the image matching templates assigned to the terminal matches the target image data, the comparison result is returned to the host.

[0144] The second access control deactivation module is used by the host to control the first identification terminal to deactivate the access control after receiving the comparison result from the target identification terminal.

[0145] As an optional implementation of this application, whether any identification terminal is in an idle state during the current time period is determined by the following steps:

[0146] The host inputs the duration of the current time period and the number of recognition operations performed by the recognition terminal within the current time period into the trained working state decision model to obtain the idle probability of the recognition terminal in the current time period.

[0147] The host determines whether the identification terminal is idle in the current time period based on the idle probability and a preset probability threshold.

[0148] As an optional implementation of this application embodiment, the access control device further includes:

[0149] The performance test certificate sending module is used by the host to send performance test certificates to all identification terminals in the access control system. The performance test certificate includes at least: test image data and N test image templates for the test image data; N is greater than or equal to 1.

[0150] The consumption time sending module is used by any identification terminal to receive the performance test certificate, compare the test image data in the performance test certificate with each test image template in the performance test certificate, and send the consumption time consumed after the comparison is completed to the host.

[0151] The performance parameter determination module is used to determine the computing performance parameters of the identification terminal based on the host receiving the time consumed and N.

[0152] The image matching template allocation module is used by the host to allocate image matching templates to each recognition terminal and the host based on the computing performance parameters of each recognition terminal and the computing performance parameters of the host itself. The computing performance parameters are positively correlated with the number of image matching templates allocated. The set of image matching templates allocated to each recognition terminal and the set of image matching templates allocated to the host is the set of image matching templates required by the access control system.

[0153] As an optional implementation of this application, the image matching template allocation module is specifically used for:

[0154] Assign image matching templates to each recognition terminal and this host according to the following formula:

[0155] P x =Mk x / (k0+k1+…+k n )

[0156] Among them, P x This represents the number of image matching templates assigned to the x-th identification terminal or the host, M is the total number of image matching templates, k0 is the computing performance parameter of the host, and k1~k n These are the computational performance parameters for the 1st to nth identification terminals, respectively, k x This represents the computing performance parameter of the xth identification terminal or the host.

[0157] As an optional implementation of this application embodiment, the access control device further includes:

[0158] The high-frequency statistics module is used by the host to count high-frequency objects in at least one time period within a specified time interval. High-frequency objects refer to objects included in the image matching template that are used more frequently than a preset frequency in at least one time period within the specified time interval. The image matching template that matches the high-frequency objects is then distributed to the host and each recognition terminal.

[0159] As an optional implementation of this application embodiment, the access control device further includes:

[0160] The comparison stop module is used by the host to stop the ongoing recognition operation on the target image data after receiving the comparison result from the target recognition terminal, and to send a comparison stop command to all other target recognition terminals; other target recognition terminals, while performing recognition operations on the target image data based on the image matching template assigned to them, will also stop performing recognition operations on the target image data if they receive the comparison stop command; and / or,

[0161] The access control maintenance module is used to control the first recognition terminal to maintain access control if the host determines that none of the image matching templates assigned to the host match the target image data and no comparison result is received from the target recognition terminal within a set time period. Alternatively, it can perform recognition operations on the target image data based on the image matching templates assigned to other recognition terminals other than the target recognition terminal that have been cached by the host.

[0162] As an optional implementation of this application embodiment, the access control device further includes:

[0163] The idle probability determination module is used by the host to count the number of recognition operations of each recognition terminal in the access control system within each divided time period, and to calculate the idle probability of each recognition terminal in each time period based on the number of recognition operations of each recognition terminal and the maximum number of recognition operations that each terminal can perform in each time period.

[0164] The training module is used to input the number of recognition operations and the idle probability of each recognition terminal in each time period into the neural network for training, so as to obtain the working state decision model.

[0165] The specific implementation process of the functions and roles of each module in the above device can be found in the implementation process of the corresponding steps in the above method, and will not be repeated here.

[0166] This concludes the process. Figure 4 Description of the device shown.

[0167] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative, and the modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this application according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0168] Correspondingly, embodiments of this application also provide Figure 5 The hardware structure diagram of the electronic device shown is as follows: Figure 5 As shown, the electronic device can be a device implementing the above-described method. Figure 5 As shown, the hardware structure includes a processor and a memory. The memory stores machine-executable instructions; the processor reads and executes the machine-executable instructions stored in the memory to implement the corresponding access control method embodiment described above.

[0169] As one embodiment, the memory can be any electronic, magnetic, optical, or other physical storage device that can contain or store information such as executable instructions, data, etc. For example, the memory can be volatile memory, non-volatile memory, or similar storage media. Specifically, the memory can be RAM (Random Access Memory), flash memory, storage drives (such as hard disk drives), solid-state drives, any type of storage disk (such as optical discs, DVDs, etc.), or similar storage media, or combinations thereof.

[0170] This concludes the process. Figure 5 Description of the electronic device shown.

[0171] Based on the same inventive concept, this embodiment also provides a computer-readable storage medium. This computer-readable storage medium is used to store a computer program; when executed by a processor, the computer program implements the above-described method embodiment.

[0172] The foregoing has described specific embodiments of this application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired results. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0173] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention filed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not claimed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this application are indicated by the following claims.

[0174] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

[0175] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A door access control method characterized by, The method is applied to an access control system comprising a host and at least one identification terminal, the host having higher computing capability than the at least one identification terminal, and the method comprises: The first identification terminal collects target image data to be identified and sends the target image data to the host; The host receives the target image data, and if there is at least one target identification terminal in an idle state in a current time period among all the identification terminals included in the access control system, performs an identification operation on the target image data based on the image matching templates allocated to the host, and sends a comparison task to the target identification terminal, and if there is no target identification terminal in an idle state in the current time period among all the identification terminals included in the access control system, performs an identification operation on the target image data based on all the image matching templates cached by the host, the comparison task carrying the target image data, and the comparison task being used to instruct to perform an identification operation on the target image data; the current time period refers to a time period including a current time point and having a set time length, and whether any identification terminal is in an idle state in the current time period is determined based on the time length of the current time period and the number of identification operations performed by the identification terminal in the current time period; the image matching templates allocated to the host and the at least one identification terminal are determined by the host according to the computing performance parameters of the host and the at least one identification terminal; the image matching templates allocated to the host and the at least one identification terminal are non-overlapping, and the set of the image matching templates allocated to the host and the at least one identification terminal is the set of image matching templates required by the access control system, and the set of image matching templates required by the access control system is cached in the host; When the host determines that one of the image matching templates allocated to the host matches the target image data, the host controls the first identification terminal to release access control and sends a comparison stop instruction to the target identification terminal; The target identification terminal receives the comparison task sent by the host, performs an identification operation on the target image data based on the comparison task and the image matching templates allocated to the terminal, and if the comparison stop instruction is received in the process of performing the identification operation, stops performing the identification operation on the target image data, otherwise, continues to perform the identification operation on the target image data to obtain a comparison result, and if one of the image matching templates allocated to the terminal matches the target image data, returns the comparison result to the host; After receiving the comparison result of the target identification terminal, the host controls the first identification terminal to release access control.

2. The method of claim 1, wherein, Whether any identification terminal is in an idle state in a current time period is determined by the following steps: The host inputs the time length of the current time period and the number of identification operations performed by the identification terminal in the current time period into a trained working state decision model to obtain an idle probability of the identification terminal in the current time period; The host determines whether the identification terminal is in an idle state in the current time period according to the idle probability and a preset probability threshold.

3. The method of claim 1, wherein, The method further comprises the following steps before the method: The host sends a performance test credential to all identification terminals in the access control system, the performance test credential at least including: test image data and N test image templates of the test image data; N is greater than or equal to 1; Any identification terminal receives the performance test credential, and compares the test image data in the performance test credential with each test image template in the performance test credential respectively, and sends a consumed time length consumed in the comparison to the host; The host receives the consumed time length, and determines a computing performance parameter of the identification terminal according to the consumed time length and the N; The host allocates image matching templates to each identification terminal and the host according to the computing performance parameters of each identification terminal and the computing performance parameter of the host, the computing performance parameter being positively correlated with the number of allocated image matching templates; and a set of image matching templates allocated to each identification terminal and a set of image matching templates allocated to the host are a set of image matching templates required by the access control system.

4. The method of claim 3, wherein, The allocating image matching templates to each identification terminal and the host according to the computing performance parameters of each identification terminal and the computing performance parameter of the host comprises: allocating image matching templates to each identification terminal and the host according to the following formula: wherein, represents the number of image matching templates assigned to the xth identification terminal or the host computer, and M represents the total number of image matching templates, is a computing performance parameter of the host computer, ~ are computing performance parameters of the 1st to nth identification terminals, respectively, represents a computing performance parameter of the xth identification terminal or the host computer.

5. The method according to claim 1 or 3, characterized in that, The method further comprises: The host counts high-frequency objects in at least one time period in a specified time interval, the high-frequency object being an object contained in an image matching template used more frequently than a preset frequency in the at least one time period in the specified time interval; and the image matching template matched with the high-frequency object is allocated to the host and each identification terminal.

6. The method of claim 1, wherein, The method further comprises: The host stops a currently ongoing identification operation on the target image data after receiving a comparison result of the target identification terminal, and sends a comparison stop instruction to each other target identification terminal; each other target identification terminal stops the identification operation on the target image data if the comparison stop instruction is received in a process of performing the identification operation on the target image data based on the image matching template allocated to the terminal; and / or If the host determines that all image matching templates allocated to the host are not matched with the target image data, and no comparison result of the target identification terminal is received within a set time length, the first identification terminal is controlled to maintain the access control, or the identification operation on the target image data is performed based on the image matching templates allocated to each identification terminal other than the target identification terminal and cached by the host.

7. The method of claim 1, wherein, The method further comprises: The host counts identification operation times of each identification terminal in each time period divided in the access control system, and calculates an idle probability of each identification terminal in each time period according to the identification operation times of each identification terminal and a maximum identification operation time that can be performed by each terminal in each time period; The host inputs the identification operation times and the idle probabilities of each identification terminal in each time period into a neural network to obtain a working state decision model.

8. An access control device, characterized by The device is applied to an access control system, the access control system comprising a host and at least one identification terminal, the host having higher computing capability than the at least one identification terminal, and the device comprising: a target image data sending module configured to collect target image data to be identified by the first identification terminal and send the target image data to the host; a host identification module configured to receive the target image data by the host, if there is at least one target identification terminal in an idle state in a current time period among all identification terminals included in the access control system, perform an identification operation on the target image data based on an image matching template allocated to the host, and send a comparison task to the target identification terminal, and if there is no target identification terminal in an idle state in the current time period among all identification terminals included in the access control system, perform an identification operation on the target image data based on all image matching templates cached by the host, the comparison task carrying the target image data, and the comparison task being used to instruct to perform an identification operation on the target image data; the current time period refers to a time period including a current time point and having a set time length, whether any identification terminal is in an idle state in the current time period being determined based on the time length of the current time period and the number of identification operations performed by the identification terminal in the current time period; the image matching templates allocated to the host and the at least one identification terminal are determined by the host according to computing performance parameters of the host and the at least one identification terminal; the image matching templates allocated to the host and the at least one identification terminal are non-overlapping, and a set of the image matching templates allocated to the host and the at least one identification terminal is a set of image matching templates required by the access control system, and the set of image matching templates required by the access control system is cached in the host; a first access control release module configured to control the first identification terminal to release access control and send a comparison stop instruction to the target identification terminal when the host determines that one of the image matching templates allocated to the host matches the target image data; a target identification terminal identification module configured to receive the comparison task sent by the host by the target identification terminal, perform an identification operation on the target image data based on the comparison task and the image matching template allocated to the terminal, stop the identification operation on the target image data if the comparison stop instruction is received during the identification operation, otherwise, continue the identification operation on the target image data to obtain a comparison result, and return the comparison result to the host if one of the image matching templates allocated to the terminal matches the target image data; a second access control release module configured to control the first identification terminal to release access control after the host receives the comparison result of the target identification terminal.

9. The apparatus of claim 8, wherein, Whether any identification terminal is in an idle state in a current time period is determined by the following steps: the host inputs the time length of the current time period and the number of identification operations performed by the identification terminal in the current time period into a trained working state decision model to obtain an idle probability of the identification terminal in the current time period. The host determines whether the identified terminal is in an idle state in a current time period according to the idle probability and a preset probability threshold; The device further comprises: The performance test credential sending module is configured to send, by the host, a performance test credential to all identified terminals in the access control system, the performance test credential comprising at least: test image data and N test image templates of the test image data, where N is greater than or equal to 1; The consumption time length sending module is configured to receive, by any identified terminal, the performance test credential, and compare the test image data in the performance test credential with each test image template in the performance test credential respectively, and send a consumption time length consumed in the comparison to the host; The computing performance parameter determination module is configured to receive, by the host, the consumption time length, and determine a computing performance parameter of the identified terminal according to the consumption time length and the N; The image matching template allocation module is configured to allocate, by the host, image matching templates to each identified terminal and the host according to the computing performance parameter of each identified terminal and the computing performance parameter of the host, the computing performance parameter being positively correlated with the number of allocated image matching templates, and a set of image matching templates allocated to each identified terminal and the host being an image matching template set required by the access control system; The image matching template allocation module is specifically configured to: allocate image matching templates to each identified terminal and the host according to the following formula: wherein, represents the number of image matching templates assigned to the xth identification terminal or the host computer, and M represents the total number of image matching templates, is a computing performance parameter of the host computer, ~ are computing performance parameters of the 1st to nth identification terminals, respectively, represents a computing performance parameter of the xth identification terminal or the host computer; The device further comprises: The high-frequency statistical module is configured to count, by the host, high-frequency objects in at least one time period in a specified time interval, the high-frequency object being an object contained in an image matching template used more frequently than a preset frequency in at least one time period in the specified time interval, and allocate image matching templates matched with the high-frequency objects to the host and each identified terminal; The device further comprises: The stop comparison module is configured to stop, by the host, an identification operation currently being performed on the target image data after receiving a comparison result of the target identified terminal, and send a comparison stop instruction to each other target identified terminal, and each other target identified terminal stops performing the identification operation on the target image data after receiving the comparison stop instruction in a process of performing the identification operation on the target image data based on an image matching template allocated to the terminal; and / or The access control maintenance module is configured to control the first identified terminal to maintain access control if the host determines that all image matching templates allocated to the host are not matched with the target image data, and no comparison result of the target identified terminal is received within a set time length, or perform the identification operation on the target image data based on image matching templates allocated to each identified terminal other than the target identified terminal and cached by the host; The device further comprises: The idle probability determination module is configured to count, by the host, the number of identification operations of each identified terminal in the access control system in each time period divided, and calculate an idle probability of each identified terminal in each time period according to the number of identification operations of each identified terminal and the maximum number of identification operations that can be performed by each terminal in each time period. The training module is configured to input the number of identification operations and the idle probability of each identified terminal in each time period into the neural network for training to obtain the working state decision model.

10. An electronic device, comprising: The electronic device includes a processor and a memory; The memory is configured to store machine executable instructions. The processor is configured to read and execute the machine executable instructions stored in the memory to implement the method according to any one of claims 1 to 7.

Images